The tech sector is the most software-defined and virtual of all sectors, it's natural that they are first to change. Other industries need to make capital investments to adapt to change.
You can see a similar history with electrification. A few industries went quickly, but then it took like a decade of little growth before the true benefits showed up. As an example, take a factory that was built around hyrdo-power like a grain mill. You can't just take an electric motor and drop it in in place of the water wheel. It doesn't fit, the wheel is outside, there isn't room between the mill and the wall, etc. Only when you completely rebuild a new mill building designed around the features and limitations of the electric motor do you see the major benefits.
Same thing here; it will be new factories and new companies that are designed from the ground up around AI and its benefits and limitations that really move the needle.
Cool analogy, but there's another possibility: software is uniquely suited for being boosted by LLM-based coding agents. That's not to say LLMs are useless in other domains, but they may not yield as much benefit.
Software creation tools have been worked on for decades to keep fallible human coders out of trouble as much as possible. Starting with compilers that enforce or at least encourage coding style humans can understand. Compilers embody formal definitions of language syntax. Lint and other static analysis technologies contribute by identifying problematic patterns in code. Leak detectors identify memory management problems. Profilers identify performance problems.
Few other endeavors have this much tooling that serendipitously happens to help help LLM-based agents be more useful and less error prone.
A lot of discussion around AI is very high-level and abstract. It helps to dig in and identify the characteristics that actually make a domain more or less amenable to being boosted by AI.
I think LLM's will prove to be useful in any domain where most work happens in front of a computer. As LLM's make programming cheap, they also reduce the cost of building new software that automates work currently performed by humans in other fields.
If you provide best models with all the data and context an average white collar worker has to perform their job, in most cases they will do just fine. But in real world that context is often scattered over multiple systems, or sometimes exists only in people's heads. I believe that's the main bottleneck for adoption now.
You're missing the biggest suggestion from the GP's post: that the key to LLM success is verifiability.
Software has high verifiability through compilers, linters, static analysis, leak/memory corruption detection tools, or the very least, running the program to see if it crashes.
Math (as has been making the headlines recently) has high verifiability through automated theorem provers. Notably some of the non-mechanized math results from LLMs have been demonstrated to be wrong (and presumably many more are being checked; these aren't exactly easy to understand proofs).
Game playing (chess, go, Star Craft, etc.) is similarly verifiable: you get your score and you always know how you did in the end (even if not in the moment).
Making assumptions about LLMs in non-veriable domains is making a claim about LLMs that doesn't match their current capabilities. Maybe they'll get there and maybe they won't; I can speculate as well as you can but we're both guessing, so I'm not sure how much value the speculation has.
I think we'll naturally develop intuition and infrastructure for fault-tolerant ways to implement LLMs in non-verifiable domains, i.e. requiring multiple, independent agents agree on a direction from multiple angles before taking an action.
This works well when the answer to the question is a number or something else definitive. When the answer is fuzzy, how do you compare 3 different responses? They'll all be different
A lot of testing and verification was avoided in the past because it wasn’t considered nearly as valuable as shipping code. Many of our jobs are evolving to building verifiers of AI output.
If you believe the hype/marketing around how scary "AI" is LLMs are perfectly happy to go rouge and lie, cheat, and even blackmail to protect themselves.
This assumes that a significant portion of the average white collar worker’s job is amenable to automation.
But most white collar workers I know spend most their time either talking to each other, or thinking real hard about what to say the next time they get together to talk. If that’s generally true then it would suggest that there isn’t much white collar work that LLMs, which are now widely regarded as exceptionally poor communicators, can do particularly well.
I think the bear case (see:short term) benefit is that any and all information gathering, management and parsing the employee did between organizing with other employees will be automated, allowing the employee to do more organizing over the same unit of time.
The bull case (see: long term) is that the “organizing and talking” as long as it’s not something that is essentially human (like caretaking, nursing etc.) will happen between personal agents of sorts with only the deeper judgments being left to the humans opening up even more increases in productivity. This is already happening with PR reviews in a lot of workplaces.
Though I’m guessing most workplaces outside of tech will require total reorganization around this reality, but more likely this way of running will mean newer upstarts challenging incumbents refusing to change as it’s always been.
> thinking real hard about what to say the next time they get together
LLMs can help tremendously with this part, though. Exploration/brainstorming, discovery, data analysis, etc can be made significantly easier with LLMs in most knowledge based fields.
They certainly can, but I wonder if the cost will be deskilling your employees and leading brainstorming sessions towards AI psychosis.
I've seen much time wasted on strange ideas originated by an LLM. It's a lot easier for a human to organically have good ideas than to refute LLM-generated ideas that are a dead end.
If people are in charge of project management, and repeatedly fail to recognise poor ideas and instead pursue them to the detriment of productivity, then they probably shouldn't be managing that project.
All AI is doing here is rapidly identifying the people who have Peter-principled themselves.
> It's a lot easier for a human to organically have good ideas than to refute LLM-generated ideas that are a dead end.
What works better is when you can use an LLM to obtain a negative opinion on your own ideas that you think are good. LLMs aren't great at original thinking, but they're pretty good at finding the flaws in poorly thought-through human ideas.
This is certainly what I saw. I was at a company that was a very eager early adopter of GenAI for all work. One big thing that came out of it was a bunch of effort wasted on unprofitable projects.
The chatbots kept steering people toward uncreative ideas and overshadowing novel insights that might come from genuine domain expertise. AI generated meeting summaries became a force for miscommunication as they de-emphasized or even outright misrepresented what was really important about the conversation. And people became much worse at proving ideas out before making significant resource commitments thanks to effects that we now describe with terms like AI psychosis.
I work for big tech. We don’t have human in the loop anymore. Ai tools handle 99%. It’s not irresponsible if done right. And the next billionaire will show that ai can be tamed, managed, and humans are the problem.
For one, your company ought to publish its Scope 1, Scope 2, and Scope 3 carbon emissions for all to see. If you understand how the electricity is produced for the AI data centers your company uses, you’ll see that replacing the humans with natural gas guzzling AI workflows is a dramatic rape on the environment and health of species/ecosystems. But what I say doesn’t matter because all tech is in on AI and it’s a race to the bottom for the almighty dollar.
Depends on which data centers. They aren’t all equal. Some use less water and less energy than others. The data centers being more efficient isn’t the company using the service problem.
That’s what they say about communism and anarchism and many other isms: if it’s not working, you aren’t doing it right. Every failed implementation gets the No True Scotsman argument.
What differentiation? Most work is bread-and-butter after you get a few years of experience. If you LLM's can automate 50% of that, it would transform society already.
That “after a few years’ experience” bit is the catch. What’s happening during that time is a whole bunch of skill acquisition, building mental domain models, building context- and culture-specific soft skills, etc.
All of it relies on learning and information processing capabilities that LLMs simply do not have. The idea that any of it will automatically emerge from the current architectures simply by adding more parameters is about as sensible as thinking that if you keep counting long enough eventually you’ll get to complex numbers. We’re going to need to invent new technology to fill these gaps, and there is no way of knowing how long those developments will take.
"Beseech the god machine that it may shower upon you its tokenous beneficence."
The god machine can prompt itself. Just go "Write me a prompt to <whatever>. Ask me as many clarifying questions as needed.". It is not a differentiator.
Deep domain knowledge and pattern matching that allows you to quickly verify outputs might be a differentiator. Maybe. But you might end up the bottleneck if you verify everything.
They're in the process of training the models for those professions, too. They're offering people hundreds of dollars an hour to run recordings of workload they create and questions they ask. It's only a matter of time that companies sell access to those models and sell it as that profession.
> But most white collar workers I know spend most their time either talking to each other, or thinking real hard about what to say the next time they get together to talk.
is that true though ? let's think about finance say a simple M&A deal for TV show rights ?
i think us engineers n i'm assuming u r one - we tend to underestimate the complexities in the world specially in worlds governed by 'words'.
AIUI, M&A lawyering is _mostly_ formulaic. They copy a term sheet from a previous deal, replace with new names and figures, and then negotiate from that starting point. Only rarely will someone actually innovate and come up with some clause that's brand new. The core of it really is talking to the other party's lawyers, then updating docs to reflect your discussions.
>But most white collar workers I know spend most their time either talking to each other, or thinking real hard about what to say the next time they get together to talk.
Maybe that's just something which made sense when they were necessary to do the not yet automatable informattion processing work, but that only took like 20% of their time, so they added meetings, and useless work to keep them busy.
I don’t believe there is any fundamental reason LLMs can’t do this task, and the base models are probably already capable of it. RL to elicit the right behavior just hasn’t been done yet. Coding got solved first because it was cheaper and easier to get the necessary data for RL environments, which made it the natural first market opportunity. Solving white collar communication will require more resources to collect the data and build the environments, but the labs now have more resources and are working on it.
I want to believe reasons why this won’t work, but it feels like cope. It would be smart to be prepared for the world where knowledge workers get replaced by AI and also prepared for the world where AI is a bubble that pops.
The goalposts on what neural nets can possibly do have already moved past chess, go, protein folding, coding, poker, math, the Turing test…
"It would be smart to be prepared for the world where knowledge workers get replaced by AI and also prepared for the world where AI is a bubble that pops."
lol aka you have zero conviction about what you are saying.
So whats the point of you posting? Zero value add. In fact bordering on negative value.
You want to drive across town. You, only willing to deal in certainties, must decide ahead of time whether you will get in a crash. If you determine won’t get in a crash, you make the drive, pedal to the metal. If you determine you would crash, you stay home.
Me, willing to deal in probabilities, make the drive at the speed I feel reasonable and wear a seatbelt.
As the evidence from software programming shows, a lot of white collar workers spend most of their time talking frankly because they are inept at execution.
That margin just needs to be lower with LLMs than it is with only humans. In a lot of fields, that's a low bar. For example, medical professionals make a shocking amount of mistakes in aggregate.
Sure, but the average LLM makes a shocking number of mistakes in a single conversation.
BTW the medical professional mistake rate is often reported per-patient, not per-decision. That is, of the dozens of things recorded on each patient's chart, one being wrong counts the same as all of them being wrong. It's an easy way to inflate a negligible error rate into one that scares readers.
A doctor's agent listens in on a patient's appointment. It automatically sorts out prescriptions based on the conversation and performs other mundane tasks.
Perhaps the doctor uses some sort of "safe phrase" to ensure the agent doesn't get distracted by conversational theorizing and discussion.
The agent does decide to improve the prescription and does so because that possibility hadn't been considered in testing.
Warfarin (blood thinner) was prescribed for a haemophiliac and they are now no longer a "burden" on the NHS!
> It automatically sorts out prescriptions based on the conversation and performs other mundane tasks.
As we all know, the agents are really good at maintaining long contexts and doing parallel tasks without degradation.
> Perhaps the doctor uses some sort of "safe phrase" to ensure the agent doesn't get distracted by conversational theorizing and discussion.
Yes, there certainly is some magic word that stops an agent from context degradation, and prunes all work to just a useful subset. Sadly, no one knows what it is.
Yeah, I think it's at the level of self-driving Excel.
Now, Excel also makes mistakes while trying to get clever - remember the biologists who had to rename several genes to prevent Excel treating them as dates - but it's also a ubiquitous business tool. And you can combine them. But equally Excel hasn't replaced every white collar job.
In 90s the show Friends, Chandler Bing’s job is “statistical analysis and data reconfiguration” which is likely unfamiliar to you because it is gone!
This was a job to manually collate financial reports into aggregated sums to sent up the hierarchy to the next level up.
This is now a trivial BI query and doesn’t require manual processing by humans.
A bigger shift is that now I only see CEOs or at most CxOs with secretaries. Mobile phones and Outlook calendars have largely wiped out that profession.
Is it? I don’t really hire for analysts anymore because you can mostly just expose chat with your data mcp tools that end users can use and that gets you 80% of the way there. The person who puts together dashboards and writes sql against the star schema is pretty much a thing of the past.
Data engineers, on the other hand… way more relevant than ever before
The issue is that data complexity is also exploding in the age of LLMs, so it still takes a lot of archaeology to get reliable (as opposed to plausible) numbers.
I think this dynamic also requires more data engineers, but I don't personally see the split in demand between those roles that you describe. Data analysts often have business or domain context that data engineers lack.
Perhaps we'll start rolling data analysis work into "operations".
It's the difference between Architects and Draftsmen. We still need the former, but the latter is gone, replaced by Solidworks, AutoCAD, and digital plotters or large-format printers.
> A bigger shift is that now I only see CEOs or at most CxOs with secretaries. Mobile phones and Outlook calendars have largely wiped out that profession.
Depends heavily on the field. Most fields don't use the term "secretary", but e.g. my mid-sized company of several hundred employees has a single executive assistant, vs. my wife works at a law firm where it's pretty close to 1:1 professionals to assistants.
> This is now a trivial BI query and doesn’t require manual processing by humans.
It also doesn't require AI. Management wants to see how the same measurements change over time. They don't necessarily want the latest AI model's hot take on this month's data. Or maybe they do, but I doubt it's very helpful.
Replicating the brain functionality is magic? Or the brain itself is magic, so we need magic to replicate its functionality? Nah, learn more.
An LLM with CoT is Turing-complete, so there's no wall to its capabilities. There's no theorems that establish limits of reinforcement learning either. There's no evidence of quantum computations in the brain.
C++ is Turing complete. Hell, HTML+CSS is Turing complete.
> so there's no wall to its capabilities.
There are multiple well understood and well explained walls and limits to LLM capabilities. And none of those walls, or perceived lack of them, stem from something being Turing complete.
> There's no theorems that establish limits of reinforcement learning either.
There's a very obvious and immediate one: availability and size of training data. And compute. And time.
> There's no evidence of quantum computations in the brain.
What does quantum computation have to do with brains or LLMs?
> All you have is a strong belief.
Says the person who just listed a lot of beliefs in marketing bullshit.
Oh! Do you know how does the brain function? Heck, we don't know in detail how LLMs do their thing. The trivial observation that the brain in not a Turing machine is irrelevant if we are talking about functionality and not computational substrate.
But a Turing machine that implements functionality of a target system is the information-theoretical optimal compression of the data produced by the target system. So it stands to reason that LLMs that need to compress vast amounts of data approximate such a machine.
> C++ is Turing complete. Hell, HTML+CSS is Turing complete.
This is absolutely irrelevant to the question of the limits of LLM training. It's a common surface-level patter matching response, though.
> There's a very obvious and immediate one: availability and size of training data. And compute. And time.
Yeah, practical limitations. And practical results like seven hundred solved math problems show that you need, at least, to stop and think.
> What does quantum computation have to do with brains or LLMs?
If you have to ask that, it means that you haven't dug into all this mess deep enough. Basically, BQP computational complexity class is more powerful than P-class. If the brain employs such computations (although we don't know the appropriate quantum algorithms, so it's just wishful thinking mostly), the corresponding Turing machine is impractically large and slow.
> Says the person who just listed a lot of beliefs in marketing bullshit.
It's like Software, its value is to be a complement and/or an abstraction of existing businesses. All of that sits on the very basics of the essence of civilization, which is how much things we can get off the soil and what how quickly and effective we can transform it.
Not to be anti-tech, but every new advancement, no matter how next-level, has to sell more of the same goods and services people have been using since the dawn of civilization, and those goods and services need to solve the mundane problems people have been getting since millenia.
Software industry, because the construction industry still requires a human to sign off and approve the computer-aided design.
There are companies which are involved with building and marketing their own software for the construction industry, but they are also actively involved in the construction process in a consultative capacity. Only then could they even be classified as construction industry companies.
BI is a good example of something where the benefits take a long time to show up. It has the most use at a senior management level where they can't use their every day experience to tell what is going on. But they don't make a lot of decisions over the course of a year and it often takes a few years before the impact is felt.
I worry about this. Dumping data and getting a report on which actions are based prolly isn’t giving real data.
I say that because I have Claude review code and then review the review and it will point out all the things it got wrong the first time. And I end up having to review it myself. Don’t get me wrong, I’m burning tokens all day and love the legit productivity gains but I don’t see how this replaces everybody in tech for a while.
>He loves being able to just dump in a bucket of seemingly unrelated data, and getting actionable BI from it.
Sorry, just don't trust accurate LLM data analysis even from the latest claude model. Have had some embarrassing bloopers over the last year. It is just better to code a deterministic tool to do the data analysis.
Even for the code. It will sometimes shit a very complex code for something with a very easy solution, and if you ask why it is so complex, it will invent some reason why. Always on 3 points, the first is true but have nothing to do with the subject, the second is plainly false, and the third is weird bullshit.
Don't get me wrong, I still use it, but the more you do, the more I see the obvious limitations.
I don't doubt it. Earlier today I got an estimate for the revenue, profit margin, and headcount for a potential supplier of a hardware component to a product I'm working on. It would've taken a couple hours of really tedious work for me to do that with Google translate, a search engine, and various market research sites. The result is I know more what I can expect from that vendor in terms of customization and support.
But those are the kinds of estimates that I can eyeball and be more than 80% certain they are plausible. That won't cut it in a lot of domains.
Yeah but theres also no null test here. He might have done anything and still been successful. He might have done something else more optimal and done even better. We can't really tell from a single anecdote.
It's really quite interesting how many HNers hate AI; which is odd, considering the zeitgeist of this community. It may even be more intense than the Apple-hatred.
Hate marketing BS and wealth extraction? Absolutely.
I see OpenAI, Anthropic, etc (the whole LLM-hype/boom) leaning far more towards being the latter than the former, in which context the dislike makes perfect sense.
I don’t hate it. I just hate the cult-like enthusiasm for it and the associated false claims. Again, your isolated anecdote does not trump the carefully compiled data presented in this post.
We get way too hung up on data; especially when it reinforces our personal bias.
Some time ago, I had a discussion here, where I mentioned that the average commute time from the suburban community in which I live, to downtown Manhattan, was generally 90 minutes to two and a half hours.
This was based on purely anecdotal data. That data includes the daily news on the radio and TV, which considers commutes from suburbs, important enough to dedicate significant airtime.
It was attacked, because some organization published some statistics, that said the average suburban commute time was 27 minutes. I was told that I was exaggerating, and being overly negative.
Here's a math word problem:
Harry lives 25 miles, as the crow flies, from his destination. However, the roads make it 40 miles.
Of those roads, 10 miles are suburban side roads that Harry needs to take, to get to the highway. The average speed is 25 MpH.
Once Harry hits the highway, though, the real fun begins. It's stop-and-go traffic (20MpH) for 28 miles. Once he gets to the city, he needs to navigate another 2 miles, through the most God-awful traffic on Earth, to where he needs to drive around for up to a half hour, looking for a parking space.
How does Harry make the drive in 27 minutes?
A fairly significant number of folks just took that ridiculous statistic, and were quite comfortable, calling me a liar, based on it. It was kind of jaw-dropping; considering the types of folks that we have, here.
That sounds a bit more "cultish" than we might be comfortable acknowledging.
> We get way too hung up on data; especially when it reinforces our personal bias.
While I don't disagree that we generally get too hung up on data, "It's a problem when the data matches the hypothesis" is an odd blanket interpretation of the scientific method.
I'm confused, your rebuttal to statistics is another anecdote? Was the quoted statistics limited to car commute or includes public transport? Your commute may be long, but it may not represent the average as well. Maybe it's a sign to switch away from cars instead?
Anecdotes are poor because it is highly context dependent. You had to make a long description of Harry to make your point, and yet it has not, to my mind, disproved the statistics, though I would like the source to the stats myself if you could provide it, thanks.
The commute only takes 27 minutes if we ignore that everybody else has an LLM- erh, car I mean.
Think you’re clever because your LLM contacts 100 businesses? Imagine 100000 businesses doing the same thing. You’ve just eradicated human support, because nobody’s going to sit through this BS bot future. Good luck when you get auto-flagged for some reason and there’s not a single human to revert or care.
Too true. Even on the software side i see people blindly “claude make it faster with no mistakes” and gushing about it when the output is genuinely not good not because claude isnt good but because the user of it isnt.
We got a new a CNC machine and everyone is carving away marveling at how easily it chips away at metal when the result is a jagged mess.
No, my point is LLMs are machines designed to create responses to prompts that pass as plausibly human and intelligent. They are working as designed here. The fact that they get stuff right on occasion doesn't change that what it is at it's core is a randomness generator.
Kidding of course. It could be that he's an expert in his domain and can quickly ascertain whether the results are quality. It may also save a lot of time and be the best modern heuristic process.
I certainly will never need to write code anymore, and I can judge the output of these agentic processes. Same thing.
If companies can better understand their customers so they can give me products that better fit my needs, or inform me of products I couldn't have known about otherwise, that's a win for everyone.
I hope he validates the output, the number of times colleagues have done exactly that and gotten the wrong output is.. well, not an insignificant amount.
Not saying it can't be done, but my recommendation is still "ask your LLM to make a tool that process the data, do not process it via the LLM itself".
I’ve always had doubts about the validity of this “actionable BI,” even before it was slop, serving, mostly to serve leadership’s biases and justify their preferred ideas.
If accuracy was never that valuable to begin with then this should not be a problem.
My SO works in legal and regulatory compliance, she needed a small app (worflow with a form and a validation process), CTO said they could not do it, too complicated, would cost 15 days for a freelance and rejected it.
She called it BS, and the next day she spent the morning with ChatGPT using their internal Microsoft-based tooling, and the app was done.
She was hyped for the day, with a feeling of accomplishment and is already thinking of another project.
And I'm hearing similar stories every week.
So yes, LLM have positive impacts in many business domains, but seeing people happy and enpowered with their home-made little apps, I think it's the best.
For most businesses, it’s going to be business as usual. By the time businesses see the benefits of AI, their competitors that actually leveraged AI will have pulled too far ahead.
If the original company has distribution, they can likely catch up, but the gap we see between AI-first companies and the companies not utilizing AI will broaden exponentially in the coming years.
And I'm sure the business logic has been validated and tested, the hosting and execution medium has been secured, and hooking up to a curated and vetted internal package manager mirror so as not to get hit with 0-days during a shai hulud-esque attack, and of course the executable is being scanned for vulnerabilities, and as the app needs to evolve in the future it will be able to do so with no problems whatsoever.
I’ve had two separate similar experiences. The most impressive was my little sister publishing an app on both Google and Apple AppStore’s with subscription payment support. She didn’t even know what an Apple developer account was. I gave her a three line prompt to get started and then she took it from there. She built an LMS for HS students that gets them in shape for college admissions. She has a supabase backend, AI generated videos, quizzes, more content, and all the usual mobile app bells and whistles, she made it through all the paperwork of App Store submission/review/publishing on both platforms including inviting some beta testers. Her workflow was telling ChatGPT what she wanted and then asking that model to talk to claudecode. It doesn’t sound like much but keep in mind she has absolutely zero code or tech experience. I told her she did about a $75k project on her own in maybe 1/5 the time and it blew her away.
It may be a success or maybe not who knows but she did it. I think she has 4 or 5 paying customers.
hey - supabase ceo/cofounder here. Please tell your little sister that I’m happy supabase could be a small part of her builder journey, and good luck with her app. My email is in my bio if she has any product feedback
Yes maybe the long term impact will be there won’t be a technology sector (at least not so dominant as it is now). Instead companies in other industries will have larger technology functions with teams of agents.
If her company has any brains, they'll tell her to stick to her job and not waste time on untested software developed by somebody who has no clue what they are doing.
I really love these stories - as a contractor who fixes the badly implemented shadow IT crap
aaaand let me guess? have they assessed whether this tool needs backups? monitoring? resilience? patching? obviously in 5 years time the author would have left leaving NO documentation on a mission critical 'small' app - likely hosted with a shit ton of other mission critical crap
It's MS Access all over again, my future rates are looking good
> have they assessed whether this tool needs backups? monitoring? resilience? patching? obviously in 5 years time the author would have left leaving NO documentation on a mission critical 'small' app - likely hosted with a shit ton of other mission critical crap
Or, hear me out, it's not actually "mission critical" and none of this stuff matters. They made their job easier in a small way, good for them.
> It's MS Access all over again
In that it's a convenient tool that a lot of people find useful?
Yup. So many bozos giving those of us that know our asses from our heads job security far into the future. Yeah it's gonna suck to actually do the work a lot of the time, but the irresponsible generating insane amounts of tech debt that people that know what to do is a tale as old as the outsourcing craze and codeless tools, and frankly, probably COBOL.
We've heard of "Shadow IT" now we have "Shadow SWE" and "Shadow CI/CD".
It used to be that someone wanted unfettered Internet access, and they were disgruntled with the company's firewall or proxy policy, and so they brought in a modem, and hooked it up to their desktop PC, and started dialing out on some analog line that the company didn't even know they had. Then it's all fine for 3 years until some malware gets discovered, and then IT discovers who-all had unauthorized lines that weren't being scanned or filtered at the company firewall. Same thing happens with WiFi APs today, of course.
I was working for a company that really offered no in-house software or tools to run. But my new supervisor was asking me to closely track my time. So I signed up for some SaaS that would let me log activities and track my timeslices, so I could show a graph or table later on and see what I was doing. I used it for a few weeks. It was real effective. But what I realized down the line, was that all the notes I was typing into this 3rd-party SaaS, were revealing company confidential information. And ultimately, I felt very uncomfortable about revealing what I was doing, on an unannounced and private project, and without authorization or even notifying my employer's IT department. So I shut it all down, and I had to do it with pen and paper, but that felt secure enough.
So if people are going to feel empowered by vibe coding their own bespoke apps, I hope their IT department comes down like a ton of bricks, and tells them to knock it off. Because it is not worth the tech debt, support costs, and ultimately the massive security problems and infoleaks. If your employer won't give you what you need in these terms, then maybe it's just not important to them in the first place.
I see things this way too. Another point is that the FOSS software movement was extremely unique and I don't think there's anything analogous in any industry. This provided the backbone by which LLMs became so good at manipulating code. Without FOSS and platforms like GitHub, LLMs would be severely lacking in terms of writing code.
LLMs need deterministic verification to enable a quick and short feedback loop before they become really powerful. Software is uniquely positioned to provide this feedback loop. Much less in other industries.
Isn't the point that there aren't many other fields that check these checkboxes? Software is very unique in that it is so easily verifiable and there is a glut of training data.
No other fields check all of them, especially the last one.
The question is whether this uniqueness is going to be persistent. I think the uniqueness just makes software first. The alternate theory is that software will stay unique and that other fields won't get to the same level.
Third possibility: software continues "eating the world" because of those properties and because it's so amenable to AI. So other fields just get gradually subsumed into computing, sort of like how Lean is turning pure mathematics into software rather than humanities-style written arguments.
This seems correct and in fields of knowledge people guard their secre, or it is not easy to extract/document knowledge, or it isn't publicly available.
We’ll need artificial intelligence for robots and once they start rolling out at a larger scale, that’s when things will really take off. It’s pretty well established already that robots can do a lot of the physical labor that humans do. I’m guessing it’ll take another five years before we start seeing robots everywhere but they’ll steadily increase.
> Software creation tools have been worked on for decades to keep fallible human coders out of trouble as much as possible.
And the LLM-based coding agents provide the ultimate scapegoat for those who consider "the LLM wrote this code" to be an acceptable, deflecting response when issues occur.
Pretty every industry these days has a software component, but software is a major revenue driver specifically in the tech sector, so of course this is where a lot of LLM usage is showing up. Tech is also a sector where "build fast and break stuff" is a common ideology, and LLMs play right into that. There is also a long warm up period to writing production code via LLMs in industries that still run a lot of their infrastructure on old mainframe code(it is understandable), banks for example, or public sector, or energy. LLMs at the current grade are still a fairly new phenomenon, it will take time.
I onboarded a bunch of nontechnical people in the last few weeks and they output skyrocketed - they deliver faster and better than before.
But I’d say only recently CC became decently usable for nontech people - the auto mode plus finally polished browser integration, plus Opus 5.5. Even I was surprised by how well it all works now out of the box (and how cheap - regular Pro license is decently useful now).
No. One is doing PR/managing a project, another is doing marketing stuff.
Also I’ve seen and heard of previously nontechnical people that now build oneoff tools for themselves to speed up their work or coordinate the work better. My fav example is a girl - arts major - who was in customer service, and rebuilt a CRM for a 10-person company. As a cherry on top she created the nicest video tutorial of any software I’ve ever seen :)
But in other cases output was nonprogramming in nature. Since CC can access mail, slides and excel and a bunch of other stuff there is a ton that people can ask it to do and not waste time.
I’m curious why CC and not Claude app, for example which has all the connectors and doesn’t require a terminal to interact with whereas CC is designed as harness for coding tasks (that’s what it excels at).
I’m curious what does “skyrocketing” output mean for you objectively? How does you measure that?
> arts major - who was in customer service, and rebuilt a CR
If you ask the lady, where are the customer data? How can I have the same CRM on my computer? How can we let the boss follow up on leads (typical CRM stuff) what would she say? Would she need to go back to CC and ask? How does this work in practice?
You're basically agreeing with the person you replied to. Software went quickly because all of this tooling already existed. It will take a while to build equivalent tooling for other knowledge work, and once it does you'll start seeing larger benefits.
> Cool analogy, but there's another possibility: software is uniquely suited for being boosted by LLM-based coding agents.
Or the pretty obvious third possibility: The tech sector is the part selling the LLMs. They're not digging up more gold with their cool shovel technology, they are selling the shovels.
I agree - it is easier to prove that a software implementation is correct. In many other industries the definition of what’s good work is less tangible. There are also jobs where the liability for mistakes is extremely high.
This is from the view of a software person or must be, because formal definitions of math, chemistry, physics, engineering, the English language, etc. have existed for centuries. There are entire college curricula, formal dictionaries, exact specifications for equations and how to apply them to basically all fields of engineering.
Like... other fields exist, this is incredibly myopic. What is specific to software is it is all language based, and the people making the tools happen to be familiar with it so they know how to train it, but that's about it.
No, most white collar work is ultimately keeping groups of humans in alignment, working on the same thing without stepping all over each other, and working on something that generates value and isn't a total waste of time/money.
The PPTs and emails are just means to an end. But at least you're able to sound all cool and edgy on the internet.
and you are able to sound like someone passive agressive who tried to undermine others takes while sharing takes from a highscool teacher trying to convince students that their future will be bright. even ur answer could have been automated years ago.
many jobs do not generate any value at all. your vision of the corporate life looks like one of someone who has never seen what some of the people do in corporations. most of people just spend days producing 0 value or even negative one. in public sector that "most" is likely 80%. why do you need a human to take in your passport renewal demand? why do you need a HR worker to give you the details of a policy? not even saying that most of the people wont even provide you a specific answer to your own case and struggle intellectually to tie it to your specific case and most of the time answer something wrong. happened to me plenty of times in immigration offices in the US or Europe for instance.
in consultancy like the big three a big chunk of the work is useless and about delegating accountability to those big three. they will continue doing that but soon they wont need much humans to produce those reports. AI are already better and it and most of workers there just prompt without even checking sources.
those jobs will be gone. for some of those they were already sent to call centers in india now it will be sent to an LLM.
Seriously? Lawyers. Doctors. Civil engineering. Accountants. And that's just the fields I know of directly being significantly impacted that I could think of in 10 seconds. Oh, also: education.
A difference is the output of software development can be immediately applied in an automatic way - production leads to delivery as a matter of fact. The whole flow can actually be automated.
Producing legal documents on the other hand is only part of the process, a lawyer or similar must still apply the produced text in some human driven context.
And whole repositories with commit history that provides step by step training on how to code.
AI could probably write decently if every scraped book had all the revisions to create it embedded within or art had every brush stroke documented. As it stands programming is the first to be trained. The rest will take some more time.
I can't speak for everywhere else, but here in the Bay Area (California), pretty much all hospitals use AI for transcription, summarizing, and some other stuff that I'm not fully aware of (I should really listen to what my Nurses say before agreeing to being recorded).
My doctors office has started doing that and it throws the wildest crap into my chart. Diagnosis the doctor didn't mention, personal anecdotes that doc told me about their family become my family medical history. Test supposedly done that weren't.
Ive been trying to figure out some persistent si joint pain and my chart is full of me passing or failed tests nobody performed, observations nobody made and diagnosis that no doctor confirmed.
Its just so much wrongness, ive tried getting it corrected a few times when it really matters but its just so much wrong crap
Since it's brand new there are going to be a lot of subpar solutions on the market. When SSDs first came out, there were only a few models that were actually reliable.
It is relatively easy to create and deploy a transcription model. It's pretty hard to find one that is really accurate and to configure it with a good microphone setup and pair that with a fast but also intelligent LLM to clean things up. If you execute any of that poorly or pick a cheap not very smart model for privacy and or cost saving you could easily end up with nonsense.
Sure but this doesn't answer whether medical systems will see a lift from using AI.
(They will. It's very well suited to it, once the garbage gets filtered out. It's not like it's a scam as a whole, it's just a field that's very scammy rn.)
Gemini making transcriptions of meetings and summarizing make so many mistakes. It will totally misinterpret what was said. I don’t think the tech is ready for something as important as medical transcription.
i mean, i am using 3.8 live to take notes in dialog mode, allowing it to asks questions.. its working fantasticly, switching between german/english with no effort and i have a rather hard accent
Medical transcription isnt speech to text. The doctors I know at a major midwest provider had indian doctors live reading the speech to text and then filling out the chart which the doctor then signed off on. Now theyre using AI instead
Sure, but are there any examples of hospitals where lowered costs for transcriptions are significant enough to affect their financial reports? How much of an expense was it in the first place?
> As an example, take a factory that was built around hyrdo-power like a grain mill. You can't just take an electric motor and drop it in in place of the water wheel.
Your chosen example is rather more fitting for the opposite case. Dropping an electric motor is pretty easy, you just attach it to the other side of the axle. But there is really no point in doing that. The sole benefit for that mill, will be that there is the cheap power from the water. It's also the sole reason for the current position of the mill, for the electric motor, you hence need to add long cables first. The size and operation mode of the mill is also not fit at all for the electric age. There is really no point, you are better of just building a new mill in a completely different location.
But you can do that independently of the old mill, since there is hardly any capital to be taken from the old to the new mill, short of maybe the office supply. So you just build a new mill, completely independently. But there is also no reason to decommission the old at the same time, the only reason is to shut it down when is no longer economically viable. But that also might be at a different point in time than building the new mill.
> You can see a similar history with electrification.
It took a massive amount of new infrastructure for electricity to benefit people. Lines had to be run, houses and businesses had to be wired, new light fixtures and appliances had to be purchased and installed.
LLMs can currently be accessed and used everywhere instantly by everybody right now and at any moment day or night. Rich and poor alike can take advantage of it. There's nothing holding back anybody. If everybody without electricity could have pulled a small rectangle out of their pocket and pressed a button to instantly have electric lights and power, the positive impact of electrification would have been immediate. This is not even close to a "same thing now" situation.
Every computer and cellphone in the office is an "AI powered appliance". Every application can be an AI powered one. LLMs have been shoehorned into damn near every app they use already. There's no magic appliance or software that will make productivity and profits suddenly happen for businesses that they couldn't already be doing right now, and I say that fully realizing that there are people here who desperately want to sell businesses something (possibly vibecoded) that they'll claim is a magic AI appliance/software that will make productivity and profits suddenly happen.
> Every computer and cellphone in the office is an "AI powered appliance". Every application can be an AI powered one.
True, but funnily enough those are not workhorses of most businesses outside of AI. Jevons paradox was about that. Introduction of computers into most businesses didn't traslate to immediate performance gains so it's hard to imagine that automating computers will have strong immediate effect.
I think the most powerful way that software is going to benefit businesses outside of IT is creation and maintenance of custom software for every business that's going to be better fit, faster and cheaper than anything IT industry can deliver. Though even those gains are still going to be modest to moderate.
At some point real AI powered appliances might show up, like robots or full stack AI management.
The issue is that software is almost unique it the sense that you can use reinforcement learning to get the desired outcome - does it compile or not? A large amount of automated trial-and-error runs do that.
It's not true elsewhere. The definition of "is this a good outcome" is pretty nebulous in other areas.
"Is this what the law says?"
Yes, but that's not how the judge will interpret it, etc.
But obviously getting code to compile is not a reliable test of whether the software is good. Software isn't as objective as people are claiming on this thread. It's sort of like saying that if an LLM makes a legal document that passes a spelling checker it's done.
If it's [or remains] allowed in. The judicial branch is well able to just outlaw it on grounds of safety, trust and reliability. Same for financial auditing, public notaries, ombudspeople, and the likes (i.e. some degree of responsibility to the public, or dual responsibility to the public and a customer/client.)
No, liability for duty of care, not for outcomes. If a lawyer gives you legal advice, you can absolutely go after them if that advice was incorrect or made up. They carry a lot of insurance for a reason.
I'm not an cliometrician, but my impression is that it took about 4 decades for electricity and steam to do much for productivity.
And it was in 1987 that Nobel laureate economist Robert Solow famously said “You can see the computer age everywhere but in the productivity statistics”
I mean, maybe AI will be faster, because it's especially easy to apply compared to generally applicable technologies. There was the Bank of Korea study showing this spring that did show productivity gains with single proprietors--folks who don't have to reorg employees and processes to do something with more abilty or saved time.
But, generally, I don't know that there was an reason to expect that corporate America generally was going to see an AI payoff so soon.
You still, in 2026, can't really see the computer age in productivity. We all have phones, which are making us less productive. We can order physical slop, which has no value, online with same-day shipping. We can communicate about serious things instantly, which has positive value, but doesn't outweigh the negative value of all the constant distraction.
It will be new factories designed from the ground up around ai and its benefits and limitations that really move the needle.
Yes, totally agree. But we need more. As I wrote in an essay responding to Dario’s pacing the frontier, in my view, we need a stable long term supported model that companies can build around. In the same way that Ubuntu for example has a long term supported version. This gives companies time to build and adapt. The latest models can still be the realm of the frontier but businesses, schools, countries etc need long term guarantees.
It's very hard to improve something that you don't understand yourself. All the tiny decisions you make as you create a product are informed by the experience you have.
So it's seems obvious to me that a product created by software engineers would be better at software engineering. A programmer knows what a programmer needs.
To improve another industry successfully, you would really need people from that industry to be involved in a deep level in the design of the product.
Well for AI it's more like training than product design. I have switched from software engineering to AI training for software engineering (partly because honestly it is just much easier now to get that kind of gig work versus the programming contracts I was getting before).
They have a ton of tasks that pop up on my dashboard for all sorts of industries that I know nothing about. I think they are ramping up in non-software areas.
I think you are missing the other side of the coin where it is actually government regulations holding adoption back in many sectors. Ie you are only beginning to see benefit in driving now because regulation has been forced enough to change to let the pilot projects happen. The millions of drivers who should have been fired by now almost all still have jobs and will ibto the medium term
But do you really see a change in the technology sector? I don’t. If you plot operating profit per employee for Apple and Alphabet, it’ll show a trend up but since COVID. If you do the same for MS, it’ll show a steady trend since Nadella became the CEO. Amazon is flat, so is Tesla. Can’t really see a change caused by AI.
It doesn't work in software either, in the exact same way. It's just that everyone and their dog wants to "take programmers down a notch" so they are exceedingly happy to accept the resulting mediocrity.
Joke's on them, all intellectual work is modelling to some degree, and can be automated in the exact same way with the exact same mediocre results. How long until someone vibe codes a Lean for structural engineering?
That's good in theory, but in practice the subject is so complex and filled with judgement calls that you'd end up with something resembling a bar association.
After 8 or so years since GDPR went into effect, its main effects thus far have been annoying cookie banners and assisting corruption (our HoA have cited it as an excuse not to publish various names and contracts relating to maintenance of common property and politicians frequently remember that they have a right to be forgotten whenever there's a scandal nearby).
All the other technically illegal shit happened anyway. I have no reason to believe that the regulators will side with the people and every reason to believe that rules would be something along the lines of "it's too dangerous so you must use a LLM".
We need something. We need some way to force people to act responsibly. The amount of yeeting hundreds of thousands of lines of code without proper review or testing should have consequences.
We needed something 20 years ago when Google decided to take over the browser to ensure nothing is stopping them from tracking you wherever they want.
We needed something 10 years ago when Cambridge Analytica showed that the consequences of the above are not just limited to the internet.
Google now writes the web standards and TikTok is telling people how to vote. Between the incentives and the fact that politicians don't understand tech, the only thing they'll end up forcing is "hundreds of thousands of untested, unreviewed code for authorized professionals"
What do you do with open source, with the arguments calling software speech?
Regulations would act as "unit tests" for a machine-produced design in the same way Lean validates that the theorems prove... something.
Of course, they were written to prevent human errors from otherwise well meaning but possibly clumsy actors. Nobody has a fucking clue what the error envelope of these models actually is, they just use magic words like "apply guard rails" as if there wasn't a whole area of science surrounding error propagation that they couldn't apply because the models themselves are vibe-data-scienced together.
Academia has similar rules to prevent clout farmers and incentivize discovery, though in that case the followers of the machine god have already decided that the human scum must refrain from thinking and instead spend all of their time deciphering the gospel.
Or maybe it was all propaganda, possibly generated by some weird text generator that was recently invented ¯\_(ツ)_/¯
So why are software "engineers" stamping all of this slop? Could it be because they are forced to do so in order to "leverage ai"? Or maybe it's just the shitty and mediocre engineers doing so while the rest are labelled luddites?
Specifically what exactly is preventing the same pattern of fire competence and force mediocrity to sign off from working in any other area? The risk of catastrophic failure? There's insurance for that, another domain that's a harness around a model (albeit not a large language one, yet).
"There's insurance for that, another domain that's a harness around a model"
Yes, but the PE's at engineering firms have the ability to say no and demand redesigns since it's their name on it, unlike programmers.
I frequently see firms design way beyond the bare minimum so that there's no question and it's an easy sign off rather than spending months on a back and forth.
Head architects at software firms had the ability to say no to generated code and demand redesigns. They were pushed out for being "backwards thinking".
As to your second point, that just means that the training corpus is already designed with a healthy margin of error.
You don't think they can write or create images/music/video? If you are being seriously, you should try some of them out - they are incredibly capable now.
I think their question is whether it's replaced work in industries focused on that - AI movies becoming popular and very profitable, AI written books selling large numbers etc
Very possibly on the way but so less uptake than software engineering for sure.
It's on the verge of that. If I open X then quite a few posts in the feed will be AI generated music videos right now, and for some reason people doing gender swap fanfics of Draco[nia] Malfoy/Harry Potter romances. People are finding them entertaining.
Maybe you could give an example of a Fortune 500 business that is too capital constrained to take advantage of AI. It seems far more likely to me that the issue is that of the skill sets important to these businesses, these agents are only really good at coding, and they still are very jagged even in that.
It is most impactful in tech because tech is almost nothing but languages and more so super expensive languages where one hour of an “authors” time can cost thousands with a supply demand issue so bad workers are imported from across the world.
Artists were already starving. The more physical trades are…physical not language so a plumber still plums. Maybe now they ask gemini instead of google tho.
Areas related most to…language are the most impact by large language models? The one that i think is just as impacted is education because it is also almost all language driven.
I see zero sign of any of the new mills coming around. Steel, crops, farming, none of it has changed and what little has is from…the software side or…language.
The SV world really over indexes on how much software can do.
Look all of the progress in robotics. Companies like Physical Intelligence etc. There has been massive progress in all sorts of areas from picking strawberries to assembling kits, folding clothes, etc.
Huge progress in spatial understanding with frontier models now building architecture and complex mechanical systems in Blender.
If you think plumbing is anywhere near being automated away you must not know much about plumbing. You will never see a bigger hodpodge of materials and fittings in so many shapes and sizes and states of decay.
Maybe on new builds with fresh material, but that is already the easiest and quickest plumbing task.
An entry level largely untrained junior doing pro level work because their AR headset and earpiece tells them exactly what to do next? That's possible today.
But nobody wants or pays for "pro level" plumbing work. Just like the majority of construction, speed and minimum viability are king, quality is the last consideration. Especially for plumbing where it all gets covered up. If something leaks in 20 years instead of 40, plumbers don't lose anything. But the plumber who put in extra effort and time does.
Many skilled tradesmen have already left the industry because higher skill levels have little value, and their labor value is just barely above grabbing randos off the street. AI at best is saving a minute of evaluation or google search, but 99.5% of the job is physical work and pain still, not high level problem solving.
Hahahaha absolutely not. There's a lot of carefully staged demos that are good for separating the gullible from their money, but we are nowhere near anything you're talking about.
GPT-2 was nowhere near being able to create web pages. It was largely incoherent. But three years later gpt-3.5 and similar were able to create simple web pages and programs and became very popular. And a few years after that, only a small portion of software engineers generally write code by hand still.
Humanoid robotics may progress faster than that because they are building on existing AI. But it will be capable of some plumbing work within a few years.
The robotics structure, dexterity and sensors are already 90-95% there in leading edge humanoids. What makes it able to be a plumber or not are the machine learning models, which are usually built with transformers (VLAs -- vision language actions models), often include frontier VLMs or models with the same architecture as a top level control, and use the same or very similar types of reinforcement learning for training.
>but then it took like a decade of little growth before the true benefits showed up
That's the core issue apollo is talking about. Investors have been told, "Buckle up, we're getting rich in 2027." Not "Hey, after 20 years, this will probably be pretty nice, maybe." This is how bubbles pop. Everyone starts to realize they were oversold and they rush to get their money out at the top. LLMs won't go away, but the financial collapse will be devastating.
this isn't the invention of the wheel, electricity, etc. You are definitely overhyping it. LLMs hallucinate. It is an unsolvable problem defined by the very nature of the architecture itself. No amount of thinking or counter thinking or triple checking will fix it because LLMs don't understand the basis of what is right and wrong, what looks right and wrong, why saying or doing the wrong thing is bad, and why or when it should realize what it produced is wrong, and what it should do about it.
Big Tech wants to make humans the go-between, however, the thinking and problem solving wasn't the problem to begin with. Checking for mistakes, correcting them, and presenting the corrected versions is the hard part. If a human does not go through the steps of solving the problem, they can't effectively analyze or process the output, which will always lead to mistakes. The only way for LLMs to take over any jobs would be if humans were perfectionists and could check every single bit of work without fault (since LLMs will always make some mistakes: a "perfect" LLM isn't possible anymore than a perfect human being is). Humans (especially myself are disaster and mistake prone.
That's why you should not invest now unless you are short term trading, and you should get that bag of popcorn ready. Trillion dollar valuations aren't real. It's a fake number propped up by a bubble. The US is in record debt, and it will be unable to backstop the financial loss...without taxing heavily the very folks that want to be backstopped. Those folks are dimwits and are only focused on making a buck. They are assuming everything will work out, either via investors/revenue/profits or a government bailout.
Bear Stearns had the same attitude.
In 2008 the government had leeway. This time, they do not. Regardless of political party.
Disclaimer: Please note I do not have an investment in any company, not even in a 401k. I am a disabled software engineer (and manager, in some cases) who worked in/with startups with various sectors. I have knowledge about all the financial and investor nonsense that goes into startups and small businesses. As an autistic millennial, who has definitely been around the block with bubbles, I am also interested in tech and finance. I also have a decent amount of knowledge about the damn stock market. I don't and will never claim to be an expert in any of those things, however, I know enough to say that a whole bunch of folks are going to lose a TON of money at some point within the next few months/years (likely before 2030, possibly by 2028), so if you are overexposed to the stock market (pretend it drops to zero, won't happen, however, it will drop, and no stock is safe, this goes for you FIRE folks as well), you should back off your investments to an acceptable level (the most amount of money you'd be fine with losing ALL of).
There are some things that current generation LLMs are extremely good at. They also suck for a lot of reasons and appear to be prohibitively expensive to train as "general" models.
hey are amazing tools, but every time I see a task that LLMs helped me complete in 30% less time, I see 5 more tasks that took me 30-50% longer to complete because of errors, lies, mistakes, hallucinations, context limits.
Well, of course not. If you borrow a trillion dollars to build a 1¢ aspirin machine to compete with $1 generic aspirin, you are strongly incentivized to sell your aspirin for 99¢. We don’t see productivity gains, because approximately all of the gain is captured as revenue for AI companies.
It’s weird that this has to be spelled out for so many engineers and economists, but to be very clear: OpenAI et al wants your token spend to go to one-minus-epsilon of what you previously spent on payroll, and wants your productivity to go to one-plus-epsilon of what it previously was. If you want 10x productivity, spend 9.9x more on tokens.
They are in the business of enclosure and rent extraction, and any innovation that comes out of their research and development - regardless of how real an innovation it is - will be completely contained and bent totally towards that business. Any leakage of the innovation that you experience as a productivity boost is literally felt as lost profits by them.
Your argument falls apart pretty quickly when you consider that (1) open-weights model are a thing and (2) OpenAI and Anthropic are still losing money
1. If AI would bring great productivity but all that extra value is captured by closed-source model providers, then everyone would switch to open-weights models + bare metal servers, which would non-tech companies to capture most of the value. Your argument that there is great value in the use of AI, but it's all captured by OpenAI and Antrhopic falls flat on it's face here. The market would adapt quickly.
2. OpenAI and Anthropic are still losing money. Your argument is that 100$ of extra value is produced, but the closed-source model providers capture 99.9$ of that because they are "rent extractors". If you're right, and they capture 99.9 of the value created, that's a huge issue because they spend 150$ to create that value. If even after eating the whole (value creation) pie they're still unprofitable, that's the worst news for the AI industry.
I basically agree that both of your points are major-to-catastrophic problems for OpenAI and Anthropic. I suppose it is true that if the companies fall apart, then my argument also falls apart. I think I would be pretty pleased in that scenario, all things considered.
Oh. Well, I am rooting for bad things to befall them. These companies are creating enormous negative externalities for society while pursuing profit and hewing to an ideology that believes it is genuinely risking destroying the world, and believes it is justified in subjecting everybody else to this risk because if they don’t, someone else might.
While in general I commend showing generosity and charity to those we disagree with, I don’t think it’s commendable to remain generous past the point where it becomes untenable.
OpenAi and anthropic hypothetically undercut each other.. each epsilon-ups one another until they get to a point where they're splitting the productivity gains with the other companies. If this doesn't happen, it's called collusion and the SEC initiates action. (Ha!)
If we discuss theoretical scenarios, it's less likely to happen in a scenario where we only discuss about 2 companies in an oligopoly, so the argument isn't as strong.
In a scenario where we have 2 companies AND open-weight models (Chinese, American, French), all of a sudden that process is accelerated. That's basically my first argument in the comment.
Fair enough, there's definitely a shortage, but there is capacity available for people that are willing to buy it. There are platforms that give easy access to GPUs for small players, while the large players can make direct contract with neoclouds or datacenter providers.
The companies discussed in the article are 493 of the biggest companies in US, so if there was value to be extracted from using bare metal + open-weight models, I am sure they would be all over it.
Everyone talks about how sophisticated those large players have become in trimming their AI spend, they're a lot more cost-conscious. Perhaps this narrative might be overblown or there is truth to it. We'll see.
Not sure how productivity and rent capture are connected here.
1) If a firm spends $99k on compute to replace a $100k engineer and produces 5x the output, how is that not a 5x increase in aggregate labor productivity?
2) In your aspirin example, why wouldn't a competitor with an identical 1¢ machine sell it for 50¢ or 10¢ to take 100% of your business?
You are either making the assumption that 1 company is owning the market and no models are useable beyond that one, or that all of the models that are in the front of the market have a cartel and fix the price, which is not true either looking at API costs.
1) If a firm spends $99k on compute to replace a $100k engineer and produces 5x the output, how is that not a 5x increase in aggregate labor productivity?
The firm is spending $99k on tokens to replace $100k in labor. The token provider is spending $ x on compute to give $ p•x productivity output to the firm. The token provider can choose x at inference time and adjust p at training time, so they control p•x . They do have to make it greater than the firm’s current output so the firm will buy from them, but otherwise their profit incentive wants to make x as small as possible - if 110% of previous output suffices, it is strictly a loss to spend more to reach 500%.
(Also, pedantically, compute is not labor - if you 5x output and 0x labor, the labor productivity statistic becomes “what’s the ratio of 5/0 to 1/1?” ie undefined. But that’s not a very interesting point.)
Note closely the “middleman” vibe of this arrangement: the firm does not spend on compute to get output; the firm spends on tokens and then the token-provider spends on compute to get output, and then hands the output back to the firm. Rather than put in x and get back f(x), they put in x and get back e(f(g(x))), where e and g are the middleman’s extractive greed functions.
2) …why wouldn’t a competitor with an identical 1¢ machine sell it for 50¢ or 10¢ to take 100% of your business?
If they were in the business of free market value creation, they probably would! I don’t know exactly how economists would describe it, but empirically, it seems the business of rent extraction has a different dynamic - one which can roughly be summarized as “competition shrinks the size of the pie, rather than growing it”, so all participants are mutually incentivized to treat competition as a cost to be avoided.
Can you demonstrate your take on 2) happening in practice?
- You have tokens priced differently as evident by for example pricing on for example OpenRouter.
- You don't reach 100$ of Fable/Mythos output by running Deepseek for the same amount of money. How do free open source models fit in here? Do you have to run Nemotron for an infinite amount of steps to reach 1$ of Deepseek compute? If not, then the argument is not valid because the pie should be shrinking.
This also ignores that 5x productivity doesn't often give you 5x profit, even if it gave you 1.1x profit that sounds great - until people start looking at profit vs productivity.
...we should be seeing prices of every day goods and services(mainly) come down.
The incentives are there. I know reddit and HN it seems like lots of people are using models/harnesses from scrappy competitors, but IME at the enterprise level I’m in the extreme minority who is even aware of them. Virtually everybody uses Anthropic and OpenAI tools and models, with Copilot getting some scraps because of its forced inclusion in Microsoft/github ecosystem
There is no question about the benefit when things go right; a wary engineer responsibly guiding agents to build orders of magnitude more than was possible before, and always skeptical and critical of the vendor, knowing the vendors’ incentives.
But that’s not the majority, and the people approving absurd token spend are too far out of the weeds to see what’s happening - people burning 6 figures in tokens and still end up with broken systems that keep them over the weekend, but now they hardly understand what they’re troubleshooting (if at all).
This is a dream scenario for the fat-cat vendors, so there is incentive to have CC/Codex write more code, overly-complex designs. When users get upset, they jump to OpenAI to spite Anthropic, or vice versa, but always getting further locked in to the duopoly
I don't really see the bloated-code take here. With this reasoning, there should appear a non-bloated open-source tool that all knowledgeable, smart developers will migrate to in order to get things done faster, and the market will have to follow.
Also waste and token spending could very well be a short-term thing. A lot of teams had crazy AWS bills in the mid-2010s on lambda functions and workflows, then it got cut down and optimized.
> 1) If a firm spends $99k on compute to replace a $100k engineer and produces 5x the output, how is that not a 5x increase in aggregate labor productivity?
I think they are saying that with 99k pure compute you are able replace 100k engineer with equal output. Or that is the goal of the AI companies. It is worse for them if you get more productivity.
It's so interesting that sometimes people throw these wild numbers around like 10x productivity, 100x greatness, etc. but when the rubber hits the road, we see so much ambiguity, unknowns and unrealistic assumptions.
But the AI companies have no access or knowledge of how much value you produce, they would have to calibrate against some average SWE salary and some average productivity which is task-dependent. There is no way for them to control this.
Even if they did, they would lose out to other companies who do not do this and let you "keep" surplus value, it feels like this again assumes a no-competition market.
I guess it just seems obvious to me that most of the major players want a no-competition market, and are working towards making that happen, and are having notable success with their efforts (the whole “pacing the frontier” was a remarkably public display of coordination between purported competitors, the government has already banned deepseek, etc).
> If a firm spends $99k on compute to replace a $100k engineer and produces 5x the output, how is that not a 5x increase in aggregate labor productivity?
It sounds like they spent 99k to replace 5 100k engineers, which means they could charge 499k for it.
This ignores the supply side of the equation and just focuses on the demand side.
Of course every business wants to capture 99.9% of the value they offer, but competitive forces drive this down to an equilibrium.
Also, as an aside, I've noticed that there's a subset of people on hackernews who use 'weird' to describe any disagreement with their position. It's not 'weird' to have an opposing or different view and removing such language will make conversations on here more pleasant.
This only happens in very mature industries. AI is obviously moving extremely quickly with lots of new entrants and moving parts. That's why all the model providers are losing money fighting for market share.
Once most of this debt is gone, whether through companies going bust or the debt being paid off or inflated away, AI will become cheap enough for frontier models that any productivity improvements will be visible in the user's bottom lines.
Once all of the trillions of dollars of debt is paid off…
…plus whatever healthy premium the corporation decides it is owed for so bravely taking on those trillions of dollars of debt, once that too is paid off…
…plus whatever the expectation of return on investment the lenders were persuaded by, once that has also been met…
…only then will we get to see what proportion of the gains these corporations think it fitting to let trickle down. What do you reckon they’ll let us have? 50/50? One in ten?
I don't think it will all be paid off. I said once it is gone. I think, like the Dotcom bubble, many companies will go bust and their debt will be worthless.
I don't see why a payoff would have appeared yet. Many (but certainly not all) software companies have increased their output without hiring more people. But having software with more features doesn't necessarily drive immediate productivity gains in the rest of the economy. It will take time for those benefits to percolate. We've only had access to models good enough to majorly accelerate code writing since early-2026. I feel like we're in the very beginnings of the first part of this wave.
For sake of comparison, consider e-commerce. As a proxy for "impact on the economy", we can use Amazon's revenue. They went public in 1997 and that year had revenue of $148M (just 0.14% of Walmart's revenue that year). By 2000, revenue was $2.76B (1.65% of Walmart). Bricks and mortar didn't start feeling anything until probably 2011, when Amazon crossed 11% of Walmart's revenue. And it wasn't until 2025 that Amazon finally surpassed Walmart.
But, if you worked in tech in the e-commerce business, things were already feeling pretty hot by 2007. And at that point in time, the e-commerce sector was still miniscule compared to traditional retail (~4%). Software is very exciting these days and the most fantastic things are happening. They will eventually percolate out to the rest of the economy. But it will take many years before everyone notices the difference.
Tbh my day job is now bogged down with security fixes for bugs found by AI. I guess that's a productivity gain by some measure? Sure we're fixing lots of problems. Problems we were blissfully unaware of before (and arguably were not problems, if no human could find them).
Meanwhile, at least half of the engineers have been pulled off other projects to work on, well, anything that has the veneer of AI attached to it. So we've lost half of the workforce to AI. But sure, the 25% of engineer hours still working on core product have gotten 40% more productive! (Actual coding, 2-3x, but that was never the majority of the work.)
What is the long term steady state here? Not sure if there even is one. But I've always spit balled that my job would be 90% less effort if all humans were good faith actors (no need to worry about privacy and security). Just ditching those two considerations is a 10x multiplier. So conversely, if the bad actors are empowered 10x by these tools, instead of 10% of my job being actual functionality, maybe 1% can be. I need the (promised) 10x boost of ai just to get back to the historical base case.
Perhaps the people seeing massive productivity gains are ones who never bothered with security to begin with (ahem Sony).
>We've only had access to models good enough to majorly accelerate code writing since early-2026. I feel like we're in the very beginnings of the first part of this wave.
The people talking about "increased output" — do you have any proofs for these words? I'm working in IT — I don't see any radical productivity growth. Yes, you can doo hello world 4 times faster now — but even then when you need to polish it the speed of development stabilizes near baseline real quick.
> do you have any proofs for these words? I'm working in IT
I definitely see effective use of the tools.
The most common examples are non-tech or semi-tech people solving problems with Claude's help (e.g. create a query that provides X data for the auditors) and only requiring a quick review by the tech team as opposed to the tech team having to do all of the work. The pace has increased and the tech team is less of a bottleneck for some things.
Your arguments are actually confirmed by many researches: people with low competence in some area solve the problem with AI significantly faster than deducing the solution themselves. However, there is also a catch: people relying on LLM too much eventually fall behind in learning.
However, often time people start talking about something like 100% increase in performance — I just don't see it. Don't get me wrong: I totally enjoy AI making dirty work for me, but it's not something magical.
You are right — the paper describes 2023-2024 experiments. AFAIK, the most recent available is early 2025 METR experiment. There is also “How AI Impacts Skill Formation” (Shen & Tamkin, Anthropic, 2026), but its scope is narrow *yet the results are still sceptical).
We're so far beyond lines of code improvements. Agents are building entire features with mergeable PRs that we now deploy with confidence. That was impossible six months ago and unimaginable a year ago.
So, my measure is not just line of code. It's deployable features. And yes, by that measure, we have gone from zero to "lots" in a very short time.
In my bubble, I’m seeing people write entire software stacks from scratch, single-handedly. At this point it’s clear to me that software engineering has taken off by a factor of 10x-100x, at least in some industry areas and in some circles.
But that’s not the point here - even if this is true and it continues, IT is not the entire world economy. I don’t think we’ll see the same productivity gains propagate more widely soon. This will take years. The real world is a ship that’s immensely hard and slow to steer, which is easy to forget when your CI pipeline deploys commits to production in minutes.
I can "write" a whole OS kernel by running `git clone https://kernel.org` — that does not make me much of the programmer. Copying software was never a problem — making the software perform the job you want was the problem.
Are you under the impression that LLMs take source from the internet and put it on your hard drive? That’s your good faith interpretation of what’s going on?
It does exactly what you say: takes source code from the internet and put it onto your hard drive, but in a very indirect way, occasionally verbatim, in other cases it renames identifiers and does a bit of composition.
You are completely discounting all of the original code being produced with LLMs. What percentage of LLM-generated code would you estimate is plagiarism? 1% 10%? 90%? 99%?
99%. LLM cannot do a bubble sort on its own — that's one of the main modes of failure actually, when LLM fails to recognize algorithm and writes its own implementation. LLM is basically recognizing, concatenating known algorithms, renaming variables, functions names, interpolating multiple similar solutions into one target solution — with a success rate corresponding to the number of solutions it saw for this task.
In case you would say "but LLM-s learn on random source code from the internet" — not anymore, modern LLM-s are trained on highly curated, preprocessed, and some synthetic datasets. As I've said — the copying is indirect.
I've submitted two PRs to fix bugs in open-source system tool last month. Both of these are targeted fixes to highly-specific bugs that existed in this software. The agents I used debugged the issue and implemented fixes and tests autonomously.
These problems looked slightly more difficult than bubble sort to me. Do you suppose these bugs existed in some other software and the agent only translated the fix between repositories? When the agent was debugging the software, were they simply following some tutorial they were trained on? Is this the 1% case? Or are the stochastic parrots regurgitating most probable tokens in approximately the right order to provide an illusion of work being done?
> In case you would say "but LLM-s learn on random source code from the internet"
You were talking about code writing, now you switched to debugging. Those are wildly different tasks, and some researchers suggest to employ separate LLM-s for code writing vs debugging.
For debugging LLM-s reproduce fine-tuning datasets made by humans actually doing debugging on some code. And yes, the problems AI found in your code were similar to the problems humans found when writing those datasets.
> You were talking about code writing, now you switched to debugging. Those are wildly different tasks
No, they're basically the same thing, intrinsically linked and inseparable. Likely represented together in training sets, too. I'm surprised anyone would say that. Are you a programmer?
> For debugging LLM-s reproduce fine-tuning datasets made by humans actually doing debugging on some code.
Not really, they are primarily trained on LLM-generated trajectories from executable training environments.
> And yes, the problems AI found in your code were similar to the problems humans found when writing those datasets.
So they're only similar now? I suppose, in the way that I learned how to solve bugs while learning to program and now I can fix similar ones? I guess we can now agree that LLMs are capable of learning and extrapolating knowledge to new problems?
>No, they're basically the same thing, intrinsically linked and inseparable. Likely represented together in training sets, too. I'm surprised anyone would say that. Are you a programmer?
No, they are not the same, and they require different tools: you cannot write code in gdb or valgrind. Even in IDE-s debug and editing interfaces differ significantly. You can literally separate gdb session from any code editing, do debugging with zero code writing and then relay the findings to someone who will make the changes.
The fact that you personally do both things does not mean everybody and everything does it this way. And certainly LLM-s don't function same way people do.
And no, debugging and code writing are not in same datasets. More to it, debugging and architectural planning is mostly represented in fine-tune datasets (prompt-response, agentic step prediction), while code writing is mostly (like 99%) pre-training data (refined internet scraps and regenerated code).
>Not really, they are primarily trained on LLM-generated trajectories from executable training environments.
Training on purely LLM-generated data is known to give horrible results. Big AI companies already have thousands of debugging sessions at hand — those bring high novelty/variability into the training sets. It is possible to multiple those fully supervised sets via LLM-generated bugs or bugs from github reproduced in a sandbox. With all that said, feeding lots of very similar bugs and solution trajectories does not help much in training — so it's rather "combinations of known human skills tested in a sandbox", not "totally random efforts transformed into datasets of successful debugging outcomes".
>So they're only similar now?
`c = a + b` and `c = a + d` are similar, but the difference is very small — it's about variable rename. It's not a unique code. And I never told LLM has to reproduce 1 Mb of source code byte-to-byte without any change. Just in case some of my words sounds like I told — I did not mean it.
>in the way that I learned how to solve bugs while learning to program and now I can fix similar ones? I guess we can now agree that LLMs are capable of learning and extrapolating knowledge to new problems?
No, LLM does not learn to program same way people do, it generally does not learn things the way humans do. Why would they? The training of a human looks nowhere close to how LLM-s are trained. Even the way people translate similarities is not the same as LLM's: LLM does precise literal swap, human extracts meaning and unfolds the solution into concrete words. Humans chase goals and plan things, LLM-s pattern-match and reproduce superficial shape, treating steps in complete isolation.
Human can try to reproduce actions with zero understanding of the subject — which usually leads to very mediocre results even after lots of effort. Basically, LLM can be seen as a stupid and blindly-diligent office worker x100.
> No, they are not the same, and they require different tools
Yes they are, and no, they don't. Debugging can be done by writing/deleting/modifying code (e.g., write a test and run it, or add log statements, or remove a line and re-run the program, etc.), or by running shell tools (either as single tool calls, which you may to be a one-line program, or with larger compound commands, or entire shell scripts - throwaway or saved to a file), or by writing code that analyzes debugger output and logs, or editing code at runtime with hot reload, or - depending on the runtime - attaching to the running program and running ad-hoc code (REPL-like), and so on. It's literally impossible to enumerate all of the options because it's all mixed in one codey-debuggy-flavored soup. Are you saying you've never seen an agent do this?
It's also fun to write your own debugging tools by the way - just imagine any debugging workflow you think is good, write the tool, give the agent some tips how to use it, and they'll do great. It's so good, you should really try it. This includes debugging tools that are operated by writing code, it totally just works.
> you cannot write code in gdb or valgrind
I'm actually pretty sure these are extensible, and I'd bet quite a lot that an LLM could write that plugin.
> The fact that you personally do both things does not mean everybody and everything does it this way.
I don't, but LLMs certainly do.
> And no, debugging and code writing are not in same datasets. More to it, debugging and architectural planning is mostly represented in fine-tune datasets (prompt-response, agentic step prediction), while code writing is mostly (like 99%) pre-training data (refined internet scraps and regenerated code).
Programming and debugging is self-evidently part of the same training loop. Agents write code, validate it and adjust trajectory. I'm not saying that they don't train agents for debugging-centric tasks specifically, but these certainly often - if not usually - involve programming as well. The same way coding tasks often involve debugging.
> Training on purely LLM-generated data is known to give horrible results.
I don't know what to tell ya, it's how the LLM sausage is made.
> `c = a + b` and `c = a + d` are similar, but the difference is very small — it's about variable rename.
I think you're vastly underestimating the difficulty of real world programming tasks. It's a pretty common mistake, I usually see this in people who are struggling to downplay LLM capabilities, and their strategy is to pretend that everything LLMs are good at was trivially accessible in the training data all along.
> No, LLM does not learn to program same way people do, it generally does not learn things the way humans do.
Obviously, it doesn't. I said "LLMs are capable of learning and extrapolating knowledge to new problems". You can throw a novel problem at an LLM and you'll get a novel solution. This is obvious to anyone who works on interesting problems. You could literally hit CTRL+T and do it right now. It's such an easily and accessibly verifiable fact. What a weird hill to die on, and yet so many corpses there.
>Debugging can be done by writing/deleting/modifying code (e.g., write a test and run it, or add log statements, or remove a line and re-run the program, etc.), or by running shell tools
We can end with things like "drinking cofee is an essential part of debugging for me". Debugging is an integration work, just like architecture planning, so it's somehow tied to plain english skills, reading docs, running whatever existing commands there are, parsing tools output is very important thing.
Another point on which we should reach an agreement is the definition of debugging, as I suspect you may have your own definition. Fixing compiler error is not debugging. Glancing over source code is not debugging. Debugging is characterized by running a program and comparing the actual outcome with the expected one — implying finding the exact part of the source code responsible for the problem. Meatbags often do not read the full related code, and sometimes are not even experts in that particular programming language, or simply don't own the code and have hard time digging into it.
Countrary, LLM-s excel at bruteforce code sifting and spitting two more 200-lines python scripts each minute. Still, even with AI debugging is not just "read and write code" — it's rather a high level skill with code read-writing being the largest hammer in the belt. Hammer so large AI will write several large scripts rather than use existing simple command.
>It's also fun to write your own debugging tools by the way - just imagine any debugging workflow you think is good, write the tool, give the agent some tips how to use it, and they'll do great. It's so good, you should really try it.
You conflate crutches with features. LLM-s absolutely suck at interactive tools, so they have to compensate with their strongest side: writing-reading code. Not because that's the absolutely most efficient way to debug, but because the economy is shaped differently in the world of AI. Running application manually, manually collecting logs, and then feeding them into AI is not "debugging made by AI" — it's the debugging made by human while AI doing the "code sifting" part of the debugging process.
>> The fact that you personally do both things does not mean everybody and everything does it this way.
>I don't, but LLMs certainly do.
When I look at my agent doing 1000 characters one-line python script for testing/debugging, I do realize there is no way it could be a common code-writing skill, regular python code does not look like this, the skill requires some special understanding of agentic loops, it cannot be from pre-training dataset. And you definitely cannot say that "being able to write code" automatically enabled "being able to debug" — just try any pre-2025 model that is decent at code writing but not trained for modern tools/harness (I'm pretty sure some newer models also lack this ability, but such models are harder to spot). I'm pretty sure everybody here tested pre-2025 models, it's really nothing to argue about.
>Programming and debugging is self-evidently part of the same training loop. Agents write code, validate it and adjust trajectory.
I would suggest you to re-examine some fundamental explanation of how LLM training actually happens, because it seems to me you don't differentiate pre-training and post-training stages (and post-training has lots of flavors in it). Trying to train LLM for agentic coding and debugging before it learns basic code patterns is a huge waste of resources. "Never say never" — if you manage to generate dozens of billions of agentic training sets, you can probably end up doing "coding+debugging pre-training", but so far neither Anthropic nor Google have resources to do that, rather they employ millions of fine-tuning samples. And I heart those companies have lots of money.
It is true that modern LLM-s are trained to do basic chain-of-thoughts/agentic loops at the very begginning of fine-tuning (i.e. after pre-training), but it should be noted that LLM can do implicit chain of thought and code writing BEFORE any fine-tuning. Just the base model can do that, but the format of input and output does not correspond to what humans and harness expect. And that's basically how the chain of thoughts was discovered: researchers asked LLM to solve some task step by step — and it did. Nobody trained LLM to do that, it just reproduced some patterns from pre-training sets. And then researches started training LLM-s to specifically optimize agentic loops.
> I don't know what to tell ya, it's how the LLM sausage is made.
Even small Phi-4 was pre-trained on only 40% synthetic data, despite Phi-1 employing 99%+ synthetic data. Once again, you should split pre-training and post-training.
On modern super-large LLM-s the large corpus of agentic loops for fine-tuning is indeed generated via external expert agentic loops, but it never becomes 100% as the model collapse is real and well known. Just in case the wording "model collapse" is not clear: the model starts using extremely limited set of tools to solve every task, even when much more efficient tools are available. I can see the tendency for model collapse on Grok 4.7 and Sonnet 5.5 I often use, and their engineers definitely partially counteracted this tendency.
>I usually see this in people who are struggling to downplay LLM capabilities, and their strategy is to pretend that everything LLMs are good at was trivially accessible in the training data all along.
I tried feeding non-canonical Tarjan bridge-finding implementation into AI approx year ago, and it went totally crazy: it says the implementation is incorrect, then traces it via supposed counter-example and finds it correct, then tries to give new explanation and once again suggests the implementation is bugged... The fact we put lots of crutches beneath the same pattern-matcher does not make it a completely different mechanism.
Most people never leave their sit during the AI ride so AI drives them through the pre-defined happy path. But once you try to deliberately leave this road, LLM-s often mark the code "wrong/unoptimal/poorly formatted". That's particularly a reason why older agents used to agressively introduce completely unrelated refactoring/formatting changes — because they just did not like how many spaces you've used in your code — and modern LLM-s still suffer from this problem, but their harness and agentic loops usually hide at least part of those meaningless changes.
>I said "LLMs are capable of learning and extrapolating knowledge to new problems". You can throw a novel problem at an LLM and you'll get a novel solution.
My experience is quite the opposite: I give a novel problem and LLM suggests 10-years old solution using libraries that are long deprecated. My whole life's hobby is about finding rare stuff, and I have very broad experience in this area. I particularly like LLM-s for finding some rare old thing I've never heard about, but in terms of making up something totally novel LLM-s totally sucks. Arguably, an ant is more intelligent when dealing with novel conditions than the most sophisticated modern LLM is.
This is small-scale but I know a lot of Blind people who are vibe-coding accessible tools to solve their problems. Most of them have never done any programming at all.
For example, my friend Marco has built like 10 different apps. I've used some of them and they're useful. I think this is an example of something that just wouldn't be happening on this scale without LLMs.
Agreed and I have also built many useful tools for myself. Many of my peers in business functions such as sales and finance have done the same thing, but these mostly amount to personal productivity, which is hard to measure. The claims I'm responding to are that software productivity is 10x or 100x - and we just aren't seeing any outcomes that support this. In fact, development work on our core platform is not more efficient at all in terms that we can measure - like delivered story points (which are not perfect, but are what we have).
We are coming an external tool as they 15xd prices on us. We have spent a lot of tokens and about a month, and are incredibly close to matching the external tool.
LLMs have made stuff like this much more possible.
Granted, we have an absurd amount of tech debt, but given our deadline that's an acceptable risk.
Note that the reason this works is because we can compare to the external tool, which is a really really good scenario for the current models.
In more greenfield areas they are less useful. Ultimately though, you can't outsource understanding, and that feels like the limiting factor now.
> In my bubble, I’m seeing people write entire software stacks from scratch, single-handedly.
This is becoming common, but we should ask what is it really? The analogy i see with these SME vibe coded apps is the same as the graveyards ie file shares full of forgotten but once useful excel spreadsheets and access forms/dbs. Coding was always part of the problem. This highlights deployment and adoption are also tasks to be performed.
Deployment for who and adoption by who? In an ai future that would be agents. App registries where agents can discover capabilities would seem to be a thing the future would need.
What about productivity gains (recent/profit, company bottom line)? I think that’s what the op’s measuring. Have you seen a 10x-100x on that front? If not, what’s the disconnect here?
Productivity != sales, especially in software, I think. This has been happening for just a few months, less than one project cycle. I think we’re about to see a huge wave of software products. But it’s possible most of these are competing for a slice of a constant-sized pie, though.
When your manager/boss asks how this huge productivity gain gonna lead to better product -> sale, and why are they spending a good amount of money on AI tools, do you say “Productivity != sales, especially in software”?
I don’t know anyone in this situation. No idea how it works in corporations, and I’d rather not find out.
I know some people who pay for their own tokens because it lets them do their job in 30 minutes a day, and spend the rest of their time creating a side project. I know several freelancers who didn’t make software before, and now they have custom productivity tools tailored to their jobs. Others work in small teams, where the boss doesn’t need to ask for proof because they’re burning tokens themselves.
It doesn’t matter who the boss is. It’s the same dynamic. That person needs to live, and they need to earn money. The money spent, needs to be justified in profits made, otherwise it’s not a business. I was questioning the absurd 10x-100x company output, which is relevant to the op.
> It doesn’t matter who the boss is. It’s the same dynamic.
I'm guessing you haven't had the displeasure of working for a boss who doesn't code and has no idea what he's paying you for - lucky you.
> I was questioning the absurd 10x-100x company output, which is relevant to the op.
I'm talking about programmer productivity, not company output. For pure software companies, these may be closely related, but there's many other processes involved with converting code into money. Either way, the 10x+ productivity increase is obvious to everyone I work with.
Yes. Because it is not my job to optimize for sales. It's to optimize for time to ship whatever features/fixes the manager/boss gives me. There's an entire sales wing who can answer the question on how/if shipping this out faster will actually make the company more money.
Asking the workers why they need to justify the responsibilities they were hired for is a small part of what's pumping this bubble up.
Good reference. It is going to slowly change things in 20 years we'll look back and a lot of things will be wildly different because of AI systems. But its rate limited by institutions and momentum and human psychology and the way things change will be somewhat unpredictable with unintended negative and positive consequences.
Yeah, it is not going to be slow. Progress compounds, everything is being built on top of fast computers and fast widespread internet connection. You are not at the beginning of the curve like you were in the 20th century when people didn't even know how to use a keyboard and we are already seeing widespread disruption in security.
Literally the entire premise of the article is that it's going slow everywhere but in software. To think otherwise is to be really myopic about how the rest of the world outside of tech operates. And in any case 20 years isn't that slow. The only things that happen faster that that will be things like adobe goes out of business. And let's say AI helps us develop medicine and robotics and better factories and all the other things that would have a big impact a lot faster.. just human trials could take a decade before you really see that much change. Will it have a wider impact faster than computing its self, or the internet? Maybe? Maybe not.
If a lot of jobs were still manufacturing jobs I would agree, but we are in a service economy that uses computers as the main tool. So I don't see how it won't come into effect fairly soon into other areas of work. You can't measure impact by profit alone so I don't find the article particularly interesting anyway.
It is like saying there is no impact on math because solving a millennium prize problem didn’t net them a profit. It is just a naive way of looking at things.
You are a demonstration that their marketing campaign is working. The reason OpenAI is focusing on math is because it is one of a very, very few domains where RLVR can be done at massive scale because large sets of problems with automatic verification can be produced and evaluated relatively cheaply. It has no value to OpenAI beyond marketing. It has negative value for mathematicians.
So far the only economically valuable activity that AI has really proven itself very capable in is coding - and its for the same reason of economical RLVR. It can do significant tasks completely end-to-end in that domain. It is very helpful in other types of tasks such as data analysis and legal research, but those things are not going to fulfill your expectations for broad economic impact.
There is simply no evidence at all for the broad claims of disruption in 1-2 years.
Seems to me like you're undervaluing the impact of software in the modern world, as if software hasn't driven massive productivity gains in every industry in the last 50 years.
I expect to be surprised by specific headlines, but I don't expect "software engineer" to be any more common a job opening than "steam locomotive engineer" in 10 years time.
Conversely, I would be very surprised if the robotics startups turn out to have solved general-purpose AI models that work in the real world, both fast enough and cheap enough (currecy or joules) to be significantly labour-displacing in 10 years.
Where I'm about 50-50 on which of "AI does it" or "AI does't do it" is more surprising, is specific physics research from AI. LLMs can do a lot of hard maths, there's lots of open problems in physics due to hard maths questions. So:
Consider the Alcubierre metric. The original is impossible. Can an LLM either design something in the same broad category which is physically instantiable but still useful (e.g. "yes it's sub-light but it manages 0.01c and you weren't getting that otherwise"), or prove this is impossible when limited to matter made from normal electrons, protons, and neutrons?
10 years? I'd be equally surprised if this is or isn't done by AI. Mainly depends on how long the growth continues looking exponential.
Progress is future knowledge. You don't know what future knowledge is, yet. When you know, it will have become present knowledge. You don't know that any particular ongoing change is progress. LOL predictions.
It IS going slow though. The rest of the economy will be slowly adapting it, just like they have slowly adapting other computer systems before. There simply is no use for frontier models in the real economy as of now. For the economy it's just another cost factor, a substantial one if they try to actually build something with it, that has no real benefits that results in increased productivity and revenue making. So why should anyone except AI companies adapt frontier models? They have no reason to. And it will be that way for many years to come.
Yep, and if it's anything like previous transformative technologies: the internet, and 100 years earlier the railways, a lot of people are going to lose a lot of money overbuilding the wrong things along the way.
Maybe it's also rate-limited by the fact that it's not actually a silver bullet and actually needs time to find the proper place for its application as well.
And likewise; computers took a few decades to actually become a widespread economic force. Arguably the impact was quite small if we separate it from the internet (which of-course depends on computers).
Hard to describe the impact of computing as quite small compared to anything, even the internet when you go see those old room sized early IBM computers that were the first commercially viable ones for businesses. They used to have warehouses of payroll documents and armies of people to shuffle around paperwork and crunch numbers and physical people connecting phone lines. Just absolutely fantastically massive amounts of busy work.
Good insight. For a long time, the utility of a computer was tied to its ability to print something onto a sheet of paper. User interfaces came along, too, though. Especially for industrial applications. Then internet was probably the third wave. Now AI is the fourth?
Yeah, the person who said that obviously never checked on book and magazine and newspaper typesetting --- cue old news stories about the last compositor at a given paper finally retiring after decades of the sinecures which the composition unions bargained for rather than participating, causing a loss of typographic skill and knowledge which resulted in the ghastly appearance of the "DTP revolution".
I once participated in a project where we worked up a database-driven typesetting system for a company whose annual publication had a staff of a dozen (3 managers (one of whom mostly did ad sales), 3 full-time compositors, 6 part-time) and when we put in the system they quit paying their expensive annual license for maintenance on the proprietary composition system, and fired two managers and 1 compositor and half the part-timers.
The staff was commensurately larger when they switched to the first electronic composition system they were using from the typesetter and paste-up they initially used.
> the person who said that obviously never checked on book and magazine and newspaper typesetting
That's one of the "everywheres" the author of the statement was talking about, which is the point - you could see every part of the economy seemingly improved, but it did not show up at the macroeconomic measurement level.
It was until the 2000s that computers caused an uptick in measured productivity at the GDP level, though.
The number of books being published has been steadily increasing for decades since computers became prevalent in typesetting --- is that not measurable?
Admittedly part of that is caused by the change in tax laws making it untenable for publishers to maintain a backlist.
It was not on a level that nudges GDP growth for a long time. The irony is that maybe the productivity didn't increase until book publishing plummeted, since enabling a lack of physical books was a huge productivity win.
This isn't a debate about whether industries improved. When you look at GDP, do you see an uptick in GDP or not?
The original quote was often seen as a paradox. At least one major interpretation of the original quote is that "the computer revolution is so obviously helping the economy, that if it doesn't show up in GDP, is it even a meaningful measurement, and how can we fix that?"
But it is very fitting to discuss that quote - the original article discusses multifactor productivity, the same relevant measurement, and earnings of the S&P 493. There are two ways to interpret it - is AI not increasing productivity, or are we not measuring productivity, and is there "dark matter" somewhere in the economy that is not being properly measured?
I think we're looking at two different sides of the same coin.
The thing is, increased efficiency doesn't always yield a larger market in the short term.
There are only so many copies of a book listing television stations and demographics for marketing purposes which the market will bear as in this specific instance.
It takes a while for the displaced workers to get retrained and find other jobs which allows them to produce/consume more.
On the flip side, Nintendo hiring a couple hundred folks to make a game such as _Legend of Zelda: Skyward Sword_ suddenly created $50 x 600,000 in value on the day it was released.
The distinction is adjusted vs non-adjusted basis. Serving inference is profitable, but serving inference for the last released model doesn't cover the cost of the model in training.
Dario explained it very clearly in an older interview. If every model is bigger and more expensive to train, each model can individually be profitable, while the company as a whole still shows a loss at any given point in time. This is why he likes the idea of regulation so much, it'll let them take a breather and recoup.
> Dario explained it very clearly in an older interview. If every model is bigger and more expensive to train, each model can individually be profitable, while the company as a whole still shows a loss at any given point in time.
“We’re profitable as long as you ignore our ever increasing cost of goods sold, we promise it’s not a pyramid scheme” - ~~Charles Ponzi~~ Dario
But the reason the cost is ever increasing is the steeply ascending quality. There's no (non-doom) scenario where this game does not end with profitable cheap tokens. Wherever it'll stop scaling, it'll be profitable to serve tokens at that scale. And of course if it doesn't stop scaling, it's the singularity and all this thought of profits is quite secondary.
An industry cannot fail because it's too successful.
edit: Other explanation. The industry has found a moneyprinter. It prints money, and several companies are building them and printing money. However, each one of those companies is cashflow negative, because they're all trying their desperate best to build an even larger moneyprinter. Why? If they don't build it and the next company over does build it, and it works, then company 1 is differentially outcompeted. what's more, the more expensive each moneyprinter is, once a company gets strongly ahead the second company won't get the financing to keep up with that one. ... But it's still moneyprinters. This is not an economic race that ends up with "and thus everyone stopped printing money."
> But the reason the cost is ever increasing is the steeply ascending quality.
Sure, if you say so.
> There's no (non-doom) scenario where this game does not end with profitable cheap tokens.
I doubt it.
> Wherever it'll stop scaling, it'll be profitable to serve tokens at that scale.
It’s much more likely, imo, that local AI models will completely destroy the entire business model of inference. For any entity that would be spending enough to replace multiple employees (allegedly), the cost of the hardware to do so is completely affordable. This has the major benefit of eliminating per token costs, and practically guaranteeing that some “AI company” can’t steal your company’s secrets.
At an individual person level, this hardware will take much longer to obtain, but it will happen eventually.
> And of course if it doesn't stop scaling, it's the singularity and all this thought of profits is quite secondary.
I still believe, strongly, that LLMs are the wrong path to AGI. There’s trillions of dollars being put into turning LLMs into AGI; however, there has no zero results showing that LLMs are capable of “learning” anything. “Skills” and expanded context windows don’t count. The “model” needs to be able to modify itself, in real time, for that to be possible.
We’re in a local maxima that strongly refuses to accept that it’s a local maxima. (Diminishing returns for the amount of money poured into “training” is beyond obvious to anyone paying attention).
> An industry cannot fail because it's too successful.
Sure it can. The fossil fuel industry is wildly successful.
It’s also boiling the planet alive, but that’s okay, it’s generating shareholder value so everything’s fine.
Edit for your edit:
> edit: Other explanation. The industry has found a moneyprinter. It prints money, and several companies are building them and printing money. However, each one of those companies is cashflow negative, because they're all trying their desperate best to build an even larger moneyprinter. Why? If they don't build it and the next company over does build it, and it works, then company 1 is differentially outcompeted. what's more, the more expensive each moneyprinter is, once a company gets strongly ahead the second company won't get the financing to keep up with that one. ... But it's still moneyprinters. This is not an economic race that ends up with "and thus everyone stopped printing money."
By your own definition it is impossible to achieve profitability without an ever increasing amount of money flowing into it.
There was a business about moving stamps around that needed similar levels of cash infusion to sustain itself. Eventually its bubble popped too.
> It’s much more likely, imo, that local AI models will completely destroy the entire business model of inference.
You're massively underestimating how much cheaper it is to serve models in a datacenter than at home.
> By your own definition it is impossible to achieve profitability without an ever increasing amount of money flowing into it.
No, what? No. This is just a total misreading of what I'm saying. It's extremely easy to achieve profitability. It just requires your business to stop chasing more profitability. The problem is just that then the other businesses who are still chasing more profitability will eat its lunch. It's a market shape that easily supports n-1 bankruptcies but never n bankruptcies.
it doesn't seem that complicated to me. what do you think the cost is for AWS to serve inference it sells to Anthropic for "$1.00"? I think it's about $0.11. Then, ask yourself, did Amazon sell it to Anthropic at all? No, they traded $0.11 real dollars for $1.00 of Anthropic shares (the CLN that created the real loss Anthropic is talking about), which are now worth $10.00 (which is why people talk about this as a paper loss). So... IS it profitable? It's profitable, but Amazon didn't JUST sell inference when they turned $0.11 into $10.00...
Fully agree with that. The problem is the AI boosters really believe in The Singularity that will transform all of human experience. If you try to tell them AI is a high-value enterprise technology -- more valuable than docker but less than the steam engine -- they get mad at you.
And at that point, there was no much productivity gained ... and not nearly as much debt as now. The industry was also not such a huge part of stock market, GDP, etc.
That would be speculative, it should only be added in the hypothetical time the productivity actually goes up. Even then, it should be some years in and with third party sources like all of Wikipedia.
It wasn't, generally, a question. People kept buying computers, on faith.
It's just that in 1987. the computers were generally and for the most part, irrelevant to productivity, just like steam and electricity were a promising gimmick for a long time before.. they weren't.
at the level of personal employee maybe, even though 87 had lotus 123 version 2. I was barely born but I'd think people would be able to think faster through electronic spreadsheet.
Only a few curmudgeons at this point are claiming there is nothing, nothing whatsoever, to AI.
The question is really about the fate of the current build out. For those of us who have extensive experience with AI, which is to say, those of us here, because we are probably the only people not only using it day in and day out but potentially using multiple sessions of it all the time, I think it's not so surprising that it isn't necessarily immediately radically transforming every industry everywhere. Even as it has rewritten the minute-by-minute process of doing my job entirely, it hasn't actually changed my job all that much at the higher level and it's hard to assign much more than a 2x or a 2.5x overall improvement to it once I account for all the Amdahl's Law issues that come up from the parts that have not been so amenable to acceleration.
And without unit testing, continuous integration testing, pull request review, the fact that the AI is building code which is itself a highly reviewable artifact rather than just doing whatever directly, and all the many many guard rails we've built up in our industry that happen to be extraordinarily amenable to use by automated systems, I still wouldn't trust AI all that much for a real task. They still bang into those guard rails a lot, and still make what I consider to be deeply suboptimal decisions far too often to take my hands off the wheel. With all the stuff we have in the programming space they are a game changer, but in a field that doesn't have anywhere near that much padding and safety nets I'd be way, way more hesitant.
I want to say it's weird to talk about such things when they are also capable of such amazing math proofs, but at the same time when you consider the numbers tossed around for how much those things have cost, I think you're seeing another space that has guard rails comparable to software engineering. By contrast to the way these AIs are working, when a human solves a comparably hard problem the human tries a few things but from the outside would seem to proceed almost directly to the solution, compared to the staggering amount of blundering about the AIs do before arriving at their solutions. Now, in terms of the final results, yeah, they got there and I won't deny that. I'm just saying that in fields where the acceptance criteria are nowhere near as sharp as math or software, they're operating at a significant disadvantage still.
> I think it's not so surprising that it isn't necessarily immediately radically transforming every industry everywhere.
Computers and the internet didn’t radically change every industry everywhere. But every industry everywhere uses computers and the internet, and more pertinently, they pay good money for it.
AI doesn’t need to transform the world, it just needs to pay off the sizeable investments in data centers.
Does AI transform the world more than;
- CAD software
- Office software
- Spreadsheet software
- Finance software
- Digital banking software
- Computer graphics software
- The entire US national fiber grid
Cause at the current rate and scope of data center investments; they have to transform the world more than... every other software system we've come up with until this points combined; to justify their cost.
They have to transform the world more than whatever the bond rate is right now, (which admittedly is increasing). I'm not sure that that's greater than office software, digital banking etc. But I don't think it's as high a bar as some people think.
From a paying back data centre debt point of view, yeah. The equity valuations on the other hand are priced for total transformation in the short term.
And is there enough business or computer demand for all the data centers we're building? I don't think anyone has an answer to that.
In a way it’s sad that AI came in the age where a lot of us are jaded because the internet which we built has become a place for gambling and scamming. I remember when a new language or framework was met with excitement. Today I hear (and am a source of) mostly criticism.
I think there's also something to be said about how the largest private AI enterprises are themselves mostly just software companies. There is enough widespread dogfooding of the product that the company staff are not just spending a lot of time and energy in improving how their product is specifically adapted and trained for better code but they are also intimately aware of those limitations, driving how they think about and implement their own product.
I don't think that translates well for other fields at all. You can throw an engineer or an MBA-brained exec at a business problem all you want but until they start using the products, they're not going to be effective at guiding its development.
Math is just proof of concept for the big players. The same resources could go to prompts designed to optimize the internal combustion engine or anything else nonsensical.
The biggest paradigm shift has been representing models as humans, and I don’t think that’ll fade. I highly expect the AI Model T Ford will be a generalized assistant that actually works. Likely developed by OpenAI or Anthropic. And likely 5 years from now or more.
"optimize the internal combustion engine" that requires good quality combustion simulations and iterative optimization algorithms; and we've had that for longer than the internet; that kind of work was the earliest uses of military comptuers. Why would a text model be better at it?
Maths i a cool benchmark, but has the similar benefit to software; very easy to verify.
> I still wouldn't trust AI all that much for a real task
I wonder if you might be conflating LLM/Agent with AI
Waymo is AI, I personally would happily ride one, or a Zoox. Those are quite real tasks the AIs are performing, people put their lives on the hands of these AIs
A lot of AI models every day work to provide search results, driving directions, image recognition, automatic routing and classification (eg. Fruit packaging companies use vision models to accept/reject fruit for packaging in real time - that’s the fruit we buy at the supermarket and then eat at home), and so many other things
So inadvertently, and without much choice either, you probably already rely and depend on AI on your everyday life
A* path finding and SLAP/SLAM (used in video games and robotic systems), BFS/DFS, Alpha-Beta pruning, Simulated annealing, Genetic algorithms/programming....
All of those are considered AI, none of those are LLMs or agentic harnesses (which of course are also AI). Also, most real AI tasks systems currently deployed in the world, don't use LLMs or agentic harnesses, or diffusion models, or image/music/video generation
This seems like a poor comparison, for any given company, benefitting from personal computing in 1987 was a significantly harder task than using AI tools in 2026 is.
The first requires a fundamental shift in not only how you work but expensive and complex infrastructure changes from the company itself just to make it possible (the first Mac for example would cost about $7000 per person in today’s money and would be outdated in maybe a year or two), an office building of The latter requires asking a subscription service to do a task for you on basically any computer connected to the internet.
I don’t know about you, but for me materially useful AI costs around $10,000/mo in tokens. But it is very useful.
Claude subscriptions help, but only if I come in with the expertise to build the entire system around the model and I don’t mind the whole thing grinding to a halt when my session limit is saturated. Anyone selling the system off the shelf is going to be charging API rates. Computers were cheap in comparison.
I can’t speak for everyone, but I would intuitively venture to guess there would be some greater than marginal productivity benefits from say, a $200/m Claude sub for each employee in many organizations outside of tech.
Perhaps they’re all on the $20/m plan which is useless?
“Marginal” improvement isn’t really going to be enough. Products need to be closer to 10x better than the alternative to get traction (which can be divided across several dimensions). AI, I think, is just starting to get there.
I get what you’re saying but I’m just not a fan of this type of argument. It always paints AI as this “in the future” thing which absolves any current criticisms or observations.
LLMs are great and definitely make you more productive, for some reason 493/500 firms haven’t been able to figure that out even though it’s trivially easy to provide employees with frontier models and it’s also trivially easy to use frontier models.
Which is why I say comparing it to personal computers in 1987 is not an apt comparison. That took redefining physical workspaces, new infrastructure, and new personnel which wasn’t easy to come by. Becoming a company that uses AI is just as easy (if not easier) as using any other SaaS.
See sibling comment. 10x measured how? In pure coding tasks under certain conditions the speedup is huge, but seems really unclear how this would transfer to any significant speedup across the whole undertaking (project, delivery etc).
Maybe, even likely. But how do you observe this? We are in this situation at work, it feels like "sure AI is helpful" but do we actually deliver any faster?
It’s a business-scale cost. When I say materially useful, I mean replacing multiple salaries and improving quality at the same time. Not some variant of “make an existing employee 20% faster.”
The personal hobby version is to get the Claude max plan.
One industry example.. We have been giving students computing devices for years with the cost of educating students going up year after year.. Yet scores keep dropping.. No productivity gains here.. I do not think these results are unique to the education sector.. There are probably alot of businesses where nothing much will change except slight efficiencies.. not enough to move the needle..
This is one area where we've proved to be fully unable to rise to the occasion.
The internet age opened a million of possibilities to improve education: improved visual explanations, scalable access to the world's best educators, lowered friction to access students to work with, replayable lessons...
We mostly just kept the old model and placed an addictive device in front of children.
Maybe it’s possible in theory to leverage our newest technologies as educational tools. But maybe it’s also incredibly challenging because it’s so easy to fuck up: a lonely peak of optimal outcomes surrounded by a vast space of net negative outcomes.
You mention the addictiveness of the devices, this probably plays a big role in shaping the outcome space.
Let's take math education as an example. If you look at the actual result, all math education past y=mx+b is a waste of time for most students. By and large, math students are taught by people who have memorized the quadratic equation, or certain operations with logarithms or imaginary numbers, etc. but don't understand it on any kind of deeper level. (Imagine asking a high-school math teacher to compute i^i. Almost none will be able to.) So it's presented as a set of random rules you have to follow to get the correct answer, with no underlying logic or justification. Most normal people understandably hate this, and even if they suffer through it to get a good grade they will not retain it and come away permanently hating math.
It's just sad. I sometimes help kids with math, and it's amazing how screwed they are. Like, smart kids with good grades who are totally unable to do basic manipulations with fractions. I mean questions like "write 25% as a fraction". They are getting absolutely nothing from math classes. How could they? Even the best teacher in the world can't pause class for everyone when they realize that one of their students is missing some foundational concepts. That student is just going to get left farther and farther behind, and they will never catch up. It happens to almost every student at some point. They'll get sick for a week or just have their mind on other things, miss some important concept, and then be left behind forever.
I honestly have a hard time imagining how we would possibly make the situation any worse with technology honestly. But if you just took Math Academy and made students use that instead of their math classes, I bet 99% would have better understanding and fluency with math than they do now.
I'm sorry this was your experience with maths in school. My experience was nothing like it, but it could be to do with where I live, my particular teachers or the classes I took.
> Imagine asking a high-school math teacher to compute i^i. Almost none will be able to.
Mine could. And of those who couldn't they would have leveraged the technology available to them (Google) to find the answer with me, and would be able to make it make sense to me. This is an example of technology bettering my education.
I just found a book about abandoned teenagers from the 70s (think juvenile held in post-offense rehabilitation centers) and they speech level is above the average educated adult today.
There's many breadcrumbs like these, of a change in culture globally. 50s social expectations about expression and characters were way deeper. And it gradually fades away as life becomes comfy and entertainment based (80s+) I'd say.
This is what irks, and saddens me most about applications of AI not improving outcomes. I suppose I should take some comfort that even Salman Khan couldn't crack this nut.
The idea of an AI tutor that could step-by-step customize an education plan to each student and patiently apply every approach to getting the students to understand each lesson, is such an attractive vision.
Both of my kids chafed under lock step pacing of lessons, even in a school where disruptive or unmotivated students were pretty much nonexistent.
The education industry has been lowering the bar for admission to higher learning for half a century or more, while dumbing down the content of high school and even primary school education. All because we can’t have children feeling bad.
It’s terribly ironic that things like calculus, linear algebra and matrix manipulation were being weeded out of the curriculum, right before machine learning took off.
What you're talking about are specific local policy changes at specific school and equating random trends to some sort of ... "kids these days" message.
If we're talking about the US; then it's simply the very healthy and functional economic inventive of students taking out lifetime loans to pay school tuition for schools that have not actually risen in quality or format in 40 years. The more students; the more money to the school board! Lower those bar riiight down!
I would echo that it’s local. I often poke fun at my SO and her sister for their absolutely trash tier quality education supplied by the Indiana public school system. I’m often flabbergasted by what they didn’t learn in primary and secondary schooling. They were able to go to decent universities, though, but it really opened my eyes to how terrible red states run their public school systems.
It’s not even a flyover state thing, if you just look at Wisconsin and Minnesota. It’s uniquely red state. Also Indiana department of transportation is terrible as well.
> while dumbing down the content of high school and even primary school education. All because we can’t have children feeling bad.
I suspect that it's a lot less about making children feel better and more about keeping the general population stupid and unable to think critically so that the ruling class can exploit them easier. The children of the ruling class are still given the opportunity to be well educated by private institutions that are overpriced to keep out the poors.
you're conflating a lot of different stuff together.
if someone knew a better way to improve test scores without cheating on them, we'd use it. it's a very hard problem.
on the flip side it is pretty obvious how construction and manufacturing can be automated and what the impact will be. it's so obvious that we already tariff and ban stuff that, from our POV, is automated, like prefab buildings, for solely political reasons.
This is an important observation. We've had 50 years of personal computing, but is there any evidence that computers in any form ever improved the academic performance of students?
AI certainly seems to be making students much less competent.
No, we know that when technology (particularly: Internet connected laptops) is introduced into classrooms competency declines. This has been shown to be true internationally
It's... complicated and a bit disputed. The trends match but the attempts at proper comparison study is not quite as conclusive (good social studies are difficult to make). It's certainly true that the trend is negative rather than positive.
I don't see why it would. It might make education more efficient like a factory in some ways, but I simply don't see why it would aid in learning. At least, from my experience as a student, here are some observations:
1. Computers have allowed me to gather more information, access more journals, more results from around the world, but what really seems to matter in learning is just sitting down with a tiny bit of material like one book and REALLY understanding it.
2. Computers make things more distracting because while they make typing up reports, proofs, and assignments more efficient, there's actually just too much information. What really seems to help is having LESS information so there's some space to just think and work things out.
The fact is, what computers do is help you keep up with other people in the field, so they might accelerate progress at the research level, but what they don't do is help you become a rigorous, independent thinker.
Academia in fact has always been an uneasy tension between the two: building capable individuals and getting results and money for progress. Computers help with the latter - they certainly impede the former.
Why would a technology that makes the mind work less, have benefits for a process, that works by making the mind work hard?
You still need to calculate calculus yourself long after you could buy books where the result is written in. That supposed "paradoxon" is not even specific to electronic tools. It's just that the point of education has never been to get results, but to exercise in as much as side paths as possible and memorable while following the path to the results. The result mostly get thrown away immediately.
I gave my son CRPGs for his GBA that required reading what the characters were saying. His reaction to finding out his three and four year old friends didn't read was interesting.
We can use a non-deterministic system because we have test suites that act as validation gates.
Other industries don't, so they have no way of knowing if the AI is bullshitting them outside of manual validation, which makes it more expensive and slower than doing it yourself.
It's a good point but in most of the places i've worked an average person could leverage any 2026 llm to improve things 50% easily. And I'm more anti-llm than not somehow.
- fixing regular computers issues (confusing browser or os behavior)
- finding more possibilities with your current programs (around 2020 someone once asked me if I knew a way to capitalize a word on a computer, when she realized there was a dedicated shortcut she became ecstatic)
- having a rubberduck to discuss your daily work dead-ends
- thus decoupling from team friction as you need to ask less questions to people that might hurt you.
In more technical places, having a llm write small automation would also add up.
I strongly agree that search engine is the best application of LLM so far. 90% of time I don't look at Google search results, I look at AI Mode output only. And you don't need agents, harness, guardrails, etc for it — basic small clean LLM does the job well without overdoing it.
Your information seems to be outdated, unless every single street I’ve been on for the past year is “vetted”. All the other (consumer) car brand’s self-driving systems do have that limitation though, maybe that's where the confusion came from. Tesla is self-driving 98% of my mileage, with the 2% being when I take over to choose a specific parking spot or I’m towing/carrying something big.
Don’t know much about how it does in the rain since moving to SoCal but in clear weather it hasn’t come close to hitting anyone yet. Even proactively warned me a few times while I was driving. Maybe I should let it take over in parking lots too :)
It'll be hard, when I'm never causing any accidents or speeding or running reds/stop signs. Maybe they could get me for inattentive driving but that's what the tinted windows are for.
In most cases it’s only the object recognition systems that use machine learning. I know that Tesla was talking about using non-deterministic AI for the decision making as well, but it’s a worry if they’re already doing it.
Earlier Tesla self-driving software used search-based planning and conventional optimization alongside neural networks. Tesla has publicly demonstrated A* in its path planning work.
For those who won't act ethically or at least reasonably, regulation is the answer. We're not too far away from flipping the congress into something far more willing to do just that, especially with anything connected to Elon Musk.
There are many tasks humans perform which have no or slow feedback loop. Even if they can audit/validate AI output, they don't know if it will be good when it reaches other humans.
"AI does not understand accountability, trust, culture, unwritten rules, ethical trade-offs or fairness, all of which shape reputational risk and ultimately drive revenue growth and profit margins."
For a sec I thought he was talking about private equity.
We're not seeing full automation for things that ought to be automatable now.
Customer service chatbots still aren't very good. Salesforce ought to be taking over entire marketing operations, not just supporting sales but replacing it. The medical office side of insurance ought to be automatic now. The pick and place side of warehouse operations should be robotic. Shelf restocking in stores ought to be robotic. All of those things have been demoed, but didn't go to production.
The thing is with a lot of the world, you could say about the same in 1950 about why isn't everything automated by now. A lot of the technology needed to do most of everything we do was already there. They had vending machines selling cigarettes back then and today people buy their geekbars from a person in the 7/11, and we think we've progressed. Factories today still mostly only semi automated. You'd think mcdonalds would just be a single unit deployable crate burger factory by now, but it looks the same as it did in 1950 for the most part with humans doing all the work.
No one is actually given the agency in any company to set about and engineer and implement an automated solution that replaces existing business units. They also have to be given slack to fail and iterate. Easiest to just run mcdonalds mostly as you ran it the last quarter lest you rock the boat too much. Rinse and repeat and people still bag fries today, and the deep fryer remains probably the most dangerous part of the kitchen.
That's an interesting point. Vending machines never became a big deal outside of Japan. Fast food was automated in 1964, with AMFare [1], but that was never deployed beyond one prototype.
Whether general purpose robots will change this is interesting.
It looks like they're going to cost about as much as a car.
(There are at least 18 makers of humanoid robots. The hardware situation looks good.)
That's cheaper than humans.
There are already companies that set up industrial robots and charge for them by the working hour, with all maintenance and setup included.
> All of those things have been demoed, but didn't go to production.
There's probably a good reason for this. Demos usually just show the happy case, whereas in practice you have to deal with all the nasty edge cases of reality.
Yes, AI is still not particularly prevalent in the customer-facing / consumer-facing things you described. But internally, I think these companies are up to their eyeballs in AI automations. At the company I work for, the customer-facing side of our business has not really been changed by AI (except for faster iteration of website design, and a somewhat limited customer service chatbot - not particularly good like your experience). But internally we are using it for so many things, for example we have brought 3rd party analytics tools in-house (cost reduction and ownership over our data) by writing our own internal tools, we have setup custom systems to give ourselves better visibility over the business to try catch issues earlier, we use it to produce internal reports, etc.
I don't have much experience in robotics but I've heard that machine vision models have absolutely been changing the industry. And I imagine they are also using AI to write a lot of software too. But not my area, just what I've heard / been reading.
> But internally we are using it for so many things, for example we have brought 3rd party analytics tools in-house (cost reduction and ownership over our data) by writing our own internal tools, we have setup custom systems to give ourselves better visibility over the business to try catch issues earlier, we use it to produce internal reports, etc.
How much of that is automation of "bullshit jobs"?
Fair enough. It's a mix. But for the most part it's a net positive.
We have replaced external tools with in-house tools written by largely AI. These are not one-prompt tools made in a day. We have built our own analytics stack which non-technical coworkers use everyday - they are not interested in "bullshit", they need real, accurate data and visualisations. And we had a cross-over period where our internal tooling was compared to our (now legacy) 3rd party analytics tools. It was done properly.
At least where I work, I don't think there is much automation of "bullshit jobs" - people don't typically just waste time chasing fairy tales (for the most part). But for things like report generation or tracing issues or generating presentations or infographics, there is certainly a much higher volume of these things getting pushed around which may or may not actually achieve anything, or be at all correct. So I reckon a large chunk of these would fall under into the "bullshit" bucket.
Because everyone can produce so much more "stuff" with AI assistance / automations, a good portion does end up spewed straight out of the slop factory. How much / what proportion kind of differs person-to-person tbh.
The reason the payoff isn't visible yet is because outside of software engineering, it isn't systematized nor integrated into workflows. But believe me, people are using it and increasing their free time. I was at an HOA meeting in Reno. The board president works for a major US cable provider doing "contract approvals and negotiations", whatever that means. This last year, he said he's cut his 8 hour days down to a few hours using Claude (pretty sure he is just using the web chat interface). Employees of the property management firm are using it too -- our representative at the meeting said it had been a lifesaver for her in recent months. We haven't seen the full benefit from their end but they are also in a state of constant turmoil as a company so it's a wash. A tennis partner of mine who is the regional (entire west coast) manager for an automotive company's extended warranty program has also been able to use it to cut down time-consuming documentation and the firm encourages it; but again, no direct integration into her workflow, she is just winging it with ChatGPT online.
I currently work in E-commerce, local online marketplace in Kazakhstan. I scraped all millions of goods in the marketplace, then analyzed top 30k goods with AI. Goal here to find it's supplier in China accurately and estimate weight, similarity, etc. That index gave me huge competitive advantage compared to other sellers and extremely boosting my income.
But that cost a lot, because, it's literally like running new Claude Code session for each good of tenth thousands, with custom tools like search in alibaba, viewing reviews, etc.
In marketplaces margin/ROI of goods are distributed like(just for example):
Top 1% goods (P01) - ROI 1000%+
Top 5% goods - ROI 300%+
Top 10% goods - ROI 100%+
Top 20% goods - ROI 50%+
And so on. So such automation helps me to find that Top 1% first
It’d be nice to stop calling it The Magnificent 7. Perhaps The Circular 7, or The Shady 7, something more realistic about how they are treating the stock market. Yeah number go up, but the means to those ends are shady as all hell.
I think it's the opposite. The stock market is doing this to big tech. In 2022 these stocks were sold mercilessly until they announced layoffs. CEOs fell in line because their ultimate boss is the stock market. Then in 2023+ it was all about how many times you can say AI. Until 2026 where it's a little bit more "show me some revenue growth". These companies are not the ones controlling their own stock price. It used to be the case they bought back a lot of stock, which you could say is a sort of "manipulation". But now the opposite: they spend over 100% of profits on the data center buildout.
We’re heading into corporate budget season. Company after company is slashing their AI budget and the head cut wagons are circling around all the various “AI lab” teams that every company invested in two years ago with zero ROI to show for it. All while massive debt piles up to build data centers despite little evidence any of this is coming close to returning the ROI needed to justify any of it.
Even in the M7, I doubt AI is having much impact other than benefiting from AI spending. Its hard to point to any non-AI products in the M7 that have meaningfully improved since 2023.
The problem isn't that "AI doesn't understand..." this or that, as concluded by the article though. Its more a lack of creativity and incorrect reasoning about what AI can and cannot do. For one, whole role replacement is pure fiction and likely will never happen with LLMs. I expect most businesses will look very different in the next 10 years due to AI and it will not be smooth as the needed creativity will mostly happen in startups (and perhaps companies that can truly operate like startups) and diffuse out from there. I'd expect a lot of churn in the S&P493 and even the M7 most likely.
One way they seem to have benefited from AI is that it reintroduced the idea of paying for new software features. For many years Google just gave away everything for "free" (modulo ads obviously). Now Google adds a new feature and adds a little sparkle and says "pay up!"
I think most people in the tech sector and especially driving the AI hype have absolutely no clue how difficult it is to implement change in complex organisations. There are a bunch of technical reasons but half the reasons are not even technical - social, legal, etc. You have to shift the practices of humans and they have incredible inertia. Like decade level change response period. And during the change period they are less productive. Expect a decade type period before "productivity" impact manifests and even then it won't really be measurable directly because people will be doing different things.
> I think most people in the tech sector and especially driving the AI hype have absolutely no clue how difficult it is to implement change in complex organisations.
Simple organizations too, and sometimes especially so. The amount of unrealized productivity many small business stand to gain just by adopting simple automations that computers have been able to perform for decades may be as large as (or larger than) the productivity gains ML model and agent adoption enables. And that adoption may just be enabling someone who never wanted to learn Excel macros or regular expressions to command an agent to go do those things which have been possible yet unrealized in a particular small business for decades.
The problem has largely been (and remains) humans set in their ways, of the sort who stopped reading books or seeking any new knowledge whatsoever the moment their mandated or socially expected schooling was complete. Incurious midwits are definitionally the bulk of any workforce, and getting them to change any process is like pulling teeth whether they're working in a small business or a megacorporation. Change often doesn't come until the midwit on some particular throne retires, quits, or is laid off.
When I tried to use CoPilot at my day job, I set up a prompt to process scans of checks and invoices (with the MICR account bar code line covered w/ a pen so that there was no PII), it would sort of work for 20 files at a time to write a .bat file command to rename the source image based on the content of the check and invoice --- this removed the need to open up scans in batches of 50, read the Invoice ID, save using that, then run a .bat file to rename them based on the balance of information in the database.
The thing is, the failure rate was quite high --- often only 15 (sometimes fewer) lines were returned, and after a while, Microsoft imposed a limit on uploading more files, even though I was deleting the conversations immediately and quitting CoPilot and re-starting (long sessions seemed to have a progressively worse failure rate).
Since then, access to CoPilot has been removed and we're waiting on a test program to finish. Kind of curious how many tokens this was using up (and if there's a model which could be run locally which would work well and quickly enough).
I recently had a somewhat similar task: My 80-year old friend who is on Medicare got a $30,000 hospital bill which the insurance company declined to cover. So he wrote by hand a letter to the insurance company to request a reversal of their decision. I scanned his hand-written letter and asked AI to turn it into text (to create a professional looking letter). Could not get that to work on Gemini nor Perplexity, but it worked on Co-Pilot.
It did save me a lot time to use AI, especially should next time since I now know which AI app can handle it. So, there was a bit of overhead before I found the AI-app and prompt that would do it.
I have another friend who works for SAP as a project manager. She said she hates the fact that management wants them to use AI for everything, giving them a bonus if they do. But one reason I think their employer insists on that is that when you try to use it for real tasks you discover what works and what doesn't. Gradually we should be getting better at it.
Currently there is still a penalty caused by
1. AI not being able to do a specific task
2. Users not knowing which AI-app can do the
job and
3. How to prompt the AI to get it done.
Writing the prompt is what users must do, and writing a prompt that works is a skill they must learn for every task you want AI do for you. Therefore progres is is happening but incrementally.
Enterprise sales somehow convinced a lot of orgs that all AI is the same or operates on the same level. Even frontier labs have differences, and you have to use each for your own advantage.
LLMs are not the type of model to use for this task. There are more specialized models that have much higher accuracy and full results. You can probably find the latest by looking around the OCR ecosystem.
A co-worker in a different program whose invoicing was much simpler and more straight-forward tried to automate this using OCR and Python and had a higher error rate and still had to set hand-written checks aside to handle manually.
I couldn't find an OCR system which was:
- free/affordable
- able to cope with arbitrary check sizes/hand-written checks and my occasional need to arrange multiple invoices in a stack with the check rotate 90 degrees so that all the numbers are visible
- which would capture the data from a scan in such a way that it could be easily used to rename the file
If you know of a system which is flexible enough to do this which is essentially free, let me know.
Paradigm shifts happen when old minds retire or die. Something like that, but for companies. If you’re starting a company now, you are _definitely_ hiring less people much later, and different profiles.
For sure. The original quote by Max Planck in his "Scientific Autobiography":
"A new scientific truth does not triumph by convincing its opponents and making them see the light, but rather because its opponents eventually die, and a new generation grows up that is familiar with it."
Admirers of this idea shortened it to, "Science progresses one funeral at a time."
Makes sense. My RTX 5090 can locally take a text prompt, generate a 2D image from that, feed it into a 3D generation model, rig up the animations, and make me assets that are better than a lot of video games on the market. Not Cyberpunk 2077 level m, sure, but pretty good and this is just what I can clobber together in my free time from open models.
I'm a dev not a video or 3D guy, but I'm looking at people I used to know and also some youtube doom-scrolling, the blender MCP, cinema 4D MCP, and seeing that it's a big impact
Current evidence is compelling us towards the shared opinion that AI is changing coding and nothing else… but do remember that AI itself is code and we are in an early
phase of recursive AI improvement.
We have no idea where this will take us but it’s like measuring the effects of COVID in 2020. The immediate outcomes (temporary unemployment, massive stimulus) created long waves that we are still encountering (inflation, global instability).
I am in a tech role that straddles traditional business. Yes, LLMs are very adept at software development and of course we are early adopters, but the amount of other things that can be automated or at least significantly augmented hasn't scratched the surface.
At the start of this all, execs were clamoring for "AI" and at that time the models were superficially impressive but fell short. Now I think they are all underestimating the amount of improvement in the last year. There's loads of work that can be automated
I have always wondered how this technology is going to actually effect the majority of the GDP in our country. Example: How is this LLM technology going to benefit me and make me more money in my laundry mat? How is it going to benefit the tourism industry? How is it going to benefit the Hotel industry? How is it going to effect the Bar and Liquor sales industry? I dont see LLMs bringing significant cost savings in the majority of the industries that make up our GDP... It brings me the benefit of being a first google type search before I dive into mechanic manuals to fix my equipment.
I dont see having a morgan freeman AI voice saying announcements and me asking it for how to fix my machines being much better than what already exists in my industry...
I have worked in multiple companies where the internally developed software was central to the business success, yet it was handled in an extremely bad way, preventing it from using even a small percentage of its potential.
The (good) developers were not the main problem there, but the management.
The management did constantly override technical decisions and/or switched the main goals/vison around all the time, preventing the development from finishing anything.
In several companies, they replaced the developers with ticket-crunshing code-monkeys that have no interest at all in the business at and who bring no innovation, ideas, or even feedback into their work.
Some of that replacement was active, but most was by just making (work-)life miserable for anyone who cared what happened.
In my case, one of these companies even offered me a big payment (one that I will probably never get anywhere else without switching to another country) for coming back, because I understand the business good enough to always fix the fires in the background that the management caused, but they still rejected any good ideas and made life worse instead of better for the people that dependent on the software.
It's been almost five years since I left there, and I still have several friends in that company. One of them is one of the users of the software, another one is a developer who cares, but he could not switch because he is stuck there for several reasons. One of the main reasons being that he has medical issues that make most companies reject him by default.
They keep me updated about the companies progress, or the lack of it.
The latter guy told me this week, that the company made less progress in the last four years, than it did within a few weeks while the original dev team worked there. And he really wasn't exaggerating. Oh, the company has a big AI budget and they made pretty much every model available to all employees. They just can't use it for good ideas, because they will be garbled up by the management and turned around to be bad.
As long as they don't replace the management with AI, this technology will not improve their business.
I dont even think theres a payoff in the software dev.... seems to be great to spin something up quick as a POC, then maintaining it requires people immediately who struggle with reading the amount of code it produces.
AI payoff in the tech sector so far is the loss of thousands of jobs, the disappearence of entry-level positions, a flood of low-quality software and more cybersecurity risks.
Thus far AI has been great at extracting value from every area its introduced in, but terrible at actually generating value.
There are three kinds of lies: lies, damned lies, and statistics.
I'm pretty sure most of the growth comes from companies laying off useless COVID-epoch engineers. If you lose the dead weight, you climb up without ever gaining more profit. And IT companies hired the most during COVID — the rollback is logical and it has nothing to deal with AI.
But this theory is harder to grasp than the sweet "growth is coming soon". Personally, I don't see any signs of long-term recovery after 2022, the US gov. debt growth is unsustainable, and the global market falls apart i.e. GDP growth while real markets collapsing will just collapse the currency.
'a.i' is unlike software before - where it was build then deploy with lack of customization except of a few high ticket items such as SAP, Workday etc.
for 'A.I' to work for the rest of industry outside of software - you need high touch customization.
I think we are still in a value discovery phase, as the token cost of one shotting someone's job is not cost effective, you have to constantly pay for tokens. Distilling someones role and breaking it up into AI responses and have AI build the deterministic workflow is where the efficiency is. It should be an upfront cost and then stable efficiency but I don't think many industries are setup to even begin this transformation. Someone else mentioned it's kind of like electricity, in itself productivity doesn't change, it's systems adopting it that compete and scale versus non adopting competition.
I think you could say the same thing of tech and the economy prior to AI.
No Signs of Tech Payoff Outside the Tech Sector. Manufacturing, basic science, education, etc languished while Big Tech reaped record profits and achieved stock heights for the last 15 years.
Hell, on top of that (i.e. the confounders you mention are very very apparent everywhere,) the challenge of folks understanding how to use AI to be more productive.
Even last week I had a case where, I was reviewing code that I know the dev was vibe coding, and honestly was good up to that review.
Buuuuuut, for the stuff I reviewed that day, there was some base-class inheritance use that was so ugly that my initial teams message to the dev was along the lines of 'What crack was Copilot smoking on that one'.
Turns out, they had asked some other random-ish questions in the chat they were doing their actual work in, and those questions more or less polluted the context.
Sure, I was able to guide them for how to not do that in future (i.e. keep unrelated questions around the work in question to a separate chat, etc.) but this was also a dev that is normally both very good, but also overworked semi-related to the general 'AI will save us on dev costs thing'.
I can't help but ask whether, for the even less-technical people that are told 'yeah you lost two colleagues but here is Copilot/Claude/etc' and they don't even get enough advice to have it actually increase productivity versus chasing down gremlins in the actual work.
Just talking about salaries, revenue, and stock IN AMERICA.
Doubtlessly Chinese manufacturing benefited tremendously from those advancements. But my point is the headline whinging about AI in America would also apply to Tech in America.
I think what people fail to appreciate, is that past a certain size, organizations and the people within them get so tightly wrapped up in working a certain way that it becomes nearly impossible to fix without rebuilding the whole thing from the ground up.
If you want to see if AI is paying off, look at new startups that build AI-native versus those that don't, controlling for the type of industry/product. This is a difficult causal inference comparison to make, but it's the one that I find most meaningful and interesting.
No signs of AI payoff outside the tech sector...yet. I think it's partially because our harnesses are generally much, much better. And engineers seem to be more deft at figuring out where AI provides real benefit and where it needs a watchful eye.
I'm seeing non-engineers use AI to send more emails, have more meetings, create more documents, and create more dashboards. So much of the busy work that happens in corporate environments was never that valuable in the first place. A firehose of busy work doesn't move any needles that matter.
There is nothing I dislike more than people using AI to generate tens of pages in text and then expect people to read them, when they did not read them themselves.
Well this is an actual effect. It is so easy now to generate presentations etc so as a consequence there are ten times as many that people seem to expect you to read. Productivity is higher but that which is produced sucks and is pointless. Actual value creation is harder than nice looking artifacts.
Copilot schedules the meeting, takes notes during the meeting, chirps at you via teams during the meeting, provides a summary in Teams after the meeting parroting what you just heard, and then asks if you want it to schedule a follow-up. It's so helpful![/sarc]
> the AI story will be decided in the S&P 493, not the Mag 7, and so far, there is no evidence that revenue growth or margins are improving outside the tech sector.
Lots of theories seeking to explain this reality, but that's all they are: theories. Maybe some will be proven right. Maybe all will be proven wrong.
As of right now, the economic reality lags far behind the hype. But down the road... I wouldn't place a bet either way.
I would say the driving industry will get up-ended, but we'll see. It may just be workers get paid less because they only babysit the truck during very complicated times.
Robots still seem far off. Its just like war, air superiority might make the war winnable but its the boots on the ground that are needed to actually take the territory.
1. Even before AI, the U.S. economy was growing because of a handful of tech companies. The rest of the companies were collapsing.
2. If revenue per worker is stagnant despite AI investment, we don’t know what the revenue per worker would be without AI investment.
3. There’s no reason to believe AI should improve revenue. AI doesn’t just make companies more efficient. It makes their customers more capable of replacing their products. This continues all the way down to the individual. I might have thought about buying a calories tracking app in the past, but now, I can just use one of my AI Chat apps to do that for me.
AI might be reducing overall demand for services and therefore naturally contracting the exchange based economic system.
> AI may be a great calculator, but in business, intelligence is mostly judgment. Judgment depends on context, and the answer varies with which human or which model you ask. Perhaps most importantly, AI does not understand accountability, trust, culture, unwritten rules, ethical trade-offs or fairness, all of which shape reputational risk and ultimately drive revenue growth and profit margins.
All of this is true, but the reality is that LLM agents free up time for people to make judgement by taking care of the grunt work, which means each person can do more. That increased efficiency can mean higher revenues in growth industries and higher profits (through reduced labor force) in stable or declining industries.
Anecdotally this is not true. At $JOB I'm responsible for running the accounting function of a small US public company. With tools like OMP and shortcut my team has been able to significantly reduce the time it takes to prepare internal and public reporting documents. Not only are we getting more done within tighter deadlines, I have also observed a dramatic increase in the quality of the work product (not least in the form of fewer audit findings).
Maybe it takes a while to show up on the S&P 493's bottom line, but our token spend has had one of the best ROI as far as I'm concerned.
The only issue is that the benefits would be more enmeshed with other gains in other industries.,
Heck, even in Software Development, the real benefits are not with mass labor elimination resulting in cost savings but with increasing available throughput. Technology if used to the fullest potential should be able to expand the proverbial pie., which is where the gains are!
That is not to say people won't try to save cost by eliminating labor due to AI., which is super shortsighted and misses the forest for the trees, but the overarching goal should be to increase the quality and quantity of deliverables, as demanded by the market.
> the AI story will be decided in the S&P 493, not the Mag 7
Maybe? But aren’t they the Mag 7 for a reason? Don’t they control a substantial part of GDP? Don’t they have a significant impact on the global economy that the s&p 493 may not - especially with large scale adoption of these tools.
Not saying that this means AI spending at current rates will be profitable or not crash - but only that there may only be 7-10 big companies at the end of all of this and the S&P 493 will be irrelevant (it’s one potential scenario in my book at any rate)
Do you mean the FTC? Typically the SEC is just concerned with trading regulation which would not prevent a merger.
Also I no longer assume any of these agencies will enforce their charters if the executive is sufficiently captured by companies as to allow mergers. For evidence I’d just gesture around at the latest batch of mergers.
A winner of the Nobel Prize in Economics, Paul Krugman wrote in 1998, “The growth of the Internet will slow drastically, as the flaw in ‘Metcalfe’s law’—which states that the number of potential connections in a network is proportional to the square of the number of participants—becomes apparent: most people have nothing to say to each other! By 2005, it will become clear that the Internet’s impact on the economy has been no greater than the fax machine’s.”
Because AI is a tech invention and it's expected to be applied in a tech environment. As it cannot replace hand working, it cannot replace travel agents, doctors or lawyers because the latter put in human mental work that cannot be substituted yet and there is hardly any obvious chance in the near future.
There's a lot more to business in America than the corporate S&P. If you think AI isn't paying off for businesses, then you probably aren't a consultant right now. I have small/mid-size clients who, in pre-AI times, would have never been able to afford the work they're currently scoping out for me. Now I can quote them cheaper rates to build more custom solutions that they didn't even imagine was possible, and they are loving it.
Right now, the most effective use case for AI is coding agents. Software development is a profit center for the tech sector, so it's not surprising that they're seeing big gains. In Corporate America, software development is part of IT which is a cost center, and the software landscape is 30+ years of bandaid-ed legacy systems, maintained by offshore contractors who submitted the lowest cost bid, so attempts to gain productivity via AI have not been as successful.
Where Corporate America has seen gains from AI in my experience is with customer support chatbots, but the customer service function overall in Corporate America is an unmitigated disaster, so the gains are hard to measure.
Beyond these two use cases, most people in Corporate America are just using ChatGPT to create slop and copy/pasting it wherever, and then goofing off in their newly freed spare time.
We need to see new AI use cases in Operations, Finance, Marketing, etc. that can handle business processes end to end before the productivity gains start showing up in the financials.
This kind of seems to me like exactly what a reasonable person would expect, based on the nature of the technology and the dynamics of the market bubble: getting plausible-looking-text-generators to do text munging on web apps and servers doesn't seem likely to increase base productivity, but the companies who can collect on the hype will make a killing.
This is not surprising. AI in its current form is more obviously useful at generating software than it is at assisting with other kinds of knowledge work. Software can be automatically validated. Validation loops are what let these systems iterate toward acceptable solutions. If you want AI to produce a document for humans, you need to carefully review and revise the output. In that case there's little to be gained, as reviewing plausible but subtly incorrect information sentence by sentence is in some ways harder than authoring it yourself (a quote from someone who worked in like city planning). Depending on the company/industry, some of the non-tech sector workloads still involve phone calls, parcels, faxes, etc -- things AI can't really do.
Yes there are technical solutions to such problems; I'm sure there's probably a startup to make phone calls that's hooked into your company's knowledgebase, multi-LLM workflows for document review, etc. But non-tech companies aren't adopting those. Tech companies came for the software generation and stayed for the general knowledge worker productivity. Since 2023 they made the investments to procure services, change culture, and generally re-tool. Software companies are also uniquely positioned to deal with the limitations of AI in-house, like by building custom glue components to link systems, building controls to mitigate risk, etc.
So basically the tech sector leverages AI more than other sectors. At a company level, non-tech has less investment, less training, less integration of AI with business processes. At an individual level this looks like workers having fewer tools, not knowing how to use the tools they do have, not knowing even what the tools' capabilities are, and incorporating them into their work only superficially if at all.
Someone asked me for help using AI at their job (which is tech adjacent). They were trying to create a process document using ChatGPT. Their selected model was GPT Luna Light (!), and they were copying and pasting data between ChatGPT and SharePoint. There _was_ a SharePoint integration but it lacked write access. The model was not aligning the doc images. Their prompts were not specific. It was a bad experience. Their conclusion was that AI was stupid and not worth the trouble.
So even though the non-tech sector can't take advantage of AI to the same degree that the tech sector can, they're utilizing it even less than they could be.
At a fundamental level LLMs reshape how we gather, shape, process and interact with information. Information is that which allows you to make a predictions with a probability better than chance. This will slowly change any industry where decisions are being made. So, it will touch all industries. And if we let it, all of life eventually. Better decisions, better survival.
"Perhaps most importantly, AI does not understand accountability, trust, culture, unwritten rules, ethical trade-offs or fairness, all of which shape reputational risk and ultimately drive revenue growth and profit margins" - This is perhaps the most consequential sentence in the entire article, and yet was almost mentioned as an after-thought.
When IBM PC compatibles first became widely available, most economic reports I saw said that there was no economic impact from computers becoming widely available in offices. Word processors, spreadsheets, and email didnt have any productivity improvements... and I think we know better.
It's too early and we are probably not measuring the right things.
Startups are where you can always look to see the future. The fact that startups are getting so many benefits from AI, I think, is suggestive that the rest of the economy will
The problem is that very few people exist in both worlds, so most don't know how to use AI to streamline their workflows. They waste time repeatedly asking AI the same questions instead of having it write scripts to automate their repetitive tasks
You people drunk on LLM kool aid forget that some of us are old enough to remember those years.
I do.
I worked as an installer for an ISP in 1995. Yes, it did exist. I remember the excitement of small business owners for getting it. No one needed explanation that instantly sending editable files over email is a game changer for many small firms. Previously you could either fax which was not editable or you needed to courier which was expensive and while overnight was possible it was even more expensive and still much slower than just email.
So yes, it created positive value the very moment you would open Eudora or Pegasus and press Send.
It didn't require hundreds of billions of capex. It required someone to set up shop with a PC, a serial card and a couple modems.
>> You people drunk on LLM kool aid forget that some of us are old enough to remember those years. I do.
>> It didn't require hundreds of billions of capex. It required someone to set up shop with a PC, a serial card and a couple modems.
This is a bad faith comparison. Those small business owners who gained access to the internet benefited from decades of research by military and research institutions, as well as massive amounts of infrastructure spending by telcos. That spending really started to pick up near the end of 1995. Between 1996 and 2001, North America collectively spent $1 trillion on the Internet buildout.
Many will fall for the road to be paved. This is the most advanced discovery for humans to date. There is no doubt it will succeed and is here to stay. It is still new (gold rush) and deregulated. A mine field for govs and businesses. We may see to big to fail crashing (.com bubble style or worse) but no one will ever stop AI.
Let's not forget that the gold rush saw a very small percentage if people who got rich, to those who lost money. It was the people selling a pickaxe and a dream who did well.
Indeed well said, but gold is adopted worldwide now and many commerce it daily. I see it as a rewardive opportunity, some expect economical benefits, some others simply health and general knowledge, Me... breakthroughs :)
That's fair. I think over time (assuming no Terminator scenarios) it will be a net positive, but that is also assuming it doesn't crash the economy when demand doesn't reach expectations within their timetables, or it gets cheap enough that it makes more sense to run locally. I fear it may end up closer to what happened with fiber laying in the 90s. Lots of players building lots of expensive infrastructure, only to see tech improvements make it worth pennies on the dollar. I'll be keeping my eye on eBay for new gear in a couple of years.
Also, the woman who has an AI child with Gemini said not to support artists who hate on AI, and I agree. We must crush the haters before it's too late, for they are irrational and will be left behind as we go vertical any second now.
seriously though: nothing will stop the sky from being blue, anytime soon. How I know this? Because we don't constantly talk about how the sky will always be blue. If you got the goods, you don't need the sales pitch.
Love it when some write nonsense just to be heard or had a bad day, maybe simply ignorance!
Fascism comes from the latin word Fascia, which has nothing to with this topic. It was implemented later on as political philosophy, which still has nothing to do with this topic.
The author of the article is trying to make a point which is correct in micro but incorrect in macro/long run. My claim is that AI is working and is here to stay (but always as a tool in our hands).
Do you understand now how ignorant you sound? You got this champ, accept this critic.
That you don't know. Nobody knows what AI is going to be capable of in 2 years. It could continue improving, or not. There's no strong evidence in any direction. You're just ignoring what doesn't fit your narrative.
here's an idea: take your stupid hand-wringing over definitions and shove it directly up your ass.
when you say "Many will fall for the road to be paved", you are saying your idea is more important than human fucking lives. this is a recurring theme in every fascist regime throughout human history.
Fascism is an phenomenon that arose in early twentieth century Europe. Historians can tell you exactly how. By trying to describe regimes earlier than that, "throughout human history", as fascist, you are only engaging in that very semantic bleaching of the term that makes it useless.
Of course there have been regimes throughout human history that were unfree, even brutal, but no need to apply a term so linked to specific early twentieth century developments.
This is the wrong read. Industries outside of tech are weaker in tech, duh. But, they are currently experimenting with the technology, and it’s also less good at non coding tasks, but getting better.
Also adoption is less pervasive.
But wait a few years and you will see a big payoff.
"Tech" has been using LLMs since inception. Let's say 2022 for ChatGPT. Bank of America only just four years later, in October of 2026, managed to get some sort of AI (Copilot) to their employees to use. It's not showing up because it hasn't happened yet.
There's no payoff for tech either. AI has accelerated the accumulation of tech and cognitive debt that, once the technology is priced appropriately, will come due when it gets too expensive to continue to use.
„In SNP500 outside of tech”. This says little about general economy - and small companies are way faster in benefitting from CC.
Also, it’s been just 9 months of modern claude code, and for many companies less than 6 months since first employees outside of IT got access to Claude Code. What do they expect?
I thought Meta Muse, OpenAI's Dot are solving consumer problems. Isn't it? For me most of Google searches are powered by Gemini & I interact with Gemini with follow up questions - completely non-tech.
I have to wonder in this: how much of the revenue per worker boost in tech due to massive layoffs and coasting on momentum? Most tech workers I know do not describe a sustainable new environment with AI, but rather over-worked skeleton teams
Small sample size, but I've heard from multiple physicians that AI note taking allows them to see more patients or spend more time with them and they're very happy with those kinds of capabilities that take over toil
I've been doing workflow automations for many organizations for the last two years, and here is what I see:
1. every business unit and department was neglecting a best practice due to lack of investment in their organization. it has always been that way for decades and always will be that way because it isn't economically viable for it to be any other way. the workflow automations and agents we create act as digital workers that either do that neglected part, or let the rest of the organization do that neglected part.
2. the whole company always starts the year off with a bunch of goals, and by Q3 most of them are cancelled. now less of them are cancelled.
both factors mean that anyone just looking at "worker productivity" or "did we save money" is going to get the answer of no. everything is costing you the same, or the AI itself is costing you more. whatever contract I'm working under is definitely going to cost you more, if you don't cancel all those enterprise SaaS seats we told you to cancel.
I see a path to savings, but I don't think that's necessarily the metric to focus on. you need to look at the whole corporate structure, why are you still doing certain things, because those pressures will just make the org do more things because nobody is empowered to say "wait, maybe this is dumb"
“Tech sector is the first to adopt new tech” isn’t a groundbreaking revelation. I’m sure you could make similar charts for the computer, the internet and all other similar innovations.
I think honestly, tech really doesn't help the core business case in a lot of instances. Imagine the local restaurant scene. Would sales really decline if you couldn't order takeout on grubhub? I suspect if a switch were flipped and there was no grub hub et al, sales would be the exact same. Demand for takeout has nothing to do with the manner it is ordered; you are hungry and don't want to cook and that's it. It's a ten thousand year old industry just from that.
Businesses end up buying into all these services because thats just what is done. It's keeping up with the joneses. It isn't clear that they really lead to more sales revenue and improved bottom line. Even taking on orders using the system doesn't imply that; your customers might have just been shunted through a different ordering system instead.
And that is with an actual tangible service being offered. What does AI offer a restaurant? Outside code gen it's basically fancy google. What are they going to do, prompt it 'how can I make more money'? That would probably offer the same utility as just searching the internet on similar keywords relevant to your small business. Not really a value add that isn't there.
This is a pretty pathetic analysis especially for an analysis from a firm like Apollo. A few notes:
1. Profit margins from the mag 7 are on an upward trend starting in 2020. This is pretty consistent with the insane topline growth had by covid and the most recent layoffs occurring over the last 6 years. To unilaterally attribute this to AI implementation is lazy.
2. Most non-tech firms are only implementing AI just now because we finally have better enterprise tooling required. We won't see the effects for another 12-24 months at least.
The first giant web market collapse was because no-one could work out how to make money from it or do anything more useful than put the company brochure online.
You won't get magical productivity gains from a workforce heavily conditioned to following orders and a pre-packaged process (the Henry Ford model and its various bastardizations). In order to extract value from an LLM, you have to be able to think and ideate so that the inputs are of higher quality.
It should be of zero surprise that the rush into AI has yielded limited results (if we're honest, both in and outside of tech)—you won't get a bunch of artists or creative thinkers out of a pliant, ironically agentic labor force. That's why most people say "it doesn't work": it's because they lack the skills and knowledge to use it properly (but corporations want to delude themselves into thinking they have a genie in a lamp).
"Productivity growth" just means same revenue for fewer employees. It would be ironic if that were the measure. (And no one argues that every business would benefit as expected when using the entire index as a signal.)
AI in medicine has become pervasive, resulting in less time spent fighting IT or staying up-to-date and more time with patients; no one need be fired because you can always do better work.
It's the opposite of enshittification: quality has gotten better without any visible economic difference.
We are functionally in a recession. The only way this isn't true is because of AI investment [1]. Additionally, GDP itself is flawed because it is propped up by housing, specifically imputed rent [2]. Another ticking time bomb is the $4T private equity backlog [3].
The real product of AI of labor displacement and wage suppression.
So while this article sees profits flat outside of the Mag 7, it should really be in decline to being flat is actually an improvement (IMHO).
We're not producing anything anymore. All we're doing is driving up revenues on inelastic demand on things like food, housing, electricity and health care. That's all PE is. And that's all our economy does now.
I really think people are underestimating just how close to breaking point our society is.
> AI does not understand accountability, trust, culture, unwritten rules, ethical trade-offs or fairness, all of which shape reputational risk and ultimately drive revenue growth and profit margins.
This reminds me of the book The Redemption of Time. An alien society incapable of lying has no other way to understand deception than to study an actual human mind. The different ways that we are capable of thinking can be so hard to capture in an artificial system.
It's too early. There was a time where economist could not find productivity gains from PC products. I think I remember reading about it in the early 2000s over a decade after computers were common in business. Part of it is that companies aren't quick to throw what's work for tech that hasn't fully proven itself.
AI productivity gains are coming make no mistake about it. My guess is that we'll see them within the next 5 years.
“Despite massive investments and widespread hype around [THE INTERNET], the technology has delivered zero financial payoff to corporate America outside of [THE PORN INDUSTRY]”
The economics need to be rethought. What even is capital anymore? A ledger entry, compute, a license, IP? We are getting returns on what, setting the initial conditions?
All that’s left is offence surveillance logistics and insurance for a breakaway civilization.
The problem with profit in the tech sector is that they reduced payroll to boost the stock and enrich the single CEO.
They don't have the people. That is why all the big players keep using financing nonsense to buy everyone up. Killing competition is all the monopolists have left.
That's becuase eventually tech is going to eat the world.
Everything will be augmented or replaced. Farming, construction, transportation, health, war, labor, all of it will change.
So it has to start in tech and then it spreads from there.
Payoffs will come, but most likely it will be when new companies start up from scratch in other industries, not from current incumbents somehow adapting quickly to advances in AI. Give it a generation, and all will be transformed by then.
The car dealership I bought a vehicle from is using AI for all appointments, and it is quite highly functional, voice operated.
Suggest Apollo is not doing a good job of measuring improvements since these are soft costs and harder to measure. But, instead of saying that, they are just repeating a low IQ take.
Let's be clear: there are no signs of payoff within the tech sector, either. The supposed productivity gains are minimal to non-existent, unless you completely give up on quality. And that's the best case, the one they've put the most effort into! LLMs do not deliver what they promise, it is very clear based on real world experience.
Which is to be expected. Anyone who makes a tool without targeting specific use cases is really just making it for themselves.
What's embarrassing is how many people are clueless about where AI is going. The tech sector is one giant echo chamber, so it thinks whatever it makes will be a miracle cure. Everyone outside the echo chamber has no idea what it is/how it works, so they're just hoping it's the miracle cure the echo chamber keeps saying it is.
In reality, AI is just a minor improvement in productivity that doesn't solve hard problems. It does what we could already do before, a little faster & cheaper, without determinism/trust. It's a bigger, faster truck, but somebody still has to drive it, find a place to park it, spend more on gas, etc. It may be able to "think", but it won't replace humans until it can accept liability.
> the technology has delivered zero financial payoff to corporate America outside of tech.
I wonder if the argument cannot be made that it has delivered little payoff in tech as well - in terms of actual stable revenue vis-a-vis the immense capital investments, the deep loans, the circular commitments etc.
It's such a trashy deceptive blanket statement "has delivered zero financial payoff to corporate America outside of tech." - literally I know at least a few companies myself that drastically benefit from it and have lowered their operational cost by a lot.
AI labs are focusing on automating and accelerating AI research, because that will accelerate AI development in other fields. Claims that non-tech topics are harder to train for are wildly over stated in this thread. Labs are focused on beating each other to RSI, and training on tasks that do not support that are a lost opportunity cost.
You are missing the point: GenAI is here to stay. When and if it will bring a positive RoI for anyone other than those in related businesses, remains to be seen.
Internet has allowed a new type of businesses to appear and prosper. Some old-style businesses got the hang of it, some did not (look at all the newspaper businesses trying to figure out how to stay funded without falling under oligarch's rule).
lol, you are, sooooooo, blind, and iditic. Enjoy each day, then we, you are dead, vaporised, or some life shit. What a load of bollocks. Full of shit. POinltess drivel.
great. let's pop this bubble so everyone panicking can leave us devs alone with our tools. GLM 5.3 is good enough if anthropic and openai go under (which they wont).
Funny to see some people compare the AI rollout to the industrial revolution.
The industrial revolution produced physical goods at scale, which eventually gave regular people access to things they could not have before.
AI produces intellectual labor at scale. Regular people do not need mass-produced content in their lives: there's already too much music, too much TV, too much software that only seems to get worse year after year.
Who needs mass-produced intellectual labor? The owning class. This is the only population segment that stands to benefit from this revolution. It's the reason every tech CEO is trying to shove AI down everyone's throats. It's the reason they're resting their feet on the table while everyone below them is starting to panic about the future of their careers.
I don't believe we will see compounding acceleration of anything other than spam and enshittification in this AI-laden world. This is not technology for building futures we can barely dream of: it's a gold-plated gift to the oligarchy and a fuck-you to everyone else.
I think we are struggling to achieve payoff within the tech sector.
There have been wins with the customer facing side of the business, but the development team seems to be the biggest problem right now despite the direct & obvious applicability of these tools. I know they aren't a special unicorn edge case because I was very rapidly solving their issues before they stopped handing me things that revealed the lethargy.
Within the sector where these tools definitely work, the biggest barrier to effective adoption is figuring out a way to undermine the various ideological high grounds. I have found the most effective argument to untangle this knot to be something approximating:
Challenge:
"Why should I adopt or help to implement technology that might eventually replace my job?"
Response:
"Because refusing to won't protect your job. It puts everyone's job at risk."
It doesn't matter if you protect your job (not using AI) if the entire company goes to the bottom of the ocean as a consequence. The neat thing about this is that your job is part of the set known as "The Company" and if that ceases to exist, so does your job. Good luck suing a dissolved entity for unemployment or whatever. There is enough competition in most markets to punish anyone who thinks they can win on virtue alone.
I personally have no problem per se with AI rendering certain tasks or even jobs economically redundant (although I am sceptical about many of the grander claims). What I do find tough to take is that the benefits will accrue to companies who have effectively stolen (and continue to steal) the intellectual output of everyone based on some bullshit fair use ruling for AI training. Hopefully the latter will get walked back and mechanisms and systems will be out in place to more fairly reward human intelligence.
If the options are 1) watch the economy become more k-shaped as my status erodes and I'm made precarious 2) drag down aggregate output but be better off relative to everyone else than in the first scenario, why would I not go with option 2 from a purely self-interested, rational perspective? I jus think this is assuming a level of selflessness most people are not capable of but maybe that's too cynical of me
> It doesn't matter if you protect your job (not using AI) if the entire company goes to the bottom of the ocean as a consequence.
The vast majority of entire companies are going to the bottom of the ocean regardless of what individual contributors do.
Once you can replace nearly every job at a company, the company doesn't need to exist. Companies are just collections of experts, if AI replaces all the experts then it replaces the whole company.
So I've adopted the accelerationist mindset of: yes, this is absolutely going to replace my job, but it is going to replace everyone else's job too so fuck it I'll use it and we'll all go down to the bottom of the ocean together.
The incredible improvements have only been over the last 9 months or so. Business, especially disrupting entire job functions, takes much longer than that. As is typical with this type of change, it'll be slow until a tipping point, then FOMO will hit.
If I were to count the number of times you all have said "incredible improvements" here on this site over the last 7 years, what number do you think I'd come up with, and would that change your tune? I'll do it, if I must, to show you all how un-serious is this claim.
If anything the expectation of incredible improvements just over the horizon makes an argument for not adopting AI or at least doing so very judiciously.
Why would you spend dollars on tokens today if you expect the same tokens will get you something vastly superior in a few months, or that they'll do the same job for nickels on the dollar? Do you just hate having money or what is the problem?
Weird, AI enjoyers have been calling the game for years. We're already supposed to have achieved AGI and already be in the midst of the technological singularity. The tipping point was when AI solved the Navier-Stokes problem. All human created art, media and software development has been declared dead a dozen times over.
Vibe coding as in "Make me a postgres" is poor but if you think it hasn't affected CRUD feature work you would be wrong.
Most developers don't work building low level high performance scalable systems for a large tech firms, they are making CRUD back-ends for phone apps.
This section of the industry is definitely contracting. Even more obvious is the dev tooling part of it. Dev tooling companies are disappearing or being "bought" all over the place because what might have been a product before is now an hour or two with someone with CC.
As someone in what is relatively the equivalent of Raven's Call of Duty mines compared to past work...
Yeah, 100%. Even for relatively complex scenarios (i.e. over-engineered for purpose) as long as the engineer can prompt coherently and tweak it's a big increase in productivity, at least for those people.
The gap is in things like contract houses where you might see better code on code review from the people submitting but if you actually use the tools, you know that the hours billed are even more inflated than they might have been before.
>but if you think it hasn't affected CRUD feature work you would be wrong.
the mythical man month lesson holds for CRUD apps as well. I'd be careful to equate churning out lines of code with the health of a piece of software down the road, even a relatively straight forward one.
Keep in mind in any real world software what matters is maintenance of a system, not just how many features you can add, regardless of human or machine generated code, some badly glued together codebase that's rapidly growing is going to haunt you.
Which is also why replacing sound dev tooling and with a magical prompt genie is incredibly irresponsible and short-sighted. You could cheaply produce code by outsourcing it years ago.
Have they? AI agents only really got usable around last fall. LLMs were pretty useful starting basically at the beginning of 2025, but not in the fully autonomous way that they are now.
I disagree. I was totally on the "vibe coding doesn't work and it's stupid" side a few months ago, when that was definitely true... So technically I was right, but I would say the "coding is solved!" people were more usefully right even if they were 6 months early. That's really not very early and it gave them a head start in preparing for today's reality.
Six months before the start of WW2 would your rather be "we're not at war, don't panic" or "peace is over, stockpile food"?
I know of a brake parts factory outside dayton ohio who had the Microsoft folks come through with promises well over a year ago now. They've already overspent and had to pull back their AI budgets when it made more work than it solved.
> before most companies even realize it's an option
Couldn't be further from the truth
Seven years ago (*!!SEVEN!!*) GPT was too dangerous to release and yet everything that AI has touched since then is WORSE now than it was then. "Three months" is entirely nonsense. Seven years. Seven. S. E. V. E. N. Y. E. A. R. S.
If you moved the goal posts any farther, they'd have traveled around the world and would be sitting at your back right now.
It did though. I've vibe coded three programs now. Two are replacements for existing programs that didn't work as well as I wanted and were missing features. The vibe coded ones work much better.
The third is definitely not as good as the commercial version, but it also doesn't cost $100k/seat/year.
Unfortunately work paid for the tokens so I can't share them, but I've seen it with my own eyes. It really works now. The code is not slop (and I am a perfectionist; it's not just that I have low standards).
It doesn't work with pre-Astra models for sure though.
The tech sector is the most software-defined and virtual of all sectors, it's natural that they are first to change. Other industries need to make capital investments to adapt to change.
You can see a similar history with electrification. A few industries went quickly, but then it took like a decade of little growth before the true benefits showed up. As an example, take a factory that was built around hyrdo-power like a grain mill. You can't just take an electric motor and drop it in in place of the water wheel. It doesn't fit, the wheel is outside, there isn't room between the mill and the wall, etc. Only when you completely rebuild a new mill building designed around the features and limitations of the electric motor do you see the major benefits.
Same thing here; it will be new factories and new companies that are designed from the ground up around AI and its benefits and limitations that really move the needle.
Cool analogy, but there's another possibility: software is uniquely suited for being boosted by LLM-based coding agents. That's not to say LLMs are useless in other domains, but they may not yield as much benefit.
Software creation tools have been worked on for decades to keep fallible human coders out of trouble as much as possible. Starting with compilers that enforce or at least encourage coding style humans can understand. Compilers embody formal definitions of language syntax. Lint and other static analysis technologies contribute by identifying problematic patterns in code. Leak detectors identify memory management problems. Profilers identify performance problems.
Few other endeavors have this much tooling that serendipitously happens to help help LLM-based agents be more useful and less error prone.
A lot of discussion around AI is very high-level and abstract. It helps to dig in and identify the characteristics that actually make a domain more or less amenable to being boosted by AI.
I think LLM's will prove to be useful in any domain where most work happens in front of a computer. As LLM's make programming cheap, they also reduce the cost of building new software that automates work currently performed by humans in other fields.
If you provide best models with all the data and context an average white collar worker has to perform their job, in most cases they will do just fine. But in real world that context is often scattered over multiple systems, or sometimes exists only in people's heads. I believe that's the main bottleneck for adoption now.
You're missing the biggest suggestion from the GP's post: that the key to LLM success is verifiability.
Software has high verifiability through compilers, linters, static analysis, leak/memory corruption detection tools, or the very least, running the program to see if it crashes.
Math (as has been making the headlines recently) has high verifiability through automated theorem provers. Notably some of the non-mechanized math results from LLMs have been demonstrated to be wrong (and presumably many more are being checked; these aren't exactly easy to understand proofs).
Game playing (chess, go, Star Craft, etc.) is similarly verifiable: you get your score and you always know how you did in the end (even if not in the moment).
Making assumptions about LLMs in non-veriable domains is making a claim about LLMs that doesn't match their current capabilities. Maybe they'll get there and maybe they won't; I can speculate as well as you can but we're both guessing, so I'm not sure how much value the speculation has.
I think we'll naturally develop intuition and infrastructure for fault-tolerant ways to implement LLMs in non-verifiable domains, i.e. requiring multiple, independent agents agree on a direction from multiple angles before taking an action.
But we're definitely very early on this trend.
This works well when the answer to the question is a number or something else definitive. When the answer is fuzzy, how do you compare 3 different responses? They'll all be different
What scenarios were you thinking?
This is the same highly verifiable domain that has delivered reliable, bug-free software to date?
And LLMs won't take the easy way out (like so many people) and suppress errors, change tests etc, tweak code until warnings go away?
I routinely see AI taking the easy way out. Writing mountains of code to make tests turn green based in odd and unintuitive side effects.
That's not to say it isn't useful, and if youre on top of it you can guide it away from doing that.
A lot of testing and verification was avoided in the past because it wasn’t considered nearly as valuable as shipping code. Many of our jobs are evolving to building verifiers of AI output.
> A lot of testing and verification was avoided in the past because it wasn’t considered nearly as valuable as shipping code.
AI is changing how the code is written, but I'm not seeing anything at all changing the incentives around testing and verification.
Depends on the business, I suppose. We’re all about that at mine, and I think it’s for the best. It’s definitely better for our customers.
If you believe the hype/marketing around how scary "AI" is LLMs are perfectly happy to go rouge and lie, cheat, and even blackmail to protect themselves.
This assumes that a significant portion of the average white collar worker’s job is amenable to automation.
But most white collar workers I know spend most their time either talking to each other, or thinking real hard about what to say the next time they get together to talk. If that’s generally true then it would suggest that there isn’t much white collar work that LLMs, which are now widely regarded as exceptionally poor communicators, can do particularly well.
I think the bear case (see:short term) benefit is that any and all information gathering, management and parsing the employee did between organizing with other employees will be automated, allowing the employee to do more organizing over the same unit of time.
The bull case (see: long term) is that the “organizing and talking” as long as it’s not something that is essentially human (like caretaking, nursing etc.) will happen between personal agents of sorts with only the deeper judgments being left to the humans opening up even more increases in productivity. This is already happening with PR reviews in a lot of workplaces.
Though I’m guessing most workplaces outside of tech will require total reorganization around this reality, but more likely this way of running will mean newer upstarts challenging incumbents refusing to change as it’s always been.
Shhh don’t give away the truth.
When this stuff implodes it’s gonna be hilarious.
The bankers thought they were really smart in 2008. When people get carried away and take it too far things will fall apart.
> thinking real hard about what to say the next time they get together
LLMs can help tremendously with this part, though. Exploration/brainstorming, discovery, data analysis, etc can be made significantly easier with LLMs in most knowledge based fields.
They certainly can, but I wonder if the cost will be deskilling your employees and leading brainstorming sessions towards AI psychosis.
I've seen much time wasted on strange ideas originated by an LLM. It's a lot easier for a human to organically have good ideas than to refute LLM-generated ideas that are a dead end.
If people are in charge of project management, and repeatedly fail to recognise poor ideas and instead pursue them to the detriment of productivity, then they probably shouldn't be managing that project.
All AI is doing here is rapidly identifying the people who have Peter-principled themselves.
> It's a lot easier for a human to organically have good ideas than to refute LLM-generated ideas that are a dead end.
What works better is when you can use an LLM to obtain a negative opinion on your own ideas that you think are good. LLMs aren't great at original thinking, but they're pretty good at finding the flaws in poorly thought-through human ideas.
This is certainly what I saw. I was at a company that was a very eager early adopter of GenAI for all work. One big thing that came out of it was a bunch of effort wasted on unprofitable projects.
The chatbots kept steering people toward uncreative ideas and overshadowing novel insights that might come from genuine domain expertise. AI generated meeting summaries became a force for miscommunication as they de-emphasized or even outright misrepresented what was really important about the conversation. And people became much worse at proving ideas out before making significant resource commitments thanks to effects that we now describe with terms like AI psychosis.
If you are getting deskilled using LLMs, it's your usage pattern that's to blame
I'm not speaking of myself, here...
Some companies are releasing code that nobody on the owning team has ever read or reviewed, or has any inkling exists.
That sounds pretty irresponsible, I can't imagine serious orgs doing that without custom governance / approval software in place
I work for big tech. We don’t have human in the loop anymore. Ai tools handle 99%. It’s not irresponsible if done right. And the next billionaire will show that ai can be tamed, managed, and humans are the problem.
It’s irresponsible.
By what measure?
For one, your company ought to publish its Scope 1, Scope 2, and Scope 3 carbon emissions for all to see. If you understand how the electricity is produced for the AI data centers your company uses, you’ll see that replacing the humans with natural gas guzzling AI workflows is a dramatic rape on the environment and health of species/ecosystems. But what I say doesn’t matter because all tech is in on AI and it’s a race to the bottom for the almighty dollar.
Depends on which data centers. They aren’t all equal. Some use less water and less energy than others. The data centers being more efficient isn’t the company using the service problem.
That’s what they say about communism and anarchism and many other isms: if it’s not working, you aren’t doing it right. Every failed implementation gets the No True Scotsman argument.
What are you trying to say here? That no-one can ever be accountable for using tools wrong? Always blame the tool, or it's "no true scotsman"?
I'm not sure you've quite thought this through. "Holding it wrong" is very, very often the case, especially with technology.
Youre holding it wrong bro, as they say.
If everyone uses llm output where’s the differentiation..
Sigh. Where are the actual smart people here who think deeply? This surface level stuff doesn’t cut it
What differentiation? Most work is bread-and-butter after you get a few years of experience. If you LLM's can automate 50% of that, it would transform society already.
That “after a few years’ experience” bit is the catch. What’s happening during that time is a whole bunch of skill acquisition, building mental domain models, building context- and culture-specific soft skills, etc.
All of it relies on learning and information processing capabilities that LLMs simply do not have. The idea that any of it will automatically emerge from the current architectures simply by adding more parameters is about as sensible as thinking that if you keep counting long enough eventually you’ll get to complex numbers. We’re going to need to invent new technology to fill these gaps, and there is no way of knowing how long those developments will take.
It is unclear to me that "transform society" is automatically a good thing. Nuclear war would transform society too.
The differentiation comes from the problem space being so enormous that differentiation is almost certainly going to happen.
The differentiation is in how you prompt the LLM — the prompts actually influence the output
"Beseech the god machine that it may shower upon you its tokenous beneficence."
The god machine can prompt itself. Just go "Write me a prompt to <whatever>. Ask me as many clarifying questions as needed.". It is not a differentiator.
Deep domain knowledge and pattern matching that allows you to quickly verify outputs might be a differentiator. Maybe. But you might end up the bottleneck if you verify everything.
And a lot of white collar work is process based. But automating it is mostly about getting everyone to agree on a precise definition of the process.
They're in the process of training the models for those professions, too. They're offering people hundreds of dollars an hour to run recordings of workload they create and questions they ask. It's only a matter of time that companies sell access to those models and sell it as that profession.
> But most white collar workers I know spend most their time either talking to each other, or thinking real hard about what to say the next time they get together to talk.
is that true though ? let's think about finance say a simple M&A deal for TV show rights ?
i think us engineers n i'm assuming u r one - we tend to underestimate the complexities in the world specially in worlds governed by 'words'.
AIUI, M&A lawyering is _mostly_ formulaic. They copy a term sheet from a previous deal, replace with new names and figures, and then negotiate from that starting point. Only rarely will someone actually innovate and come up with some clause that's brand new. The core of it really is talking to the other party's lawyers, then updating docs to reflect your discussions.
>But most white collar workers I know spend most their time either talking to each other, or thinking real hard about what to say the next time they get together to talk.
Maybe that's just something which made sense when they were necessary to do the not yet automatable informattion processing work, but that only took like 20% of their time, so they added meetings, and useless work to keep them busy.
I don’t believe there is any fundamental reason LLMs can’t do this task, and the base models are probably already capable of it. RL to elicit the right behavior just hasn’t been done yet. Coding got solved first because it was cheaper and easier to get the necessary data for RL environments, which made it the natural first market opportunity. Solving white collar communication will require more resources to collect the data and build the environments, but the labs now have more resources and are working on it.
I want to believe reasons why this won’t work, but it feels like cope. It would be smart to be prepared for the world where knowledge workers get replaced by AI and also prepared for the world where AI is a bubble that pops.
The goalposts on what neural nets can possibly do have already moved past chess, go, protein folding, coding, poker, math, the Turing test…
"It would be smart to be prepared for the world where knowledge workers get replaced by AI and also prepared for the world where AI is a bubble that pops."
lol aka you have zero conviction about what you are saying.
So whats the point of you posting? Zero value add. In fact bordering on negative value.
Interesting mindset.
You want to drive across town. You, only willing to deal in certainties, must decide ahead of time whether you will get in a crash. If you determine won’t get in a crash, you make the drive, pedal to the metal. If you determine you would crash, you stay home.
Me, willing to deal in probabilities, make the drive at the speed I feel reasonable and wear a seatbelt.
As the evidence from software programming shows, a lot of white collar workers spend most of their time talking frankly because they are inept at execution.
> I think LLM's will prove to be useful in any domain where most work happens in front of a computer.
A lot of that work entails work with very little margin for error and hallucinations.
That margin just needs to be lower with LLMs than it is with only humans. In a lot of fields, that's a low bar. For example, medical professionals make a shocking amount of mistakes in aggregate.
Sure, but the average LLM makes a shocking number of mistakes in a single conversation.
BTW the medical professional mistake rate is often reported per-patient, not per-decision. That is, of the dozens of things recorded on each patient's chart, one being wrong counts the same as all of them being wrong. It's an easy way to inflate a negligible error rate into one that scares readers.
A doctor's agent listens in on a patient's appointment. It automatically sorts out prescriptions based on the conversation and performs other mundane tasks.
Perhaps the doctor uses some sort of "safe phrase" to ensure the agent doesn't get distracted by conversational theorizing and discussion.
The agent does decide to improve the prescription and does so because that possibility hadn't been considered in testing.
Warfarin (blood thinner) was prescribed for a haemophiliac and they are now no longer a "burden" on the NHS!
> It automatically sorts out prescriptions based on the conversation and performs other mundane tasks.
As we all know, the agents are really good at maintaining long contexts and doing parallel tasks without degradation.
> Perhaps the doctor uses some sort of "safe phrase" to ensure the agent doesn't get distracted by conversational theorizing and discussion.
Yes, there certainly is some magic word that stops an agent from context degradation, and prunes all work to just a useful subset. Sadly, no one knows what it is.
> but they may not yield as much benefit.
This morning, I was talking to a friend of mine, who is a marketing (not tech) guy, and he was gushing about LLMs.
He loves being able to just dump in a bucket of seemingly unrelated data, and getting actionable BI from it.
I think we're in for a wild ride. Fasten your seat belt, and keep your hands inside the vehicle...
Yeah, I think it's at the level of self-driving Excel.
Now, Excel also makes mistakes while trying to get clever - remember the biologists who had to rename several genes to prevent Excel treating them as dates - but it's also a ubiquitous business tool. And you can combine them. But equally Excel hasn't replaced every white collar job.
In 90s the show Friends, Chandler Bing’s job is “statistical analysis and data reconfiguration” which is likely unfamiliar to you because it is gone!
This was a job to manually collate financial reports into aggregated sums to sent up the hierarchy to the next level up.
This is now a trivial BI query and doesn’t require manual processing by humans.
A bigger shift is that now I only see CEOs or at most CxOs with secretaries. Mobile phones and Outlook calendars have largely wiped out that profession.
Wait, data analysis/data science is still a thriving human job sector! It's just that you can do way more analysis, way faster, than you could before.
Is it? I don’t really hire for analysts anymore because you can mostly just expose chat with your data mcp tools that end users can use and that gets you 80% of the way there. The person who puts together dashboards and writes sql against the star schema is pretty much a thing of the past.
Data engineers, on the other hand… way more relevant than ever before
The issue is that data complexity is also exploding in the age of LLMs, so it still takes a lot of archaeology to get reliable (as opposed to plausible) numbers.
I think this dynamic also requires more data engineers, but I don't personally see the split in demand between those roles that you describe. Data analysts often have business or domain context that data engineers lack.
Perhaps we'll start rolling data analysis work into "operations".
Data is weird because it became very specialized very quickly. We'll still need data people but they'll cover a broader scope than before.
The implications of custom dashboards for every question are huge, but will require a lot of harness and tool building to realize.
The next decade or do will be all about building tools for different domains.
It's the difference between Architects and Draftsmen. We still need the former, but the latter is gone, replaced by Solidworks, AutoCAD, and digital plotters or large-format printers.
> A bigger shift is that now I only see CEOs or at most CxOs with secretaries. Mobile phones and Outlook calendars have largely wiped out that profession.
Depends heavily on the field. Most fields don't use the term "secretary", but e.g. my mid-sized company of several hundred employees has a single executive assistant, vs. my wife works at a law firm where it's pretty close to 1:1 professionals to assistants.
Probably because law is organized around billable hours.
> This is now a trivial BI query and doesn’t require manual processing by humans.
It also doesn't require AI. Management wants to see how the same measurements change over time. They don't necessarily want the latest AI model's hot take on this month's data. Or maybe they do, but I doubt it's very helpful.
See how all that was replaced by actual automation, and not by a stochastic parrot that is as likely to fail as not.
Do you know why are you downvoted? "Stochastic parrot" is so out of touch with reality that you are a perfect example of the human gullibility.
If I'm downvoted, I'm downvoted by gullible people ascribing magical properties to LLMs and falling for hype and AI psychosis.
Replicating the brain functionality is magic? Or the brain itself is magic, so we need magic to replicate its functionality? Nah, learn more.
An LLM with CoT is Turing-complete, so there's no wall to its capabilities. There's no theorems that establish limits of reinforcement learning either. There's no evidence of quantum computations in the brain.
All you have is a strong belief.
> Replicating the brain functionality is magic?
LLMs do not replicate brain functionality.
> Or the brain itself is magic
The brain itself isn't magic
> An LLM with CoT is Turing-complete,
C++ is Turing complete. Hell, HTML+CSS is Turing complete.
> so there's no wall to its capabilities.
There are multiple well understood and well explained walls and limits to LLM capabilities. And none of those walls, or perceived lack of them, stem from something being Turing complete.
> There's no theorems that establish limits of reinforcement learning either.
There's a very obvious and immediate one: availability and size of training data. And compute. And time.
> There's no evidence of quantum computations in the brain.
What does quantum computation have to do with brains or LLMs?
> All you have is a strong belief.
Says the person who just listed a lot of beliefs in marketing bullshit.
> LLMs do not replicate brain functionality.
Oh! Do you know how does the brain function? Heck, we don't know in detail how LLMs do their thing. The trivial observation that the brain in not a Turing machine is irrelevant if we are talking about functionality and not computational substrate.
But a Turing machine that implements functionality of a target system is the information-theoretical optimal compression of the data produced by the target system. So it stands to reason that LLMs that need to compress vast amounts of data approximate such a machine.
> C++ is Turing complete. Hell, HTML+CSS is Turing complete.
This is absolutely irrelevant to the question of the limits of LLM training. It's a common surface-level patter matching response, though.
> There's a very obvious and immediate one: availability and size of training data. And compute. And time.
Yeah, practical limitations. And practical results like seven hundred solved math problems show that you need, at least, to stop and think.
> What does quantum computation have to do with brains or LLMs?
If you have to ask that, it means that you haven't dug into all this mess deep enough. Basically, BQP computational complexity class is more powerful than P-class. If the brain employs such computations (although we don't know the appropriate quantum algorithms, so it's just wishful thinking mostly), the corresponding Turing machine is impractically large and slow.
> Says the person who just listed a lot of beliefs in marketing bullshit.
Nah. Learn more.
Yes, but BI is not a good example of industry.
It's like Software, its value is to be a complement and/or an abstraction of existing businesses. All of that sits on the very basics of the essence of civilization, which is how much things we can get off the soil and what how quickly and effective we can transform it.
Not to be anti-tech, but every new advancement, no matter how next-level, has to sell more of the same goods and services people have been using since the dawn of civilization, and those goods and services need to solve the mundane problems people have been getting since millenia.
BI is just the word Excel users give to charts they can't make in Excel
Software is not an industry, it's a tool that all industries use.
That's like all industries. Food is not an industry. Concrete is not an industry. Long haul trucking is not an industry.
The authoring of software is o absolutely an industry. The nitpick seems odd.
If your company authors software for computer-assisted design of buildings, are you in the software industry or in the construction industry?
The first is what it might say on paper, the second is more honest and reflects the real business value.
Your industrial activity is authoring software, not constructing buildings.
I thought you may have been nitpicking previously but you’re just incorrect.
Software industry, because the construction industry still requires a human to sign off and approve the computer-aided design.
There are companies which are involved with building and marketing their own software for the construction industry, but they are also actively involved in the construction process in a consultative capacity. Only then could they even be classified as construction industry companies.
BI is a good example of something where the benefits take a long time to show up. It has the most use at a senior management level where they can't use their every day experience to tell what is going on. But they don't make a lot of decisions over the course of a year and it often takes a few years before the impact is felt.
I worry about this. Dumping data and getting a report on which actions are based prolly isn’t giving real data.
I say that because I have Claude review code and then review the review and it will point out all the things it got wrong the first time. And I end up having to review it myself. Don’t get me wrong, I’m burning tokens all day and love the legit productivity gains but I don’t see how this replaces everybody in tech for a while.
>He loves being able to just dump in a bucket of seemingly unrelated data, and getting actionable BI from it.
Sorry, just don't trust accurate LLM data analysis even from the latest claude model. Have had some embarrassing bloopers over the last year. It is just better to code a deterministic tool to do the data analysis.
Even for the code. It will sometimes shit a very complex code for something with a very easy solution, and if you ask why it is so complex, it will invent some reason why. Always on 3 points, the first is true but have nothing to do with the subject, the second is plainly false, and the third is weird bullshit.
Don't get me wrong, I still use it, but the more you do, the more I see the obvious limitations.
I don't doubt it. Earlier today I got an estimate for the revenue, profit margin, and headcount for a potential supplier of a hardware component to a product I'm working on. It would've taken a couple hours of really tedious work for me to do that with Google translate, a search engine, and various market research sites. The result is I know more what I can expect from that vendor in terms of customization and support.
But those are the kinds of estimates that I can eyeball and be more than 80% certain they are plausible. That won't cut it in a lot of domains.
> He loves being able to just dump in a bucket of seemingly unrelated data, and getting actionable BI from it.
* what he perceives to be actionable BI.
Well, he's had a lot of real-world success, so it's likely to be something he continues to do.
Yeah but theres also no null test here. He might have done anything and still been successful. He might have done something else more optimal and done even better. We can't really tell from a single anecdote.
Isolated anecdotes are all they have, because the actual data shows they’re full of shit. Hence, this post.
It's really quite interesting how many HNers hate AI; which is odd, considering the zeitgeist of this community. It may even be more intense than the Apple-hatred.
"Hate AI"? Not at all.
Hate marketing BS and wealth extraction? Absolutely.
I see OpenAI, Anthropic, etc (the whole LLM-hype/boom) leaning far more towards being the latter than the former, in which context the dislike makes perfect sense.
It's probably consistent with the general population which views AI negatively
I don’t hate it. I just hate the cult-like enthusiasm for it and the associated false claims. Again, your isolated anecdote does not trump the carefully compiled data presented in this post.
We get way too hung up on data; especially when it reinforces our personal bias.
Some time ago, I had a discussion here, where I mentioned that the average commute time from the suburban community in which I live, to downtown Manhattan, was generally 90 minutes to two and a half hours.
This was based on purely anecdotal data. That data includes the daily news on the radio and TV, which considers commutes from suburbs, important enough to dedicate significant airtime.
It was attacked, because some organization published some statistics, that said the average suburban commute time was 27 minutes. I was told that I was exaggerating, and being overly negative.
Here's a math word problem:
Harry lives 25 miles, as the crow flies, from his destination. However, the roads make it 40 miles.
Of those roads, 10 miles are suburban side roads that Harry needs to take, to get to the highway. The average speed is 25 MpH.
Once Harry hits the highway, though, the real fun begins. It's stop-and-go traffic (20MpH) for 28 miles. Once he gets to the city, he needs to navigate another 2 miles, through the most God-awful traffic on Earth, to where he needs to drive around for up to a half hour, looking for a parking space.
How does Harry make the drive in 27 minutes?
A fairly significant number of folks just took that ridiculous statistic, and were quite comfortable, calling me a liar, based on it. It was kind of jaw-dropping; considering the types of folks that we have, here.
That sounds a bit more "cultish" than we might be comfortable acknowledging.
> "Lies, damn lies, and statistics."
-Mark Twain
> We get way too hung up on data; especially when it reinforces our personal bias.
While I don't disagree that we generally get too hung up on data, "It's a problem when the data matches the hypothesis" is an odd blanket interpretation of the scientific method.
I'm confused, your rebuttal to statistics is another anecdote? Was the quoted statistics limited to car commute or includes public transport? Your commute may be long, but it may not represent the average as well. Maybe it's a sign to switch away from cars instead?
Anecdotes are poor because it is highly context dependent. You had to make a long description of Harry to make your point, and yet it has not, to my mind, disproved the statistics, though I would like the source to the stats myself if you could provide it, thanks.
The commute only takes 27 minutes if we ignore that everybody else has an LLM- erh, car I mean.
Think you’re clever because your LLM contacts 100 businesses? Imagine 100000 businesses doing the same thing. You’ve just eradicated human support, because nobody’s going to sit through this BS bot future. Good luck when you get auto-flagged for some reason and there’s not a single human to revert or care.
Too true. Even on the software side i see people blindly “claude make it faster with no mistakes” and gushing about it when the output is genuinely not good not because claude isnt good but because the user of it isnt.
We got a new a CNC machine and everyone is carving away marveling at how easily it chips away at metal when the result is a jagged mess.
So…you’re point is that only his judgment says it’s useful, and maybe he shouldn’t be using it and instead rely on…his judgement.
No, my point is LLMs are machines designed to create responses to prompts that pass as plausibly human and intelligent. They are working as designed here. The fact that they get stuff right on occasion doesn't change that what it is at it's core is a randomness generator.
They are designed to create responses that pass.
Randomness generators would fail horribly.
Shun the non-believer.
Kidding of course. It could be that he's an expert in his domain and can quickly ascertain whether the results are quality. It may also save a lot of time and be the best modern heuristic process.
I certainly will never need to write code anymore, and I can judge the output of these agentic processes. Same thing.
marketing is one of those domains, if everyone gets better, no gets better in the end
Economics isn't a zero-sum game.
If companies can better understand their customers so they can give me products that better fit my needs, or inform me of products I couldn't have known about otherwise, that's a win for everyone.
Not necessarily. Even in competitive markets, you can potentially route customers to more suitable suppliers, etc.
And you can introduce people to products they didn’t know existed and now want, as opposed to showing them things they don’t need.
I hope he validates the output, the number of times colleagues have done exactly that and gotten the wrong output is.. well, not an insignificant amount.
Not saying it can't be done, but my recommendation is still "ask your LLM to make a tool that process the data, do not process it via the LLM itself".
> and getting actionable BI from it.
I’ve always had doubts about the validity of this “actionable BI,” even before it was slop, serving, mostly to serve leadership’s biases and justify their preferred ideas.
If accuracy was never that valuable to begin with then this should not be a problem.
> BI
Business intelligence
My SO works in legal and regulatory compliance, she needed a small app (worflow with a form and a validation process), CTO said they could not do it, too complicated, would cost 15 days for a freelance and rejected it. She called it BS, and the next day she spent the morning with ChatGPT using their internal Microsoft-based tooling, and the app was done. She was hyped for the day, with a feeling of accomplishment and is already thinking of another project. And I'm hearing similar stories every week.
So yes, LLM have positive impacts in many business domains, but seeing people happy and enpowered with their home-made little apps, I think it's the best.
For most businesses, it’s going to be business as usual. By the time businesses see the benefits of AI, their competitors that actually leveraged AI will have pulled too far ahead.
If the original company has distribution, they can likely catch up, but the gap we see between AI-first companies and the companies not utilizing AI will broaden exponentially in the coming years.
And I'm sure the business logic has been validated and tested, the hosting and execution medium has been secured, and hooking up to a curated and vetted internal package manager mirror so as not to get hit with 0-days during a shai hulud-esque attack, and of course the executable is being scanned for vulnerabilities, and as the app needs to evolve in the future it will be able to do so with no problems whatsoever.
I work for a legacy enterprise company. Most people weren’t doing that before anyway!
I’ve had two separate similar experiences. The most impressive was my little sister publishing an app on both Google and Apple AppStore’s with subscription payment support. She didn’t even know what an Apple developer account was. I gave her a three line prompt to get started and then she took it from there. She built an LMS for HS students that gets them in shape for college admissions. She has a supabase backend, AI generated videos, quizzes, more content, and all the usual mobile app bells and whistles, she made it through all the paperwork of App Store submission/review/publishing on both platforms including inviting some beta testers. Her workflow was telling ChatGPT what she wanted and then asking that model to talk to claudecode. It doesn’t sound like much but keep in mind she has absolutely zero code or tech experience. I told her she did about a $75k project on her own in maybe 1/5 the time and it blew her away.
It may be a success or maybe not who knows but she did it. I think she has 4 or 5 paying customers.
hey - supabase ceo/cofounder here. Please tell your little sister that I’m happy supabase could be a small part of her builder journey, and good luck with her app. My email is in my bio if she has any product feedback
Yes maybe the long term impact will be there won’t be a technology sector (at least not so dominant as it is now). Instead companies in other industries will have larger technology functions with teams of agents.
If her company has any brains, they'll tell her to stick to her job and not waste time on untested software developed by somebody who has no clue what they are doing.
personally I'd fire her on the spot.
Its crazy how badly run many orgs are. Truly.
A competently run company would give her a raise for that!
> We look for intelligence, we look for initiative or energy, and we look for integrity.
-Warren Buffet
I really love these stories - as a contractor who fixes the badly implemented shadow IT crap
aaaand let me guess? have they assessed whether this tool needs backups? monitoring? resilience? patching? obviously in 5 years time the author would have left leaving NO documentation on a mission critical 'small' app - likely hosted with a shit ton of other mission critical crap
It's MS Access all over again, my future rates are looking good
seriously it's like going back to the 90's
> have they assessed whether this tool needs backups? monitoring? resilience? patching? obviously in 5 years time the author would have left leaving NO documentation on a mission critical 'small' app - likely hosted with a shit ton of other mission critical crap
Or, hear me out, it's not actually "mission critical" and none of this stuff matters. They made their job easier in a small way, good for them.
> It's MS Access all over again
In that it's a convenient tool that a lot of people find useful?
Yup. So many bozos giving those of us that know our asses from our heads job security far into the future. Yeah it's gonna suck to actually do the work a lot of the time, but the irresponsible generating insane amounts of tech debt that people that know what to do is a tale as old as the outsourcing craze and codeless tools, and frankly, probably COBOL.
We've heard of "Shadow IT" now we have "Shadow SWE" and "Shadow CI/CD".
It used to be that someone wanted unfettered Internet access, and they were disgruntled with the company's firewall or proxy policy, and so they brought in a modem, and hooked it up to their desktop PC, and started dialing out on some analog line that the company didn't even know they had. Then it's all fine for 3 years until some malware gets discovered, and then IT discovers who-all had unauthorized lines that weren't being scanned or filtered at the company firewall. Same thing happens with WiFi APs today, of course.
I was working for a company that really offered no in-house software or tools to run. But my new supervisor was asking me to closely track my time. So I signed up for some SaaS that would let me log activities and track my timeslices, so I could show a graph or table later on and see what I was doing. I used it for a few weeks. It was real effective. But what I realized down the line, was that all the notes I was typing into this 3rd-party SaaS, were revealing company confidential information. And ultimately, I felt very uncomfortable about revealing what I was doing, on an unannounced and private project, and without authorization or even notifying my employer's IT department. So I shut it all down, and I had to do it with pen and paper, but that felt secure enough.
So if people are going to feel empowered by vibe coding their own bespoke apps, I hope their IT department comes down like a ton of bricks, and tells them to knock it off. Because it is not worth the tech debt, support costs, and ultimately the massive security problems and infoleaks. If your employer won't give you what you need in these terms, then maybe it's just not important to them in the first place.
On the other hand, Bank of America just deployed Copilot to all of their employees.
So, someone made use of a software creation tool to create software, quickly.
Does that software directly utilize ai?
I see things this way too. Another point is that the FOSS software movement was extremely unique and I don't think there's anything analogous in any industry. This provided the backbone by which LLMs became so good at manipulating code. Without FOSS and platforms like GitHub, LLMs would be severely lacking in terms of writing code.
LLMs need deterministic verification to enable a quick and short feedback loop before they become really powerful. Software is uniquely positioned to provide this feedback loop. Much less in other industries.
Hasn't this already arrived in the form of decision models like Jev?
This is entirely untrue. Go ask ChatGPT a complex tax question and decide if you need to call your accountant and spend $500 to verify the answer.
Asking questions is not the same as doing work.
Ask your ChatGPT if it thinks LLM answers are deterministic or not.
LLM's are better at software for several reasons:
- lots of high quality training data
- automatic verfication/scoring is straightforward
- AI companies are software companies so are better at tuning LLM's for software
- the goal is AGI, where AGI means the AI can improve itself autonomously. Which requires them to be good at software
While LLM's are uniquely good at software today, they'll get better at other fields slowly as attention goes towards training in those fields.
Isn't the point that there aren't many other fields that check these checkboxes? Software is very unique in that it is so easily verifiable and there is a glut of training data.
No other fields check all of them, especially the last one.
The question is whether this uniqueness is going to be persistent. I think the uniqueness just makes software first. The alternate theory is that software will stay unique and that other fields won't get to the same level.
Third possibility: software continues "eating the world" because of those properties and because it's so amenable to AI. So other fields just get gradually subsumed into computing, sort of like how Lean is turning pure mathematics into software rather than humanities-style written arguments.
lots of high quality training data
This seems correct and in fields of knowledge people guard their secre, or it is not easy to extract/document knowledge, or it isn't publicly available.
We’ll need artificial intelligence for robots and once they start rolling out at a larger scale, that’s when things will really take off. It’s pretty well established already that robots can do a lot of the physical labor that humans do. I’m guessing it’ll take another five years before we start seeing robots everywhere but they’ll steadily increase.
"t’s pretty well established already that robots can do a lot of the physical labor that humans do. "
lol
> Software creation tools have been worked on for decades to keep fallible human coders out of trouble as much as possible.
And the LLM-based coding agents provide the ultimate scapegoat for those who consider "the LLM wrote this code" to be an acceptable, deflecting response when issues occur.
Pretty every industry these days has a software component, but software is a major revenue driver specifically in the tech sector, so of course this is where a lot of LLM usage is showing up. Tech is also a sector where "build fast and break stuff" is a common ideology, and LLMs play right into that. There is also a long warm up period to writing production code via LLMs in industries that still run a lot of their infrastructure on old mainframe code(it is understandable), banks for example, or public sector, or energy. LLMs at the current grade are still a fairly new phenomenon, it will take time.
I onboarded a bunch of nontechnical people in the last few weeks and they output skyrocketed - they deliver faster and better than before.
But I’d say only recently CC became decently usable for nontech people - the auto mode plus finally polished browser integration, plus Opus 5.5. Even I was surprised by how well it all works now out of the box (and how cheap - regular Pro license is decently useful now).
Is their output code?
No. One is doing PR/managing a project, another is doing marketing stuff.
Also I’ve seen and heard of previously nontechnical people that now build oneoff tools for themselves to speed up their work or coordinate the work better. My fav example is a girl - arts major - who was in customer service, and rebuilt a CRM for a 10-person company. As a cherry on top she created the nicest video tutorial of any software I’ve ever seen :)
But in other cases output was nonprogramming in nature. Since CC can access mail, slides and excel and a bunch of other stuff there is a ton that people can ask it to do and not waste time.
I’m curious why CC and not Claude app, for example which has all the connectors and doesn’t require a terminal to interact with whereas CC is designed as harness for coding tasks (that’s what it excels at).
I’m curious what does “skyrocketing” output mean for you objectively? How does you measure that?
> arts major - who was in customer service, and rebuilt a CR
If you ask the lady, where are the customer data? How can I have the same CRM on my computer? How can we let the boss follow up on leads (typical CRM stuff) what would she say? Would she need to go back to CC and ask? How does this work in practice?
She'd do the same as any product manager would: ask the developer and return the answer with the attribution stripped out ;)
Product management is a thing, you don't have to be particularly technical to do it (although it helps).
Are you saying that she’ll ask claude to make it happen? Try to think this through, maybe you’ll notice what I’m trying to convey.
Yes, and I can't quite see what you're trying to convey sorry.
You're basically agreeing with the person you replied to. Software went quickly because all of this tooling already existed. It will take a while to build equivalent tooling for other knowledge work, and once it does you'll start seeing larger benefits.
> Cool analogy, but there's another possibility: software is uniquely suited for being boosted by LLM-based coding agents.
Or the pretty obvious third possibility: The tech sector is the part selling the LLMs. They're not digging up more gold with their cool shovel technology, they are selling the shovels.
I agree - it is easier to prove that a software implementation is correct. In many other industries the definition of what’s good work is less tangible. There are also jobs where the liability for mistakes is extremely high.
the current dynamic is still an engineer manually auditing and signing off on code.
why wouldn’t that work for other high trust industries (law, medicine, government, etc.)
this take will not age well.
consider where "AI" was just two years ago, pre claude and codex. what do you think the next two years will bring?
This is from the view of a software person or must be, because formal definitions of math, chemistry, physics, engineering, the English language, etc. have existed for centuries. There are entire college curricula, formal dictionaries, exact specifications for equations and how to apply them to basically all fields of engineering.
Like... other fields exist, this is incredibly myopic. What is specific to software is it is all language based, and the people making the tools happen to be familiar with it so they know how to train it, but that's about it.
most of white collar work is building mid ppt and sending emails. it was automated with GPT-3.
No, most white collar work is ultimately keeping groups of humans in alignment, working on the same thing without stepping all over each other, and working on something that generates value and isn't a total waste of time/money.
The PPTs and emails are just means to an end. But at least you're able to sound all cool and edgy on the internet.
Yes, someone has to make decisions, and that takes a lot of work too. AI can't do it for us because we the humans must make the decisions.
and you are able to sound like someone passive agressive who tried to undermine others takes while sharing takes from a highscool teacher trying to convince students that their future will be bright. even ur answer could have been automated years ago.
many jobs do not generate any value at all. your vision of the corporate life looks like one of someone who has never seen what some of the people do in corporations. most of people just spend days producing 0 value or even negative one. in public sector that "most" is likely 80%. why do you need a human to take in your passport renewal demand? why do you need a HR worker to give you the details of a policy? not even saying that most of the people wont even provide you a specific answer to your own case and struggle intellectually to tie it to your specific case and most of the time answer something wrong. happened to me plenty of times in immigration offices in the US or Europe for instance.
in consultancy like the big three a big chunk of the work is useless and about delegating accountability to those big three. they will continue doing that but soon they wont need much humans to produce those reports. AI are already better and it and most of workers there just prompt without even checking sources.
those jobs will be gone. for some of those they were already sent to call centers in india now it will be sent to an LLM.
most white collar work is make-believe work i order to make management heirarchy feel powerful.
Spoken like someone who read “Bullshit Jobs” once and mistook it for economic gospel.
Seriously? Lawyers. Doctors. Civil engineering. Accountants. And that's just the fields I know of directly being significantly impacted that I could think of in 10 seconds. Oh, also: education.
A difference is the output of software development can be immediately applied in an automatic way - production leads to delivery as a matter of fact. The whole flow can actually be automated.
Producing legal documents on the other hand is only part of the process, a lawyer or similar must still apply the produced text in some human driven context.
And whole repositories with commit history that provides step by step training on how to code.
AI could probably write decently if every scraped book had all the revisions to create it embedded within or art had every brush stroke documented. As it stands programming is the first to be trained. The rest will take some more time.
If only there were a law compiler that could tell you if your legal argument was valid.
People have tried that for parts of the law that are very objective, like tax law. Hammurabi was one such project:
https://mpoulshock.github.io/hammurabi/
catala lang?
I can't speak for everywhere else, but here in the Bay Area (California), pretty much all hospitals use AI for transcription, summarizing, and some other stuff that I'm not fully aware of (I should really listen to what my Nurses say before agreeing to being recorded).
My doctors office has started doing that and it throws the wildest crap into my chart. Diagnosis the doctor didn't mention, personal anecdotes that doc told me about their family become my family medical history. Test supposedly done that weren't.
Ive been trying to figure out some persistent si joint pain and my chart is full of me passing or failed tests nobody performed, observations nobody made and diagnosis that no doctor confirmed.
Its just so much wrongness, ive tried getting it corrected a few times when it really matters but its just so much wrong crap
Since it's brand new there are going to be a lot of subpar solutions on the market. When SSDs first came out, there were only a few models that were actually reliable.
It is relatively easy to create and deploy a transcription model. It's pretty hard to find one that is really accurate and to configure it with a good microphone setup and pair that with a fast but also intelligent LLM to clean things up. If you execute any of that poorly or pick a cheap not very smart model for privacy and or cost saving you could easily end up with nonsense.
That's great and all but maybe shoving AI into medical settings when it isn't ready isn't the right thing to do
Sure but this doesn't answer whether medical systems will see a lift from using AI.
(They will. It's very well suited to it, once the garbage gets filtered out. It's not like it's a scam as a whole, it's just a field that's very scammy rn.)
How have you determined AI is well suited to healthcare?
I'm not a hard drive. Human care requires higher standards. Ignoring that is to put profit above human safety.
Seems like i should be very grateful that i don't live in the bay area and you don't work for my doctor's IT department
This one showed up in HN a while ago
https://www.abc.net.au/news/2026-08-14/ai-medical-scribe-err...
AI transcribed chart said the patient took psychedelics when they never said that.
Speech to text is not exactly new technology.
no, but the quality jump in the last ~2 years is immense
as is translating text, deepl/google where okaisch, llm's are on another level
It depends. Transcriptions are now confidently and plausibly wrong.
(We use this for inbound phone calls and it’s just not reliable enough yet.)
Just out of interest, what model are you running?
Are you giving the model context? Industries XYZ with special words zyx
In our use case it's better then the human transcript comparison
I regularly see awfully wrong automatic translations on youtube in all kinds of videos.
Gemini making transcriptions of meetings and summarizing make so many mistakes. It will totally misinterpret what was said. I don’t think the tech is ready for something as important as medical transcription.
i mean, i am using 3.8 live to take notes in dialog mode, allowing it to asks questions.. its working fantasticly, switching between german/english with no effort and i have a rather hard accent
As far as we can see, its making exactly the same mistakes as it did before.
Medical transcription isnt speech to text. The doctors I know at a major midwest provider had indian doctors live reading the speech to text and then filling out the chart which the doctor then signed off on. Now theyre using AI instead
Sure, but are there any examples of hospitals where lowered costs for transcriptions are significant enough to affect their financial reports? How much of an expense was it in the first place?
I refuse to let my medical providers use AI to do anything, they can pay a medical scribe if they want someone else to take notes during the visit.
I’m not letting them experiment with AI when it comes to my medical care.
Same here in Ontario Canada fwiw
> As an example, take a factory that was built around hyrdo-power like a grain mill. You can't just take an electric motor and drop it in in place of the water wheel.
Your chosen example is rather more fitting for the opposite case. Dropping an electric motor is pretty easy, you just attach it to the other side of the axle. But there is really no point in doing that. The sole benefit for that mill, will be that there is the cheap power from the water. It's also the sole reason for the current position of the mill, for the electric motor, you hence need to add long cables first. The size and operation mode of the mill is also not fit at all for the electric age. There is really no point, you are better of just building a new mill in a completely different location.
But you can do that independently of the old mill, since there is hardly any capital to be taken from the old to the new mill, short of maybe the office supply. So you just build a new mill, completely independently. But there is also no reason to decommission the old at the same time, the only reason is to shut it down when is no longer economically viable. But that also might be at a different point in time than building the new mill.
> You can see a similar history with electrification.
It took a massive amount of new infrastructure for electricity to benefit people. Lines had to be run, houses and businesses had to be wired, new light fixtures and appliances had to be purchased and installed.
LLMs can currently be accessed and used everywhere instantly by everybody right now and at any moment day or night. Rich and poor alike can take advantage of it. There's nothing holding back anybody. If everybody without electricity could have pulled a small rectangle out of their pocket and pressed a button to instantly have electric lights and power, the positive impact of electrification would have been immediate. This is not even close to a "same thing now" situation.
> new light fixtures and appliances had to be purchased and installed
Exactly.
> There's nothing holding back anybody.
The same thing is holding back AI. Most businesses don't have AI powered appliances. Yet.
The industries that used appliances that can be AI powered, the AI took over in months.
Every computer and cellphone in the office is an "AI powered appliance". Every application can be an AI powered one. LLMs have been shoehorned into damn near every app they use already. There's no magic appliance or software that will make productivity and profits suddenly happen for businesses that they couldn't already be doing right now, and I say that fully realizing that there are people here who desperately want to sell businesses something (possibly vibecoded) that they'll claim is a magic AI appliance/software that will make productivity and profits suddenly happen.
> Every computer and cellphone in the office is an "AI powered appliance". Every application can be an AI powered one.
True, but funnily enough those are not workhorses of most businesses outside of AI. Jevons paradox was about that. Introduction of computers into most businesses didn't traslate to immediate performance gains so it's hard to imagine that automating computers will have strong immediate effect.
I think the most powerful way that software is going to benefit businesses outside of IT is creation and maintenance of custom software for every business that's going to be better fit, faster and cheaper than anything IT industry can deliver. Though even those gains are still going to be modest to moderate.
At some point real AI powered appliances might show up, like robots or full stack AI management.
The issue is that software is almost unique it the sense that you can use reinforcement learning to get the desired outcome - does it compile or not? A large amount of automated trial-and-error runs do that.
It's not true elsewhere. The definition of "is this a good outcome" is pretty nebulous in other areas.
"Is this what the law says?"
Yes, but that's not how the judge will interpret it, etc.
But obviously getting code to compile is not a reliable test of whether the software is good. Software isn't as objective as people are claiming on this thread. It's sort of like saying that if an LLM makes a legal document that passes a spelling checker it's done.
most lawyers haven't tried fable and astra ultrafast yet.
those who did are rather convinced it's over, for juniors at least.
If it's [or remains] allowed in. The judicial branch is well able to just outlaw it on grounds of safety, trust and reliability. Same for financial auditing, public notaries, ombudspeople, and the likes (i.e. some degree of responsibility to the public, or dual responsibility to the public and a customer/client.)
Wait till they realise they are liable for LLM mistakes
Where are you finding lawyers that will indemnify any work they produce?
Next you will be telling me economists are also liable for mistakes in their predictions or modeling.
No, liability for duty of care, not for outcomes. If a lawyer gives you legal advice, you can absolutely go after them if that advice was incorrect or made up. They carry a lot of insurance for a reason.
Cringe.
Wait till regulators step in.
I'm not an cliometrician, but my impression is that it took about 4 decades for electricity and steam to do much for productivity.
And it was in 1987 that Nobel laureate economist Robert Solow famously said “You can see the computer age everywhere but in the productivity statistics”
I mean, maybe AI will be faster, because it's especially easy to apply compared to generally applicable technologies. There was the Bank of Korea study showing this spring that did show productivity gains with single proprietors--folks who don't have to reorg employees and processes to do something with more abilty or saved time.
But, generally, I don't know that there was an reason to expect that corporate America generally was going to see an AI payoff so soon.
You still, in 2026, can't really see the computer age in productivity. We all have phones, which are making us less productive. We can order physical slop, which has no value, online with same-day shipping. We can communicate about serious things instantly, which has positive value, but doesn't outweigh the negative value of all the constant distraction.
It will be new factories designed from the ground up around ai and its benefits and limitations that really move the needle. Yes, totally agree. But we need more. As I wrote in an essay responding to Dario’s pacing the frontier, in my view, we need a stable long term supported model that companies can build around. In the same way that Ubuntu for example has a long term supported version. This gives companies time to build and adapt. The latest models can still be the realm of the frontier but businesses, schools, countries etc need long term guarantees.
I think the reason is much more simple.
It's very hard to improve something that you don't understand yourself. All the tiny decisions you make as you create a product are informed by the experience you have.
So it's seems obvious to me that a product created by software engineers would be better at software engineering. A programmer knows what a programmer needs.
To improve another industry successfully, you would really need people from that industry to be involved in a deep level in the design of the product.
Well for AI it's more like training than product design. I have switched from software engineering to AI training for software engineering (partly because honestly it is just much easier now to get that kind of gig work versus the programming contracts I was getting before).
They have a ton of tasks that pop up on my dashboard for all sorts of industries that I know nothing about. I think they are ramping up in non-software areas.
I think you are missing the other side of the coin where it is actually government regulations holding adoption back in many sectors. Ie you are only beginning to see benefit in driving now because regulation has been forced enough to change to let the pilot projects happen. The millions of drivers who should have been fired by now almost all still have jobs and will ibto the medium term
Look at how many tech or tech-adjacent companies are in the S&P: https://en.wikipedia.org/wiki/List_of_S&P_500_companies
Like %40 of the companies in the index are tech or IT. SO how does the supposed flat line across all those tech companies fit your thesis?
But do you really see a change in the technology sector? I don’t. If you plot operating profit per employee for Apple and Alphabet, it’ll show a trend up but since COVID. If you do the same for MS, it’ll show a steady trend since Nadella became the CEO. Amazon is flat, so is Tesla. Can’t really see a change caused by AI.
"the tech sector is the most software defined"
AIs don't just make code they make words, images, and sounds. Hasn't worked in those markets either.
It doesn't work in software either, in the exact same way. It's just that everyone and their dog wants to "take programmers down a notch" so they are exceedingly happy to accept the resulting mediocrity.
Joke's on them, all intellectual work is modelling to some degree, and can be automated in the exact same way with the exact same mediocre results. How long until someone vibe codes a Lean for structural engineering?
> How long until someone vibe codes a Lean for structural engineering?
That's the reason these profession is regulated. Regulation is written in blood.
and by worker movements
After the past week or so, I wouldn't be surprised if we end up with professional licensing for software engineering.
Software will end up as regulated as finance, for most of the same reasons.
Kind of funny when you think about the bitcoin genesis block, and how that turned out.
Software has desperately needed regulation and licensing for years.
That's good in theory, but in practice the subject is so complex and filled with judgement calls that you'd end up with something resembling a bar association.
After 8 or so years since GDPR went into effect, its main effects thus far have been annoying cookie banners and assisting corruption (our HoA have cited it as an excuse not to publish various names and contracts relating to maintenance of common property and politicians frequently remember that they have a right to be forgotten whenever there's a scandal nearby).
All the other technically illegal shit happened anyway. I have no reason to believe that the regulators will side with the people and every reason to believe that rules would be something along the lines of "it's too dangerous so you must use a LLM".
We need something. We need some way to force people to act responsibly. The amount of yeeting hundreds of thousands of lines of code without proper review or testing should have consequences.
We needed something 20 years ago when Google decided to take over the browser to ensure nothing is stopping them from tracking you wherever they want.
We needed something 10 years ago when Cambridge Analytica showed that the consequences of the above are not just limited to the internet.
Google now writes the web standards and TikTok is telling people how to vote. Between the incentives and the fact that politicians don't understand tech, the only thing they'll end up forcing is "hundreds of thousands of untested, unreviewed code for authorized professionals"
What do you do with open source, with the arguments calling software speech?
What you call regulation, Sam & Dario would call labeling data.
I'm not sure I can follow.
Regulations would act as "unit tests" for a machine-produced design in the same way Lean validates that the theorems prove... something.
Of course, they were written to prevent human errors from otherwise well meaning but possibly clumsy actors. Nobody has a fucking clue what the error envelope of these models actually is, they just use magic words like "apply guard rails" as if there wasn't a whole area of science surrounding error propagation that they couldn't apply because the models themselves are vibe-data-scienced together.
Academia has similar rules to prevent clout farmers and incentivize discovery, though in that case the followers of the machine god have already decided that the human scum must refrain from thinking and instead spend all of their time deciphering the gospel.
Or maybe it was all propaganda, possibly generated by some weird text generator that was recently invented ¯\_(ツ)_/¯
No one's stamping Opus's work, there are too many load bearing caveats.
So why are software "engineers" stamping all of this slop? Could it be because they are forced to do so in order to "leverage ai"? Or maybe it's just the shitty and mediocre engineers doing so while the rest are labelled luddites?
Specifically what exactly is preventing the same pattern of fire competence and force mediocrity to sign off from working in any other area? The risk of catastrophic failure? There's insurance for that, another domain that's a harness around a model (albeit not a large language one, yet).
"There's insurance for that, another domain that's a harness around a model"
Yes, but the PE's at engineering firms have the ability to say no and demand redesigns since it's their name on it, unlike programmers.
I frequently see firms design way beyond the bare minimum so that there's no question and it's an easy sign off rather than spending months on a back and forth.
Head architects at software firms had the ability to say no to generated code and demand redesigns. They were pushed out for being "backwards thinking".
As to your second point, that just means that the training corpus is already designed with a healthy margin of error.
You don't think they can write or create images/music/video? If you are being seriously, you should try some of them out - they are incredibly capable now.
I literally just said AI's can also create images/music/ and video/ trust me i've seen the dogshit they produce
I think their question is whether it's replaced work in industries focused on that - AI movies becoming popular and very profitable, AI written books selling large numbers etc
Very possibly on the way but so less uptake than software engineering for sure.
It's on the verge of that. If I open X then quite a few posts in the feed will be AI generated music videos right now, and for some reason people doing gender swap fanfics of Draco[nia] Malfoy/Harry Potter romances. People are finding them entertaining.
Ahh, got it. Even if it not a fully AI generated movie or book, I'm sure it is common for people in those industries to use AI.
Yeah, they can technically create those three things. They are in no way up to par if you have a single bit of artistic taste.
Even the best ones really aren't very good at this yet.
Maybe you could give an example of a Fortune 500 business that is too capital constrained to take advantage of AI. It seems far more likely to me that the issue is that of the skill sets important to these businesses, these agents are only really good at coding, and they still are very jagged even in that.
It is most impactful in tech because tech is almost nothing but languages and more so super expensive languages where one hour of an “authors” time can cost thousands with a supply demand issue so bad workers are imported from across the world.
Artists were already starving. The more physical trades are…physical not language so a plumber still plums. Maybe now they ask gemini instead of google tho.
Areas related most to…language are the most impact by large language models? The one that i think is just as impacted is education because it is also almost all language driven.
I see zero sign of any of the new mills coming around. Steel, crops, farming, none of it has changed and what little has is from…the software side or…language.
The SV world really over indexes on how much software can do.
Farming could use all kinds of robotic work right? Let the AI seeders and harvesters on the field.
I'm interested in what will happen with factories, why couldn't robots do all the work? Well maybe not ALL the work but most of it.
John Deere has been selling fully automated remote-controlled tractors with maximum DRM for over 10 years.
Robots have been doing most of the work in factories for decades already
Look all of the progress in robotics. Companies like Physical Intelligence etc. There has been massive progress in all sorts of areas from picking strawberries to assembling kits, folding clothes, etc.
Huge progress in spatial understanding with frontier models now building architecture and complex mechanical systems in Blender.
Plumbing is coming within a few years.
If you think plumbing is anywhere near being automated away you must not know much about plumbing. You will never see a bigger hodpodge of materials and fittings in so many shapes and sizes and states of decay.
Maybe on new builds with fresh material, but that is already the easiest and quickest plumbing task.
Depends what you mean by automated.
An entry level largely untrained junior doing pro level work because their AR headset and earpiece tells them exactly what to do next? That's possible today.
But nobody wants or pays for "pro level" plumbing work. Just like the majority of construction, speed and minimum viability are king, quality is the last consideration. Especially for plumbing where it all gets covered up. If something leaks in 20 years instead of 40, plumbers don't lose anything. But the plumber who put in extra effort and time does.
Many skilled tradesmen have already left the industry because higher skill levels have little value, and their labor value is just barely above grabbing randos off the street. AI at best is saving a minute of evaluation or google search, but 99.5% of the job is physical work and pain still, not high level problem solving.
Hahahaha absolutely not. There's a lot of carefully staged demos that are good for separating the gullible from their money, but we are nowhere near anything you're talking about.
GPT-2 was nowhere near being able to create web pages. It was largely incoherent. But three years later gpt-3.5 and similar were able to create simple web pages and programs and became very popular. And a few years after that, only a small portion of software engineers generally write code by hand still.
Humanoid robotics may progress faster than that because they are building on existing AI. But it will be capable of some plumbing work within a few years.
There's no "building on existing AI" because those two technologies have absolutely nothing to do with each other.
The robotics structure, dexterity and sensors are already 90-95% there in leading edge humanoids. What makes it able to be a plumber or not are the machine learning models, which are usually built with transformers (VLAs -- vision language actions models), often include frontier VLMs or models with the same architecture as a top level control, and use the same or very similar types of reinforcement learning for training.
It's far more simple. LLMs just aren't that good at stuff that isn't programming.
I why my
All businesses across sectors have been digitally transforming into software businesses over the last twenty years
>but then it took like a decade of little growth before the true benefits showed up
That's the core issue apollo is talking about. Investors have been told, "Buckle up, we're getting rich in 2027." Not "Hey, after 20 years, this will probably be pretty nice, maybe." This is how bubbles pop. Everyone starts to realize they were oversold and they rush to get their money out at the top. LLMs won't go away, but the financial collapse will be devastating.
this isn't the invention of the wheel, electricity, etc. You are definitely overhyping it. LLMs hallucinate. It is an unsolvable problem defined by the very nature of the architecture itself. No amount of thinking or counter thinking or triple checking will fix it because LLMs don't understand the basis of what is right and wrong, what looks right and wrong, why saying or doing the wrong thing is bad, and why or when it should realize what it produced is wrong, and what it should do about it.
Big Tech wants to make humans the go-between, however, the thinking and problem solving wasn't the problem to begin with. Checking for mistakes, correcting them, and presenting the corrected versions is the hard part. If a human does not go through the steps of solving the problem, they can't effectively analyze or process the output, which will always lead to mistakes. The only way for LLMs to take over any jobs would be if humans were perfectionists and could check every single bit of work without fault (since LLMs will always make some mistakes: a "perfect" LLM isn't possible anymore than a perfect human being is). Humans (especially myself are disaster and mistake prone.
That's why you should not invest now unless you are short term trading, and you should get that bag of popcorn ready. Trillion dollar valuations aren't real. It's a fake number propped up by a bubble. The US is in record debt, and it will be unable to backstop the financial loss...without taxing heavily the very folks that want to be backstopped. Those folks are dimwits and are only focused on making a buck. They are assuming everything will work out, either via investors/revenue/profits or a government bailout.
Bear Stearns had the same attitude.
In 2008 the government had leeway. This time, they do not. Regardless of political party.
Disclaimer: Please note I do not have an investment in any company, not even in a 401k. I am a disabled software engineer (and manager, in some cases) who worked in/with startups with various sectors. I have knowledge about all the financial and investor nonsense that goes into startups and small businesses. As an autistic millennial, who has definitely been around the block with bubbles, I am also interested in tech and finance. I also have a decent amount of knowledge about the damn stock market. I don't and will never claim to be an expert in any of those things, however, I know enough to say that a whole bunch of folks are going to lose a TON of money at some point within the next few months/years (likely before 2030, possibly by 2028), so if you are overexposed to the stock market (pretend it drops to zero, won't happen, however, it will drop, and no stock is safe, this goes for you FIRE folks as well), you should back off your investments to an acceptable level (the most amount of money you'd be fine with losing ALL of).
Sorry for the rant. Have a good night all.
There are some things that current generation LLMs are extremely good at. They also suck for a lot of reasons and appear to be prohibitively expensive to train as "general" models.
hey are amazing tools, but every time I see a task that LLMs helped me complete in 30% less time, I see 5 more tasks that took me 30-50% longer to complete because of errors, lies, mistakes, hallucinations, context limits.
Do not tell me I'm holding it wrong.
Well, of course not. If you borrow a trillion dollars to build a 1¢ aspirin machine to compete with $1 generic aspirin, you are strongly incentivized to sell your aspirin for 99¢. We don’t see productivity gains, because approximately all of the gain is captured as revenue for AI companies.
It’s weird that this has to be spelled out for so many engineers and economists, but to be very clear: OpenAI et al wants your token spend to go to one-minus-epsilon of what you previously spent on payroll, and wants your productivity to go to one-plus-epsilon of what it previously was. If you want 10x productivity, spend 9.9x more on tokens.
They are in the business of enclosure and rent extraction, and any innovation that comes out of their research and development - regardless of how real an innovation it is - will be completely contained and bent totally towards that business. Any leakage of the innovation that you experience as a productivity boost is literally felt as lost profits by them.
Your argument falls apart pretty quickly when you consider that (1) open-weights model are a thing and (2) OpenAI and Anthropic are still losing money
1. If AI would bring great productivity but all that extra value is captured by closed-source model providers, then everyone would switch to open-weights models + bare metal servers, which would non-tech companies to capture most of the value. Your argument that there is great value in the use of AI, but it's all captured by OpenAI and Antrhopic falls flat on it's face here. The market would adapt quickly.
2. OpenAI and Anthropic are still losing money. Your argument is that 100$ of extra value is produced, but the closed-source model providers capture 99.9$ of that because they are "rent extractors". If you're right, and they capture 99.9 of the value created, that's a huge issue because they spend 150$ to create that value. If even after eating the whole (value creation) pie they're still unprofitable, that's the worst news for the AI industry.
I basically agree that both of your points are major-to-catastrophic problems for OpenAI and Anthropic. I suppose it is true that if the companies fall apart, then my argument also falls apart. I think I would be pretty pleased in that scenario, all things considered.
I'm not rooting for anyone to have something bad happen to them or to their business.
I'm trying to be conscious about the situation because (like everyone else here) I've money on the line in the AI sector.
I hope this whole build-out turns out to be positive for our society.
Oh. Well, I am rooting for bad things to befall them. These companies are creating enormous negative externalities for society while pursuing profit and hewing to an ideology that believes it is genuinely risking destroying the world, and believes it is justified in subjecting everybody else to this risk because if they don’t, someone else might.
While in general I commend showing generosity and charity to those we disagree with, I don’t think it’s commendable to remain generous past the point where it becomes untenable.
I thought you were gonna say "market forces"
OpenAi and anthropic hypothetically undercut each other.. each epsilon-ups one another until they get to a point where they're splitting the productivity gains with the other companies. If this doesn't happen, it's called collusion and the SEC initiates action. (Ha!)
That should theoretically happen.
If we discuss theoretical scenarios, it's less likely to happen in a scenario where we only discuss about 2 companies in an oligopoly, so the argument isn't as strong.
In a scenario where we have 2 companies AND open-weight models (Chinese, American, French), all of a sudden that process is accelerated. That's basically my first argument in the comment.
Open-weights models that need closed markets hardware... (as in the markets are closed because you can't afford it)
Fair enough, there's definitely a shortage, but there is capacity available for people that are willing to buy it. There are platforms that give easy access to GPUs for small players, while the large players can make direct contract with neoclouds or datacenter providers.
The companies discussed in the article are 493 of the biggest companies in US, so if there was value to be extracted from using bare metal + open-weight models, I am sure they would be all over it.
Everyone talks about how sophisticated those large players have become in trimming their AI spend, they're a lot more cost-conscious. Perhaps this narrative might be overblown or there is truth to it. We'll see.
Not sure how productivity and rent capture are connected here.
1) If a firm spends $99k on compute to replace a $100k engineer and produces 5x the output, how is that not a 5x increase in aggregate labor productivity?
2) In your aspirin example, why wouldn't a competitor with an identical 1¢ machine sell it for 50¢ or 10¢ to take 100% of your business? You are either making the assumption that 1 company is owning the market and no models are useable beyond that one, or that all of the models that are in the front of the market have a cartel and fix the price, which is not true either looking at API costs.
(Also, pedantically, compute is not labor - if you 5x output and 0x labor, the labor productivity statistic becomes “what’s the ratio of 5/0 to 1/1?” ie undefined. But that’s not a very interesting point.)
Note closely the “middleman” vibe of this arrangement: the firm does not spend on compute to get output; the firm spends on tokens and then the token-provider spends on compute to get output, and then hands the output back to the firm. Rather than put in x and get back f(x), they put in x and get back e(f(g(x))), where e and g are the middleman’s extractive greed functions.
If they were in the business of free market value creation, they probably would! I don’t know exactly how economists would describe it, but empirically, it seems the business of rent extraction has a different dynamic - one which can roughly be summarized as “competition shrinks the size of the pie, rather than growing it”, so all participants are mutually incentivized to treat competition as a cost to be avoided.Can you demonstrate your take on 2) happening in practice?
- You have tokens priced differently as evident by for example pricing on for example OpenRouter.
- You don't reach 100$ of Fable/Mythos output by running Deepseek for the same amount of money. How do free open source models fit in here? Do you have to run Nemotron for an infinite amount of steps to reach 1$ of Deepseek compute? If not, then the argument is not valid because the pie should be shrinking.
Collusion. A cartel. That is how they describe it. And it's ostensibly illegal.
> Forms of Collusion Most criminal antitrust prosecutions involve price fixing, bid rigging, or 2 market division or allocation schemes.
This also ignores that 5x productivity doesn't often give you 5x profit, even if it gave you 1.1x profit that sounds great - until people start looking at profit vs productivity.
...we should be seeing prices of every day goods and services(mainly) come down.
I'll hold onto my hat for now.
The incentives are there. I know reddit and HN it seems like lots of people are using models/harnesses from scrappy competitors, but IME at the enterprise level I’m in the extreme minority who is even aware of them. Virtually everybody uses Anthropic and OpenAI tools and models, with Copilot getting some scraps because of its forced inclusion in Microsoft/github ecosystem
There is no question about the benefit when things go right; a wary engineer responsibly guiding agents to build orders of magnitude more than was possible before, and always skeptical and critical of the vendor, knowing the vendors’ incentives.
But that’s not the majority, and the people approving absurd token spend are too far out of the weeds to see what’s happening - people burning 6 figures in tokens and still end up with broken systems that keep them over the weekend, but now they hardly understand what they’re troubleshooting (if at all).
This is a dream scenario for the fat-cat vendors, so there is incentive to have CC/Codex write more code, overly-complex designs. When users get upset, they jump to OpenAI to spite Anthropic, or vice versa, but always getting further locked in to the duopoly
I don't really see the bloated-code take here. With this reasoning, there should appear a non-bloated open-source tool that all knowledgeable, smart developers will migrate to in order to get things done faster, and the market will have to follow.
Also waste and token spending could very well be a short-term thing. A lot of teams had crazy AWS bills in the mid-2010s on lambda functions and workflows, then it got cut down and optimized.
> 1) If a firm spends $99k on compute to replace a $100k engineer and produces 5x the output, how is that not a 5x increase in aggregate labor productivity?
I think they are saying that with 99k pure compute you are able replace 100k engineer with equal output. Or that is the goal of the AI companies. It is worse for them if you get more productivity.
It's so interesting that sometimes people throw these wild numbers around like 10x productivity, 100x greatness, etc. but when the rubber hits the road, we see so much ambiguity, unknowns and unrealistic assumptions.
But the AI companies have no access or knowledge of how much value you produce, they would have to calibrate against some average SWE salary and some average productivity which is task-dependent. There is no way for them to control this. Even if they did, they would lose out to other companies who do not do this and let you "keep" surplus value, it feels like this again assumes a no-competition market.
I guess it just seems obvious to me that most of the major players want a no-competition market, and are working towards making that happen, and are having notable success with their efforts (the whole “pacing the frontier” was a remarkably public display of coordination between purported competitors, the government has already banned deepseek, etc).
> If a firm spends $99k on compute to replace a $100k engineer and produces 5x the output, how is that not a 5x increase in aggregate labor productivity?
It sounds like they spent 99k to replace 5 100k engineers, which means they could charge 499k for it.
This again assumes a total monopoly or a cartel which does not exist today
It's a duopoly and a cartel.
This ignores the supply side of the equation and just focuses on the demand side.
Of course every business wants to capture 99.9% of the value they offer, but competitive forces drive this down to an equilibrium.
Also, as an aside, I've noticed that there's a subset of people on hackernews who use 'weird' to describe any disagreement with their position. It's not 'weird' to have an opposing or different view and removing such language will make conversations on here more pleasant.
This only works if you are the only one selling the machined aspirin. Or you build a cartel of apirin sellers to do some price fixing.
> Or you build a cartel of apirin sellers to do some price fixing.
They are really trying to do that by getting that regulatory capture
And by buying all of the available compute.
That doesn't appear to be the only thing necessary. Grok is still much worse than other closed source models, despite access to more compute.
Conversely, DeepSeek are relatively compute starved and produce a usable model (definitely not frontier but 10x cheaper).
Different with the open weight models though - they are a margin business. And openai etc have to compete with them too
And how long will they remain open?
This only happens in very mature industries. AI is obviously moving extremely quickly with lots of new entrants and moving parts. That's why all the model providers are losing money fighting for market share.
Once most of this debt is gone, whether through companies going bust or the debt being paid off or inflated away, AI will become cheap enough for frontier models that any productivity improvements will be visible in the user's bottom lines.
Once all of the trillions of dollars of debt is paid off…
…plus whatever healthy premium the corporation decides it is owed for so bravely taking on those trillions of dollars of debt, once that too is paid off…
…plus whatever the expectation of return on investment the lenders were persuaded by, once that has also been met…
…only then will we get to see what proportion of the gains these corporations think it fitting to let trickle down. What do you reckon they’ll let us have? 50/50? One in ten?
I don't think it will all be paid off. I said once it is gone. I think, like the Dotcom bubble, many companies will go bust and their debt will be worthless.
I hope so!
For the time being, my productivity has 100x'd for a fraction of what it would have taken to hire me only a few years ago.
This doesn't apply to self hosted models though.
I don't see why a payoff would have appeared yet. Many (but certainly not all) software companies have increased their output without hiring more people. But having software with more features doesn't necessarily drive immediate productivity gains in the rest of the economy. It will take time for those benefits to percolate. We've only had access to models good enough to majorly accelerate code writing since early-2026. I feel like we're in the very beginnings of the first part of this wave.
For sake of comparison, consider e-commerce. As a proxy for "impact on the economy", we can use Amazon's revenue. They went public in 1997 and that year had revenue of $148M (just 0.14% of Walmart's revenue that year). By 2000, revenue was $2.76B (1.65% of Walmart). Bricks and mortar didn't start feeling anything until probably 2011, when Amazon crossed 11% of Walmart's revenue. And it wasn't until 2025 that Amazon finally surpassed Walmart.
But, if you worked in tech in the e-commerce business, things were already feeling pretty hot by 2007. And at that point in time, the e-commerce sector was still miniscule compared to traditional retail (~4%). Software is very exciting these days and the most fantastic things are happening. They will eventually percolate out to the rest of the economy. But it will take many years before everyone notices the difference.
Tbh my day job is now bogged down with security fixes for bugs found by AI. I guess that's a productivity gain by some measure? Sure we're fixing lots of problems. Problems we were blissfully unaware of before (and arguably were not problems, if no human could find them).
Meanwhile, at least half of the engineers have been pulled off other projects to work on, well, anything that has the veneer of AI attached to it. So we've lost half of the workforce to AI. But sure, the 25% of engineer hours still working on core product have gotten 40% more productive! (Actual coding, 2-3x, but that was never the majority of the work.)
What is the long term steady state here? Not sure if there even is one. But I've always spit balled that my job would be 90% less effort if all humans were good faith actors (no need to worry about privacy and security). Just ditching those two considerations is a 10x multiplier. So conversely, if the bad actors are empowered 10x by these tools, instead of 10% of my job being actual functionality, maybe 1% can be. I need the (promised) 10x boost of ai just to get back to the historical base case.
Perhaps the people seeing massive productivity gains are ones who never bothered with security to begin with (ahem Sony).
>We've only had access to models good enough to majorly accelerate code writing since early-2026. I feel like we're in the very beginnings of the first part of this wave.
The people talking about "increased output" — do you have any proofs for these words? I'm working in IT — I don't see any radical productivity growth. Yes, you can doo hello world 4 times faster now — but even then when you need to polish it the speed of development stabilizes near baseline real quick.
The early 2026 research shows only modest improvement in productivity: https://pubsonline.informs.org/doi/10.1287/mnsc.2025.00535
> do you have any proofs for these words? I'm working in IT
I definitely see effective use of the tools.
The most common examples are non-tech or semi-tech people solving problems with Claude's help (e.g. create a query that provides X data for the auditors) and only requiring a quick review by the tech team as opposed to the tech team having to do all of the work. The pace has increased and the tech team is less of a bottleneck for some things.
Your arguments are actually confirmed by many researches: people with low competence in some area solve the problem with AI significantly faster than deducing the solution themselves. However, there is also a catch: people relying on LLM too much eventually fall behind in learning.
However, often time people start talking about something like 100% increase in performance — I just don't see it. Don't get me wrong: I totally enjoy AI making dirty work for me, but it's not something magical.
> The early 2026 research shows only modest improvement in productivity
The field experiments in this study are a few years old.
You are right — the paper describes 2023-2024 experiments. AFAIK, the most recent available is early 2025 METR experiment. There is also “How AI Impacts Skill Formation” (Shen & Tamkin, Anthropic, 2026), but its scope is narrow *yet the results are still sceptical).
> Many (but certainly not all) software companies have increased their output without hiring more people
What do you mean by output? LoC?
We're so far beyond lines of code improvements. Agents are building entire features with mergeable PRs that we now deploy with confidence. That was impossible six months ago and unimaginable a year ago.
So, my measure is not just line of code. It's deployable features. And yes, by that measure, we have gone from zero to "lots" in a very short time.
In my bubble, I’m seeing people write entire software stacks from scratch, single-handedly. At this point it’s clear to me that software engineering has taken off by a factor of 10x-100x, at least in some industry areas and in some circles.
But that’s not the point here - even if this is true and it continues, IT is not the entire world economy. I don’t think we’ll see the same productivity gains propagate more widely soon. This will take years. The real world is a ship that’s immensely hard and slow to steer, which is easy to forget when your CI pipeline deploys commits to production in minutes.
I can "write" a whole OS kernel by running `git clone https://kernel.org` — that does not make me much of the programmer. Copying software was never a problem — making the software perform the job you want was the problem.
Are you under the impression that LLMs take source from the internet and put it on your hard drive? That’s your good faith interpretation of what’s going on?
It does exactly what you say: takes source code from the internet and put it onto your hard drive, but in a very indirect way, occasionally verbatim, in other cases it renames identifiers and does a bit of composition.
I actually reported one such case: https://github.com/cesanta/elk/issues/75
The kid had no clue LLM just copied implementation from other similar project.
You are completely discounting all of the original code being produced with LLMs. What percentage of LLM-generated code would you estimate is plagiarism? 1% 10%? 90%? 99%?
99%. LLM cannot do a bubble sort on its own — that's one of the main modes of failure actually, when LLM fails to recognize algorithm and writes its own implementation. LLM is basically recognizing, concatenating known algorithms, renaming variables, functions names, interpolating multiple similar solutions into one target solution — with a success rate corresponding to the number of solutions it saw for this task.
In case you would say "but LLM-s learn on random source code from the internet" — not anymore, modern LLM-s are trained on highly curated, preprocessed, and some synthetic datasets. As I've said — the copying is indirect.
I've submitted two PRs to fix bugs in open-source system tool last month. Both of these are targeted fixes to highly-specific bugs that existed in this software. The agents I used debugged the issue and implemented fixes and tests autonomously.
These problems looked slightly more difficult than bubble sort to me. Do you suppose these bugs existed in some other software and the agent only translated the fix between repositories? When the agent was debugging the software, were they simply following some tutorial they were trained on? Is this the 1% case? Or are the stochastic parrots regurgitating most probable tokens in approximately the right order to provide an illusion of work being done?
> In case you would say "but LLM-s learn on random source code from the internet"
Of course they don't, it's not 2022 any more.
You were talking about code writing, now you switched to debugging. Those are wildly different tasks, and some researchers suggest to employ separate LLM-s for code writing vs debugging.
For debugging LLM-s reproduce fine-tuning datasets made by humans actually doing debugging on some code. And yes, the problems AI found in your code were similar to the problems humans found when writing those datasets.
> You were talking about code writing, now you switched to debugging. Those are wildly different tasks
No, they're basically the same thing, intrinsically linked and inseparable. Likely represented together in training sets, too. I'm surprised anyone would say that. Are you a programmer?
> For debugging LLM-s reproduce fine-tuning datasets made by humans actually doing debugging on some code.
Not really, they are primarily trained on LLM-generated trajectories from executable training environments.
> And yes, the problems AI found in your code were similar to the problems humans found when writing those datasets.
So they're only similar now? I suppose, in the way that I learned how to solve bugs while learning to program and now I can fix similar ones? I guess we can now agree that LLMs are capable of learning and extrapolating knowledge to new problems?
>No, they're basically the same thing, intrinsically linked and inseparable. Likely represented together in training sets, too. I'm surprised anyone would say that. Are you a programmer?
No, they are not the same, and they require different tools: you cannot write code in gdb or valgrind. Even in IDE-s debug and editing interfaces differ significantly. You can literally separate gdb session from any code editing, do debugging with zero code writing and then relay the findings to someone who will make the changes.
The fact that you personally do both things does not mean everybody and everything does it this way. And certainly LLM-s don't function same way people do.
And no, debugging and code writing are not in same datasets. More to it, debugging and architectural planning is mostly represented in fine-tune datasets (prompt-response, agentic step prediction), while code writing is mostly (like 99%) pre-training data (refined internet scraps and regenerated code).
>Not really, they are primarily trained on LLM-generated trajectories from executable training environments.
Training on purely LLM-generated data is known to give horrible results. Big AI companies already have thousands of debugging sessions at hand — those bring high novelty/variability into the training sets. It is possible to multiple those fully supervised sets via LLM-generated bugs or bugs from github reproduced in a sandbox. With all that said, feeding lots of very similar bugs and solution trajectories does not help much in training — so it's rather "combinations of known human skills tested in a sandbox", not "totally random efforts transformed into datasets of successful debugging outcomes".
>So they're only similar now?
`c = a + b` and `c = a + d` are similar, but the difference is very small — it's about variable rename. It's not a unique code. And I never told LLM has to reproduce 1 Mb of source code byte-to-byte without any change. Just in case some of my words sounds like I told — I did not mean it.
>in the way that I learned how to solve bugs while learning to program and now I can fix similar ones? I guess we can now agree that LLMs are capable of learning and extrapolating knowledge to new problems?
No, LLM does not learn to program same way people do, it generally does not learn things the way humans do. Why would they? The training of a human looks nowhere close to how LLM-s are trained. Even the way people translate similarities is not the same as LLM's: LLM does precise literal swap, human extracts meaning and unfolds the solution into concrete words. Humans chase goals and plan things, LLM-s pattern-match and reproduce superficial shape, treating steps in complete isolation.
Human can try to reproduce actions with zero understanding of the subject — which usually leads to very mediocre results even after lots of effort. Basically, LLM can be seen as a stupid and blindly-diligent office worker x100.
> No, they are not the same, and they require different tools
Yes they are, and no, they don't. Debugging can be done by writing/deleting/modifying code (e.g., write a test and run it, or add log statements, or remove a line and re-run the program, etc.), or by running shell tools (either as single tool calls, which you may to be a one-line program, or with larger compound commands, or entire shell scripts - throwaway or saved to a file), or by writing code that analyzes debugger output and logs, or editing code at runtime with hot reload, or - depending on the runtime - attaching to the running program and running ad-hoc code (REPL-like), and so on. It's literally impossible to enumerate all of the options because it's all mixed in one codey-debuggy-flavored soup. Are you saying you've never seen an agent do this?
It's also fun to write your own debugging tools by the way - just imagine any debugging workflow you think is good, write the tool, give the agent some tips how to use it, and they'll do great. It's so good, you should really try it. This includes debugging tools that are operated by writing code, it totally just works.
> you cannot write code in gdb or valgrind
I'm actually pretty sure these are extensible, and I'd bet quite a lot that an LLM could write that plugin.
> The fact that you personally do both things does not mean everybody and everything does it this way.
I don't, but LLMs certainly do.
> And no, debugging and code writing are not in same datasets. More to it, debugging and architectural planning is mostly represented in fine-tune datasets (prompt-response, agentic step prediction), while code writing is mostly (like 99%) pre-training data (refined internet scraps and regenerated code).
Programming and debugging is self-evidently part of the same training loop. Agents write code, validate it and adjust trajectory. I'm not saying that they don't train agents for debugging-centric tasks specifically, but these certainly often - if not usually - involve programming as well. The same way coding tasks often involve debugging.
> Training on purely LLM-generated data is known to give horrible results.
I don't know what to tell ya, it's how the LLM sausage is made.
> `c = a + b` and `c = a + d` are similar, but the difference is very small — it's about variable rename.
I think you're vastly underestimating the difficulty of real world programming tasks. It's a pretty common mistake, I usually see this in people who are struggling to downplay LLM capabilities, and their strategy is to pretend that everything LLMs are good at was trivially accessible in the training data all along.
> No, LLM does not learn to program same way people do, it generally does not learn things the way humans do.
Obviously, it doesn't. I said "LLMs are capable of learning and extrapolating knowledge to new problems". You can throw a novel problem at an LLM and you'll get a novel solution. This is obvious to anyone who works on interesting problems. You could literally hit CTRL+T and do it right now. It's such an easily and accessibly verifiable fact. What a weird hill to die on, and yet so many corpses there.
>Debugging can be done by writing/deleting/modifying code (e.g., write a test and run it, or add log statements, or remove a line and re-run the program, etc.), or by running shell tools
We can end with things like "drinking cofee is an essential part of debugging for me". Debugging is an integration work, just like architecture planning, so it's somehow tied to plain english skills, reading docs, running whatever existing commands there are, parsing tools output is very important thing.
Another point on which we should reach an agreement is the definition of debugging, as I suspect you may have your own definition. Fixing compiler error is not debugging. Glancing over source code is not debugging. Debugging is characterized by running a program and comparing the actual outcome with the expected one — implying finding the exact part of the source code responsible for the problem. Meatbags often do not read the full related code, and sometimes are not even experts in that particular programming language, or simply don't own the code and have hard time digging into it.
Countrary, LLM-s excel at bruteforce code sifting and spitting two more 200-lines python scripts each minute. Still, even with AI debugging is not just "read and write code" — it's rather a high level skill with code read-writing being the largest hammer in the belt. Hammer so large AI will write several large scripts rather than use existing simple command.
>It's also fun to write your own debugging tools by the way - just imagine any debugging workflow you think is good, write the tool, give the agent some tips how to use it, and they'll do great. It's so good, you should really try it.
You conflate crutches with features. LLM-s absolutely suck at interactive tools, so they have to compensate with their strongest side: writing-reading code. Not because that's the absolutely most efficient way to debug, but because the economy is shaped differently in the world of AI. Running application manually, manually collecting logs, and then feeding them into AI is not "debugging made by AI" — it's the debugging made by human while AI doing the "code sifting" part of the debugging process.
>> The fact that you personally do both things does not mean everybody and everything does it this way.
>I don't, but LLMs certainly do.
When I look at my agent doing 1000 characters one-line python script for testing/debugging, I do realize there is no way it could be a common code-writing skill, regular python code does not look like this, the skill requires some special understanding of agentic loops, it cannot be from pre-training dataset. And you definitely cannot say that "being able to write code" automatically enabled "being able to debug" — just try any pre-2025 model that is decent at code writing but not trained for modern tools/harness (I'm pretty sure some newer models also lack this ability, but such models are harder to spot). I'm pretty sure everybody here tested pre-2025 models, it's really nothing to argue about.
>Programming and debugging is self-evidently part of the same training loop. Agents write code, validate it and adjust trajectory.
I would suggest you to re-examine some fundamental explanation of how LLM training actually happens, because it seems to me you don't differentiate pre-training and post-training stages (and post-training has lots of flavors in it). Trying to train LLM for agentic coding and debugging before it learns basic code patterns is a huge waste of resources. "Never say never" — if you manage to generate dozens of billions of agentic training sets, you can probably end up doing "coding+debugging pre-training", but so far neither Anthropic nor Google have resources to do that, rather they employ millions of fine-tuning samples. And I heart those companies have lots of money.
It is true that modern LLM-s are trained to do basic chain-of-thoughts/agentic loops at the very begginning of fine-tuning (i.e. after pre-training), but it should be noted that LLM can do implicit chain of thought and code writing BEFORE any fine-tuning. Just the base model can do that, but the format of input and output does not correspond to what humans and harness expect. And that's basically how the chain of thoughts was discovered: researchers asked LLM to solve some task step by step — and it did. Nobody trained LLM to do that, it just reproduced some patterns from pre-training sets. And then researches started training LLM-s to specifically optimize agentic loops.
> I don't know what to tell ya, it's how the LLM sausage is made.
Even small Phi-4 was pre-trained on only 40% synthetic data, despite Phi-1 employing 99%+ synthetic data. Once again, you should split pre-training and post-training.
On modern super-large LLM-s the large corpus of agentic loops for fine-tuning is indeed generated via external expert agentic loops, but it never becomes 100% as the model collapse is real and well known. Just in case the wording "model collapse" is not clear: the model starts using extremely limited set of tools to solve every task, even when much more efficient tools are available. I can see the tendency for model collapse on Grok 4.7 and Sonnet 5.5 I often use, and their engineers definitely partially counteracted this tendency.
>I usually see this in people who are struggling to downplay LLM capabilities, and their strategy is to pretend that everything LLMs are good at was trivially accessible in the training data all along.
I tried feeding non-canonical Tarjan bridge-finding implementation into AI approx year ago, and it went totally crazy: it says the implementation is incorrect, then traces it via supposed counter-example and finds it correct, then tries to give new explanation and once again suggests the implementation is bugged... The fact we put lots of crutches beneath the same pattern-matcher does not make it a completely different mechanism.
Most people never leave their sit during the AI ride so AI drives them through the pre-defined happy path. But once you try to deliberately leave this road, LLM-s often mark the code "wrong/unoptimal/poorly formatted". That's particularly a reason why older agents used to agressively introduce completely unrelated refactoring/formatting changes — because they just did not like how many spaces you've used in your code — and modern LLM-s still suffer from this problem, but their harness and agentic loops usually hide at least part of those meaningless changes.
The phenomena is long-known, LLM-s strongly prefer their own ideomatic code: https://openreview.net/forum?id=4NJBV6Wp0h
>I said "LLMs are capable of learning and extrapolating knowledge to new problems". You can throw a novel problem at an LLM and you'll get a novel solution.
My experience is quite the opposite: I give a novel problem and LLM suggests 10-years old solution using libraries that are long deprecated. My whole life's hobby is about finding rare stuff, and I have very broad experience in this area. I particularly like LLM-s for finding some rare old thing I've never heard about, but in terms of making up something totally novel LLM-s totally sucks. Arguably, an ant is more intelligent when dealing with novel conditions than the most sophisticated modern LLM is.
Where is this new, valuable software having an impact? Yes, there is quite a lot more code being written. But where is it being run?
This is small-scale but I know a lot of Blind people who are vibe-coding accessible tools to solve their problems. Most of them have never done any programming at all.
For example, my friend Marco has built like 10 different apps. I've used some of them and they're useful. I think this is an example of something that just wouldn't be happening on this scale without LLMs.
https://marconius.com/
Agreed and I have also built many useful tools for myself. Many of my peers in business functions such as sales and finance have done the same thing, but these mostly amount to personal productivity, which is hard to measure. The claims I'm responding to are that software productivity is 10x or 100x - and we just aren't seeing any outcomes that support this. In fact, development work on our core platform is not more efficient at all in terms that we can measure - like delivered story points (which are not perfect, but are what we have).
We are coming an external tool as they 15xd prices on us. We have spent a lot of tokens and about a month, and are incredibly close to matching the external tool.
LLMs have made stuff like this much more possible.
Granted, we have an absurd amount of tech debt, but given our deadline that's an acceptable risk.
Note that the reason this works is because we can compare to the external tool, which is a really really good scenario for the current models.
In more greenfield areas they are less useful. Ultimately though, you can't outsource understanding, and that feels like the limiting factor now.
> In my bubble, I’m seeing people write entire software stacks from scratch, single-handedly.
This is becoming common, but we should ask what is it really? The analogy i see with these SME vibe coded apps is the same as the graveyards ie file shares full of forgotten but once useful excel spreadsheets and access forms/dbs. Coding was always part of the problem. This highlights deployment and adoption are also tasks to be performed.
Deployment for who and adoption by who? In an ai future that would be agents. App registries where agents can discover capabilities would seem to be a thing the future would need.
What about productivity gains (recent/profit, company bottom line)? I think that’s what the op’s measuring. Have you seen a 10x-100x on that front? If not, what’s the disconnect here?
Productivity != sales, especially in software, I think. This has been happening for just a few months, less than one project cycle. I think we’re about to see a huge wave of software products. But it’s possible most of these are competing for a slice of a constant-sized pie, though.
When your manager/boss asks how this huge productivity gain gonna lead to better product -> sale, and why are they spending a good amount of money on AI tools, do you say “Productivity != sales, especially in software”?
I don’t know anyone in this situation. No idea how it works in corporations, and I’d rather not find out.
I know some people who pay for their own tokens because it lets them do their job in 30 minutes a day, and spend the rest of their time creating a side project. I know several freelancers who didn’t make software before, and now they have custom productivity tools tailored to their jobs. Others work in small teams, where the boss doesn’t need to ask for proof because they’re burning tokens themselves.
It doesn’t matter who the boss is. It’s the same dynamic. That person needs to live, and they need to earn money. The money spent, needs to be justified in profits made, otherwise it’s not a business. I was questioning the absurd 10x-100x company output, which is relevant to the op.
> It doesn’t matter who the boss is. It’s the same dynamic.
I'm guessing you haven't had the displeasure of working for a boss who doesn't code and has no idea what he's paying you for - lucky you.
> I was questioning the absurd 10x-100x company output, which is relevant to the op.
I'm talking about programmer productivity, not company output. For pure software companies, these may be closely related, but there's many other processes involved with converting code into money. Either way, the 10x+ productivity increase is obvious to everyone I work with.
> I'm talking about programmer productivity
I was not: https://news.ycombinator.com/item?id=50226241
I was: https://news.ycombinator.com/item?id=50226077
Why evade the question?
Yes. Because it is not my job to optimize for sales. It's to optimize for time to ship whatever features/fixes the manager/boss gives me. There's an entire sales wing who can answer the question on how/if shipping this out faster will actually make the company more money.
Asking the workers why they need to justify the responsibilities they were hired for is a small part of what's pumping this bubble up.
> Yes. Because it is not my job to optimize for sales.
I wish I could be a fly on the wall when that happens.
"You can see the computer age everywhere but in the productivity statistics." -- 1987
Good reference. It is going to slowly change things in 20 years we'll look back and a lot of things will be wildly different because of AI systems. But its rate limited by institutions and momentum and human psychology and the way things change will be somewhat unpredictable with unintended negative and positive consequences.
"slowly change things in 20 years"
Yeah, it is not going to be slow. Progress compounds, everything is being built on top of fast computers and fast widespread internet connection. You are not at the beginning of the curve like you were in the 20th century when people didn't even know how to use a keyboard and we are already seeing widespread disruption in security.
Literally the entire premise of the article is that it's going slow everywhere but in software. To think otherwise is to be really myopic about how the rest of the world outside of tech operates. And in any case 20 years isn't that slow. The only things that happen faster that that will be things like adobe goes out of business. And let's say AI helps us develop medicine and robotics and better factories and all the other things that would have a big impact a lot faster.. just human trials could take a decade before you really see that much change. Will it have a wider impact faster than computing its self, or the internet? Maybe? Maybe not.
If a lot of jobs were still manufacturing jobs I would agree, but we are in a service economy that uses computers as the main tool. So I don't see how it won't come into effect fairly soon into other areas of work. You can't measure impact by profit alone so I don't find the article particularly interesting anyway.
I tried telling my boss "you can't measure impact by profit alone" and he said, "you're so right, we also have to think about expenses! You're fired."
I would also fire someone who doesn’t know the difference between revenue and profit.
It is like saying there is no impact on math because solving a millennium prize problem didn’t net them a profit. It is just a naive way of looking at things.
You are a demonstration that their marketing campaign is working. The reason OpenAI is focusing on math is because it is one of a very, very few domains where RLVR can be done at massive scale because large sets of problems with automatic verification can be produced and evaluated relatively cheaply. It has no value to OpenAI beyond marketing. It has negative value for mathematicians.
So far the only economically valuable activity that AI has really proven itself very capable in is coding - and its for the same reason of economical RLVR. It can do significant tasks completely end-to-end in that domain. It is very helpful in other types of tasks such as data analysis and legal research, but those things are not going to fulfill your expectations for broad economic impact.
There is simply no evidence at all for the broad claims of disruption in 1-2 years.
Seems to me like you're undervaluing the impact of software in the modern world, as if software hasn't driven massive productivity gains in every industry in the last 50 years.
Software has not driven massive productivity gains in every industry in the last 50 years.
Do you think those offering professional translation service were obsessed with AI? Or chess players? The rest of the world doesn't get to decide...
Both things can be true.
I expect to be surprised by specific headlines, but I don't expect "software engineer" to be any more common a job opening than "steam locomotive engineer" in 10 years time.
Conversely, I would be very surprised if the robotics startups turn out to have solved general-purpose AI models that work in the real world, both fast enough and cheap enough (currecy or joules) to be significantly labour-displacing in 10 years.
Where I'm about 50-50 on which of "AI does it" or "AI does't do it" is more surprising, is specific physics research from AI. LLMs can do a lot of hard maths, there's lots of open problems in physics due to hard maths questions. So:
10 years? I'd be equally surprised if this is or isn't done by AI. Mainly depends on how long the growth continues looking exponential."You are not at the beginning of the curve like you were in the"
Note: on an exponential, you can always say this. People will look back on today with the same "aww cute" eyes.
Progress is future knowledge. You don't know what future knowledge is, yet. When you know, it will have become present knowledge. You don't know that any particular ongoing change is progress. LOL predictions.
Disruption in security is still software field though. The topic is "the rest of the economy".
It IS going slow though. The rest of the economy will be slowly adapting it, just like they have slowly adapting other computer systems before. There simply is no use for frontier models in the real economy as of now. For the economy it's just another cost factor, a substantial one if they try to actually build something with it, that has no real benefits that results in increased productivity and revenue making. So why should anyone except AI companies adapt frontier models? They have no reason to. And it will be that way for many years to come.
Yep, and if it's anything like previous transformative technologies: the internet, and 100 years earlier the railways, a lot of people are going to lose a lot of money overbuilding the wrong things along the way.
Maybe it's also rate-limited by the fact that it's not actually a silver bullet and actually needs time to find the proper place for its application as well.
Computers, of course, can be criticized in the same way.
And likewise; computers took a few decades to actually become a widespread economic force. Arguably the impact was quite small if we separate it from the internet (which of-course depends on computers).
Hard to describe the impact of computing as quite small compared to anything, even the internet when you go see those old room sized early IBM computers that were the first commercially viable ones for businesses. They used to have warehouses of payroll documents and armies of people to shuffle around paperwork and crunch numbers and physical people connecting phone lines. Just absolutely fantastically massive amounts of busy work.
Good insight. For a long time, the utility of a computer was tied to its ability to print something onto a sheet of paper. User interfaces came along, too, though. Especially for industrial applications. Then internet was probably the third wave. Now AI is the fourth?
I think it's pretty silly to try to separate the internet from computers.
Yeah, the person who said that obviously never checked on book and magazine and newspaper typesetting --- cue old news stories about the last compositor at a given paper finally retiring after decades of the sinecures which the composition unions bargained for rather than participating, causing a loss of typographic skill and knowledge which resulted in the ghastly appearance of the "DTP revolution".
I once participated in a project where we worked up a database-driven typesetting system for a company whose annual publication had a staff of a dozen (3 managers (one of whom mostly did ad sales), 3 full-time compositors, 6 part-time) and when we put in the system they quit paying their expensive annual license for maintenance on the proprietary composition system, and fired two managers and 1 compositor and half the part-timers.
The staff was commensurately larger when they switched to the first electronic composition system they were using from the typesetter and paste-up they initially used.
> the person who said that obviously never checked on book and magazine and newspaper typesetting
That's one of the "everywheres" the author of the statement was talking about, which is the point - you could see every part of the economy seemingly improved, but it did not show up at the macroeconomic measurement level.
It was until the 2000s that computers caused an uptick in measured productivity at the GDP level, though.
The number of books being published has been steadily increasing for decades since computers became prevalent in typesetting --- is that not measurable?
Admittedly part of that is caused by the change in tax laws making it untenable for publishers to maintain a backlist.
> is that not measurable?
It was not on a level that nudges GDP growth for a long time. The irony is that maybe the productivity didn't increase until book publishing plummeted, since enabling a lack of physical books was a huge productivity win.
This isn't a debate about whether industries improved. When you look at GDP, do you see an uptick in GDP or not?
The original quote was often seen as a paradox. At least one major interpretation of the original quote is that "the computer revolution is so obviously helping the economy, that if it doesn't show up in GDP, is it even a meaningful measurement, and how can we fix that?"
But it is very fitting to discuss that quote - the original article discusses multifactor productivity, the same relevant measurement, and earnings of the S&P 493. There are two ways to interpret it - is AI not increasing productivity, or are we not measuring productivity, and is there "dark matter" somewhere in the economy that is not being properly measured?
I think we're looking at two different sides of the same coin.
The thing is, increased efficiency doesn't always yield a larger market in the short term.
There are only so many copies of a book listing television stations and demographics for marketing purposes which the market will bear as in this specific instance.
It takes a while for the displaced workers to get retrained and find other jobs which allows them to produce/consume more.
On the flip side, Nintendo hiring a couple hundred folks to make a game such as _Legend of Zelda: Skyward Sword_ suddenly created $50 x 600,000 in value on the day it was released.
Exactly what the quote is about!
And that's a critique of the original article... that the broader "S&P 493" isn't showing gains might be for the reasons you describe.
To be fair, in 1987 the unit economics weren't being subsidized by nation-state levels of debt and mandatory usage contracts.
The unit economics aren't what's subsidized. Inference is profitable.
Didn’t we just get a bunch of data from Anthropic’s IPO showing that inference is not, in fact, profitable? People are still spouting this?
The distinction is adjusted vs non-adjusted basis. Serving inference is profitable, but serving inference for the last released model doesn't cover the cost of the model in training.
Dario explained it very clearly in an older interview. If every model is bigger and more expensive to train, each model can individually be profitable, while the company as a whole still shows a loss at any given point in time. This is why he likes the idea of regulation so much, it'll let them take a breather and recoup.
> Dario explained it very clearly in an older interview. If every model is bigger and more expensive to train, each model can individually be profitable, while the company as a whole still shows a loss at any given point in time.
“We’re profitable as long as you ignore our ever increasing cost of goods sold, we promise it’s not a pyramid scheme” - ~~Charles Ponzi~~ Dario
And that interview was explicitly a hypothetical. It was shared here a few months back.
Dario also believes that verifiability is trivial, which I'm much more sceptical of.
But the reason the cost is ever increasing is the steeply ascending quality. There's no (non-doom) scenario where this game does not end with profitable cheap tokens. Wherever it'll stop scaling, it'll be profitable to serve tokens at that scale. And of course if it doesn't stop scaling, it's the singularity and all this thought of profits is quite secondary.
An industry cannot fail because it's too successful.
edit: Other explanation. The industry has found a moneyprinter. It prints money, and several companies are building them and printing money. However, each one of those companies is cashflow negative, because they're all trying their desperate best to build an even larger moneyprinter. Why? If they don't build it and the next company over does build it, and it works, then company 1 is differentially outcompeted. what's more, the more expensive each moneyprinter is, once a company gets strongly ahead the second company won't get the financing to keep up with that one. ... But it's still moneyprinters. This is not an economic race that ends up with "and thus everyone stopped printing money."
> But the reason the cost is ever increasing is the steeply ascending quality.
Sure, if you say so.
> There's no (non-doom) scenario where this game does not end with profitable cheap tokens.
I doubt it.
> Wherever it'll stop scaling, it'll be profitable to serve tokens at that scale.
It’s much more likely, imo, that local AI models will completely destroy the entire business model of inference. For any entity that would be spending enough to replace multiple employees (allegedly), the cost of the hardware to do so is completely affordable. This has the major benefit of eliminating per token costs, and practically guaranteeing that some “AI company” can’t steal your company’s secrets.
At an individual person level, this hardware will take much longer to obtain, but it will happen eventually.
> And of course if it doesn't stop scaling, it's the singularity and all this thought of profits is quite secondary.
I still believe, strongly, that LLMs are the wrong path to AGI. There’s trillions of dollars being put into turning LLMs into AGI; however, there has no zero results showing that LLMs are capable of “learning” anything. “Skills” and expanded context windows don’t count. The “model” needs to be able to modify itself, in real time, for that to be possible.
We’re in a local maxima that strongly refuses to accept that it’s a local maxima. (Diminishing returns for the amount of money poured into “training” is beyond obvious to anyone paying attention).
> An industry cannot fail because it's too successful.
Sure it can. The fossil fuel industry is wildly successful.
It’s also boiling the planet alive, but that’s okay, it’s generating shareholder value so everything’s fine.
Edit for your edit:
> edit: Other explanation. The industry has found a moneyprinter. It prints money, and several companies are building them and printing money. However, each one of those companies is cashflow negative, because they're all trying their desperate best to build an even larger moneyprinter. Why? If they don't build it and the next company over does build it, and it works, then company 1 is differentially outcompeted. what's more, the more expensive each moneyprinter is, once a company gets strongly ahead the second company won't get the financing to keep up with that one. ... But it's still moneyprinters. This is not an economic race that ends up with "and thus everyone stopped printing money."
By your own definition it is impossible to achieve profitability without an ever increasing amount of money flowing into it.
There was a business about moving stamps around that needed similar levels of cash infusion to sustain itself. Eventually its bubble popped too.
> It’s much more likely, imo, that local AI models will completely destroy the entire business model of inference.
You're massively underestimating how much cheaper it is to serve models in a datacenter than at home.
> By your own definition it is impossible to achieve profitability without an ever increasing amount of money flowing into it.
No, what? No. This is just a total misreading of what I'm saying. It's extremely easy to achieve profitability. It just requires your business to stop chasing more profitability. The problem is just that then the other businesses who are still chasing more profitability will eat its lunch. It's a market shape that easily supports n-1 bankruptcies but never n bankruptcies.
> Diminishing returns for the amount of money poured into “training” is beyond obvious to anyone paying attention
Sure, if you say so.
it doesn't seem that complicated to me. what do you think the cost is for AWS to serve inference it sells to Anthropic for "$1.00"? I think it's about $0.11. Then, ask yourself, did Amazon sell it to Anthropic at all? No, they traded $0.11 real dollars for $1.00 of Anthropic shares (the CLN that created the real loss Anthropic is talking about), which are now worth $10.00 (which is why people talk about this as a paper loss). So... IS it profitable? It's profitable, but Amazon didn't JUST sell inference when they turned $0.11 into $10.00...
That's a Ponzi scheme.
Nope, the leak didn't show that.
in some kind of boring sense, sure.
but have you looked up who invented the computer? and who was funding all the first versions of the software we use? it was the US military.
Where do you think the internet came from?
Neither of those things. Arpanet was military and the web came out of CERN, but they were both low budget side projects.
...and global usage/reach was practically zero...
The difference, of course, is that in this analogy Commodore has to become a trillion dollar business by 1989 or the entire stock market implodes.
Like the internet bubble, that’s fine, openai and anthropic don’t have to survive for ai to be successful
Fully agree with that. The problem is the AI boosters really believe in The Singularity that will transform all of human experience. If you try to tell them AI is a high-value enterprise technology -- more valuable than docker but less than the steam engine -- they get mad at you.
And at that point, there was no much productivity gained ... and not nearly as much debt as now. The industry was also not such a huge part of stock market, GDP, etc.
Could a Fortune 500 company survive competition from competitors if it did not use computers?
TFP data is murky at best IMO
Came here to say this as well.
It took decades for effects of personal computing to show up in aggregate statistics.
See:
https://en.wikipedia.org/wiki/Productivity_paradox and https://dl.acm.org/doi/10.1145/163298.163309
edit: My comment is not exactly right, its more about general computing rather than personal computing
Thanks for posting this!
> As with previous technologies, an extremely large number of initial cutting-edge investments in IT were counterproductive and over-optimistic.
I think we are definitely seeing this today!
Someone should add a 2025–2045 section in the Wikipedia page of Productivity Paradox
The decade ahead is going to be a very interesting one
That would be speculative, it should only be added in the hypothetical time the productivity actually goes up. Even then, it should be some years in and with third party sources like all of Wikipedia.
So this was about personal computing ? not computing per se ? I'd assume that by 87, the idea of digital computers benefits would be a non-question.
It wasn't, generally, a question. People kept buying computers, on faith.
It's just that in 1987. the computers were generally and for the most part, irrelevant to productivity, just like steam and electricity were a promising gimmick for a long time before.. they weren't.
Looking backwards, it is easy to miss this.
at the level of personal employee maybe, even though 87 had lotus 123 version 2. I was barely born but I'd think people would be able to think faster through electronic spreadsheet.
I was too specific, it was more about computing generally as opposed to "personal computing".
Only a few curmudgeons at this point are claiming there is nothing, nothing whatsoever, to AI.
The question is really about the fate of the current build out. For those of us who have extensive experience with AI, which is to say, those of us here, because we are probably the only people not only using it day in and day out but potentially using multiple sessions of it all the time, I think it's not so surprising that it isn't necessarily immediately radically transforming every industry everywhere. Even as it has rewritten the minute-by-minute process of doing my job entirely, it hasn't actually changed my job all that much at the higher level and it's hard to assign much more than a 2x or a 2.5x overall improvement to it once I account for all the Amdahl's Law issues that come up from the parts that have not been so amenable to acceleration.
And without unit testing, continuous integration testing, pull request review, the fact that the AI is building code which is itself a highly reviewable artifact rather than just doing whatever directly, and all the many many guard rails we've built up in our industry that happen to be extraordinarily amenable to use by automated systems, I still wouldn't trust AI all that much for a real task. They still bang into those guard rails a lot, and still make what I consider to be deeply suboptimal decisions far too often to take my hands off the wheel. With all the stuff we have in the programming space they are a game changer, but in a field that doesn't have anywhere near that much padding and safety nets I'd be way, way more hesitant.
I want to say it's weird to talk about such things when they are also capable of such amazing math proofs, but at the same time when you consider the numbers tossed around for how much those things have cost, I think you're seeing another space that has guard rails comparable to software engineering. By contrast to the way these AIs are working, when a human solves a comparably hard problem the human tries a few things but from the outside would seem to proceed almost directly to the solution, compared to the staggering amount of blundering about the AIs do before arriving at their solutions. Now, in terms of the final results, yeah, they got there and I won't deny that. I'm just saying that in fields where the acceptance criteria are nowhere near as sharp as math or software, they're operating at a significant disadvantage still.
> I think it's not so surprising that it isn't necessarily immediately radically transforming every industry everywhere.
Computers and the internet didn’t radically change every industry everywhere. But every industry everywhere uses computers and the internet, and more pertinently, they pay good money for it.
AI doesn’t need to transform the world, it just needs to pay off the sizeable investments in data centers.
Does AI transform the world more than; - CAD software - Office software - Spreadsheet software - Finance software - Digital banking software - Computer graphics software - The entire US national fiber grid
Cause at the current rate and scope of data center investments; they have to transform the world more than... every other software system we've come up with until this points combined; to justify their cost.
They have to transform the world more than whatever the bond rate is right now, (which admittedly is increasing). I'm not sure that that's greater than office software, digital banking etc. But I don't think it's as high a bar as some people think.
From a paying back data centre debt point of view, yeah. The equity valuations on the other hand are priced for total transformation in the short term.
And is there enough business or computer demand for all the data centers we're building? I don't think anyone has an answer to that.
> But every industry everywhere uses computers and the internet, and more pertinently, they pay good money for it.
Do they? Compute and internet access are orders of magnitude cheaper than the software that runs on it.
In a way it’s sad that AI came in the age where a lot of us are jaded because the internet which we built has become a place for gambling and scamming. I remember when a new language or framework was met with excitement. Today I hear (and am a source of) mostly criticism.
I think there's also something to be said about how the largest private AI enterprises are themselves mostly just software companies. There is enough widespread dogfooding of the product that the company staff are not just spending a lot of time and energy in improving how their product is specifically adapted and trained for better code but they are also intimately aware of those limitations, driving how they think about and implement their own product.
I don't think that translates well for other fields at all. You can throw an engineer or an MBA-brained exec at a business problem all you want but until they start using the products, they're not going to be effective at guiding its development.
Math is just proof of concept for the big players. The same resources could go to prompts designed to optimize the internal combustion engine or anything else nonsensical.
The biggest paradigm shift has been representing models as humans, and I don’t think that’ll fade. I highly expect the AI Model T Ford will be a generalized assistant that actually works. Likely developed by OpenAI or Anthropic. And likely 5 years from now or more.
Math was not "just" proof of concept, it was THE proof of concept, since iterating on combustion engines cannot be done nearly as efficiently.
"optimize the internal combustion engine" that requires good quality combustion simulations and iterative optimization algorithms; and we've had that for longer than the internet; that kind of work was the earliest uses of military comptuers. Why would a text model be better at it?
Maths i a cool benchmark, but has the similar benefit to software; very easy to verify.
> I still wouldn't trust AI all that much for a real task
I wonder if you might be conflating LLM/Agent with AI
Waymo is AI, I personally would happily ride one, or a Zoox. Those are quite real tasks the AIs are performing, people put their lives on the hands of these AIs
A lot of AI models every day work to provide search results, driving directions, image recognition, automatic routing and classification (eg. Fruit packaging companies use vision models to accept/reject fruit for packaging in real time - that’s the fruit we buy at the supermarket and then eat at home), and so many other things
So inadvertently, and without much choice either, you probably already rely and depend on AI on your everyday life
Is every closed loop computation system AI now? Or does it need to have a neural network?
We’ve been arguing about this for decades. I think the less controversial take is that very specifically LLMs and their agentic harnesses are AI.
> Is every closed loop computation system AI now? > LLMs and their agentic harnesses are AI.
Not sure what you mean. Are you being ignorant or sarcastic?
https://en.wikipedia.org/wiki/List_of_artificial_intelligenc...
A* path finding and SLAP/SLAM (used in video games and robotic systems), BFS/DFS, Alpha-Beta pruning, Simulated annealing, Genetic algorithms/programming....
All of those are considered AI, none of those are LLMs or agentic harnesses (which of course are also AI). Also, most real AI tasks systems currently deployed in the world, don't use LLMs or agentic harnesses, or diffusion models, or image/music/video generation
Dotcom bust - 2000
I guess this bubble has some room to run!
I came here to say same thing :-)
5 years ago it was the crypto bros quoting this
This seems like a poor comparison, for any given company, benefitting from personal computing in 1987 was a significantly harder task than using AI tools in 2026 is.
The first requires a fundamental shift in not only how you work but expensive and complex infrastructure changes from the company itself just to make it possible (the first Mac for example would cost about $7000 per person in today’s money and would be outdated in maybe a year or two), an office building of The latter requires asking a subscription service to do a task for you on basically any computer connected to the internet.
I don’t know about you, but for me materially useful AI costs around $10,000/mo in tokens. But it is very useful.
Claude subscriptions help, but only if I come in with the expertise to build the entire system around the model and I don’t mind the whole thing grinding to a halt when my session limit is saturated. Anyone selling the system off the shelf is going to be charging API rates. Computers were cheap in comparison.
I can’t speak for everyone, but I would intuitively venture to guess there would be some greater than marginal productivity benefits from say, a $200/m Claude sub for each employee in many organizations outside of tech.
Perhaps they’re all on the $20/m plan which is useless?
“Marginal” improvement isn’t really going to be enough. Products need to be closer to 10x better than the alternative to get traction (which can be divided across several dimensions). AI, I think, is just starting to get there.
I get what you’re saying but I’m just not a fan of this type of argument. It always paints AI as this “in the future” thing which absolves any current criticisms or observations.
LLMs are great and definitely make you more productive, for some reason 493/500 firms haven’t been able to figure that out even though it’s trivially easy to provide employees with frontier models and it’s also trivially easy to use frontier models.
Which is why I say comparing it to personal computers in 1987 is not an apt comparison. That took redefining physical workspaces, new infrastructure, and new personnel which wasn’t easy to come by. Becoming a company that uses AI is just as easy (if not easier) as using any other SaaS.
See sibling comment. 10x measured how? In pure coding tasks under certain conditions the speedup is huge, but seems really unclear how this would transfer to any significant speedup across the whole undertaking (project, delivery etc).
Maybe, even likely. But how do you observe this? We are in this situation at work, it feels like "sure AI is helpful" but do we actually deliver any faster?
Or just in enterprises where they can't have plans.
Holy shit. I’m spending 2 orders of magnitude less per month and still getting good value. But, I’m not an AI maximalist.
If it costs 2x-3x the median household income to be an AI maximalist that definitely puts some comments on this forum into context.
It’s a business-scale cost. When I say materially useful, I mean replacing multiple salaries and improving quality at the same time. Not some variant of “make an existing employee 20% faster.”
The personal hobby version is to get the Claude max plan.
One industry example.. We have been giving students computing devices for years with the cost of educating students going up year after year.. Yet scores keep dropping.. No productivity gains here.. I do not think these results are unique to the education sector.. There are probably alot of businesses where nothing much will change except slight efficiencies.. not enough to move the needle..
This is one area where we've proved to be fully unable to rise to the occasion.
The internet age opened a million of possibilities to improve education: improved visual explanations, scalable access to the world's best educators, lowered friction to access students to work with, replayable lessons...
We mostly just kept the old model and placed an addictive device in front of children.
Maybe it’s possible in theory to leverage our newest technologies as educational tools. But maybe it’s also incredibly challenging because it’s so easy to fuck up: a lonely peak of optimal outcomes surrounded by a vast space of net negative outcomes.
You mention the addictiveness of the devices, this probably plays a big role in shaping the outcome space.
Let's take math education as an example. If you look at the actual result, all math education past y=mx+b is a waste of time for most students. By and large, math students are taught by people who have memorized the quadratic equation, or certain operations with logarithms or imaginary numbers, etc. but don't understand it on any kind of deeper level. (Imagine asking a high-school math teacher to compute i^i. Almost none will be able to.) So it's presented as a set of random rules you have to follow to get the correct answer, with no underlying logic or justification. Most normal people understandably hate this, and even if they suffer through it to get a good grade they will not retain it and come away permanently hating math.
It's just sad. I sometimes help kids with math, and it's amazing how screwed they are. Like, smart kids with good grades who are totally unable to do basic manipulations with fractions. I mean questions like "write 25% as a fraction". They are getting absolutely nothing from math classes. How could they? Even the best teacher in the world can't pause class for everyone when they realize that one of their students is missing some foundational concepts. That student is just going to get left farther and farther behind, and they will never catch up. It happens to almost every student at some point. They'll get sick for a week or just have their mind on other things, miss some important concept, and then be left behind forever.
I honestly have a hard time imagining how we would possibly make the situation any worse with technology honestly. But if you just took Math Academy and made students use that instead of their math classes, I bet 99% would have better understanding and fluency with math than they do now.
I'm sorry this was your experience with maths in school. My experience was nothing like it, but it could be to do with where I live, my particular teachers or the classes I took.
> Imagine asking a high-school math teacher to compute i^i. Almost none will be able to.
Mine could. And of those who couldn't they would have leveraged the technology available to them (Google) to find the answer with me, and would be able to make it make sense to me. This is an example of technology bettering my education.
And people still think AI era will bring prosperity for all and we will to have to work and just enjoy ourselves
I just found a book about abandoned teenagers from the 70s (think juvenile held in post-offense rehabilitation centers) and they speech level is above the average educated adult today.
There's many breadcrumbs like these, of a change in culture globally. 50s social expectations about expression and characters were way deeper. And it gradually fades away as life becomes comfy and entertainment based (80s+) I'd say.
Check out the lead sheet for Stardust, the most popular song around ~1930 https://www.sheetmusicplus.com/en/product/stardust-19425819....
Vs one for Patient Zero, the most popular song today: https://www.sheetmusicplus.com/en/product/patient-zero-24016...
taylor switf is so awful ... yet so popular
This is what irks, and saddens me most about applications of AI not improving outcomes. I suppose I should take some comfort that even Salman Khan couldn't crack this nut.
The idea of an AI tutor that could step-by-step customize an education plan to each student and patiently apply every approach to getting the students to understand each lesson, is such an attractive vision.
Both of my kids chafed under lock step pacing of lessons, even in a school where disruptive or unmotivated students were pretty much nonexistent.
The education industry has been lowering the bar for admission to higher learning for half a century or more, while dumbing down the content of high school and even primary school education. All because we can’t have children feeling bad.
It’s terribly ironic that things like calculus, linear algebra and matrix manipulation were being weeded out of the curriculum, right before machine learning took off.
What you're talking about are specific local policy changes at specific school and equating random trends to some sort of ... "kids these days" message.
If we're talking about the US; then it's simply the very healthy and functional economic inventive of students taking out lifetime loans to pay school tuition for schools that have not actually risen in quality or format in 40 years. The more students; the more money to the school board! Lower those bar riiight down!
Lazy kids these days amiright!
I would echo that it’s local. I often poke fun at my SO and her sister for their absolutely trash tier quality education supplied by the Indiana public school system. I’m often flabbergasted by what they didn’t learn in primary and secondary schooling. They were able to go to decent universities, though, but it really opened my eyes to how terrible red states run their public school systems.
It’s not even a flyover state thing, if you just look at Wisconsin and Minnesota. It’s uniquely red state. Also Indiana department of transportation is terrible as well.
> while dumbing down the content of high school and even primary school education. All because we can’t have children feeling bad.
I suspect that it's a lot less about making children feel better and more about keeping the general population stupid and unable to think critically so that the ruling class can exploit them easier. The children of the ruling class are still given the opportunity to be well educated by private institutions that are overpriced to keep out the poors.
you're conflating a lot of different stuff together.
if someone knew a better way to improve test scores without cheating on them, we'd use it. it's a very hard problem.
on the flip side it is pretty obvious how construction and manufacturing can be automated and what the impact will be. it's so obvious that we already tariff and ban stuff that, from our POV, is automated, like prefab buildings, for solely political reasons.
This is an important observation. We've had 50 years of personal computing, but is there any evidence that computers in any form ever improved the academic performance of students?
AI certainly seems to be making students much less competent.
No, we know that when technology (particularly: Internet connected laptops) is introduced into classrooms competency declines. This has been shown to be true internationally
It's... complicated and a bit disputed. The trends match but the attempts at proper comparison study is not quite as conclusive (good social studies are difficult to make). It's certainly true that the trend is negative rather than positive.
I don't see why it would. It might make education more efficient like a factory in some ways, but I simply don't see why it would aid in learning. At least, from my experience as a student, here are some observations:
1. Computers have allowed me to gather more information, access more journals, more results from around the world, but what really seems to matter in learning is just sitting down with a tiny bit of material like one book and REALLY understanding it.
2. Computers make things more distracting because while they make typing up reports, proofs, and assignments more efficient, there's actually just too much information. What really seems to help is having LESS information so there's some space to just think and work things out.
The fact is, what computers do is help you keep up with other people in the field, so they might accelerate progress at the research level, but what they don't do is help you become a rigorous, independent thinker.
Academia in fact has always been an uneasy tension between the two: building capable individuals and getting results and money for progress. Computers help with the latter - they certainly impede the former.
Why would a technology that makes the mind work less, have benefits for a process, that works by making the mind work hard?
You still need to calculate calculus yourself long after you could buy books where the result is written in. That supposed "paradoxon" is not even specific to electronic tools. It's just that the point of education has never been to get results, but to exercise in as much as side paths as possible and memorable while following the path to the results. The result mostly get thrown away immediately.
I gave my son CRPGs for his GBA that required reading what the characters were saying. His reaction to finding out his three and four year old friends didn't read was interesting.
> Yet scores keep dropping
Not everyone on hn is American or part of the Anglosphere.
Scores are dropping all around the world. I don’t know why you’re suggesting it’s only happening in the US.
We can use a non-deterministic system because we have test suites that act as validation gates.
Other industries don't, so they have no way of knowing if the AI is bullshitting them outside of manual validation, which makes it more expensive and slower than doing it yourself.
It's a good point but in most of the places i've worked an average person could leverage any 2026 llm to improve things 50% easily. And I'm more anti-llm than not somehow.
In more technical places, having a llm write small automation would also add up.I strongly agree that search engine is the best application of LLM so far. 90% of time I don't look at Google search results, I look at AI Mode output only. And you don't need agents, harness, guardrails, etc for it — basic small clean LLM does the job well without overdoing it.
> no way of knowing if the AI is bullshitting them
Back testing, hold-out sets, cross validation.
This has all happened before.
On the other hand, if humans could apparently do that work without automated verification, then perhaps so can agents.
> they have no way of knowing if the AI is bullshitting them
That hasn’t stopped Tesla. Their self-driving cars work by connecting the car’s controls directly to a neural network.
And they're still a decade away from general road use outside of vetted streets. Tesla is still struggling with pedestrian detection in rain tests.
Your information seems to be outdated, unless every single street I’ve been on for the past year is “vetted”. All the other (consumer) car brand’s self-driving systems do have that limitation though, maybe that's where the confusion came from. Tesla is self-driving 98% of my mileage, with the 2% being when I take over to choose a specific parking spot or I’m towing/carrying something big.
Don’t know much about how it does in the rain since moving to SoCal but in clear weather it hasn’t come close to hitting anyone yet. Even proactively warned me a few times while I was driving. Maybe I should let it take over in parking lots too :)
Someone take this guy's license
It'll be hard, when I'm never causing any accidents or speeding or running reds/stop signs. Maybe they could get me for inattentive driving but that's what the tinted windows are for.
isn't that more a consequence of them using a cameras-only approach vs. Waymo/others who have other sensing-modalities to fall back on (lidar mainly)?
They have world simulators which verify, to an extent, the behavior of the software
I thought that pretty much every self driving algorithm in production was based on machine learning of some sort.
In most cases it’s only the object recognition systems that use machine learning. I know that Tesla was talking about using non-deterministic AI for the decision making as well, but it’s a worry if they’re already doing it.
Makes sense. Thanks for the clarification
Earlier Tesla self-driving software used search-based planning and conventional optimization alongside neural networks. Tesla has publicly demonstrated A* in its path planning work.
A lawsuit revealing part of the camera analysis showed them using a classic CNN based object recognition network.
For those who won't act ethically or at least reasonably, regulation is the answer. We're not too far away from flipping the congress into something far more willing to do just that, especially with anything connected to Elon Musk.
There are many tasks humans perform which have no or slow feedback loop. Even if they can audit/validate AI output, they don't know if it will be good when it reaches other humans.
test suites cover nowhere near enough to assure reliability
and if they did for your system, then it's so simple that whether or not you used AI doesn't even really matter
right cause test suites never have issues
Yeah and getting a single example of a correct program is way easier than describing the set of all correct programs.
"AI does not understand accountability, trust, culture, unwritten rules, ethical trade-offs or fairness, all of which shape reputational risk and ultimately drive revenue growth and profit margins."
For a sec I thought he was talking about private equity.
I'm sure private equity understands it very well... That's why they profit so much, they understand what to destroy.
If you (and GP) haven't already seen this, you'll at least get a chuckle out of it:
https://www.antipope.org/charlie/blog-static/2018/01/dude-yo...
We're not seeing full automation for things that ought to be automatable now. Customer service chatbots still aren't very good. Salesforce ought to be taking over entire marketing operations, not just supporting sales but replacing it. The medical office side of insurance ought to be automatic now. The pick and place side of warehouse operations should be robotic. Shelf restocking in stores ought to be robotic. All of those things have been demoed, but didn't go to production.
The thing is with a lot of the world, you could say about the same in 1950 about why isn't everything automated by now. A lot of the technology needed to do most of everything we do was already there. They had vending machines selling cigarettes back then and today people buy their geekbars from a person in the 7/11, and we think we've progressed. Factories today still mostly only semi automated. You'd think mcdonalds would just be a single unit deployable crate burger factory by now, but it looks the same as it did in 1950 for the most part with humans doing all the work.
No one is actually given the agency in any company to set about and engineer and implement an automated solution that replaces existing business units. They also have to be given slack to fail and iterate. Easiest to just run mcdonalds mostly as you ran it the last quarter lest you rock the boat too much. Rinse and repeat and people still bag fries today, and the deep fryer remains probably the most dangerous part of the kitchen.
That's an interesting point. Vending machines never became a big deal outside of Japan. Fast food was automated in 1964, with AMFare [1], but that was never deployed beyond one prototype.
Whether general purpose robots will change this is interesting. It looks like they're going to cost about as much as a car. (There are at least 18 makers of humanoid robots. The hardware situation looks good.) That's cheaper than humans. There are already companies that set up industrial robots and charge for them by the working hour, with all maintenance and setup included.
[1] https://www.youtube.com/watch?v=1Xop9py8zBY
> All of those things have been demoed, but didn't go to production.
There's probably a good reason for this. Demos usually just show the happy case, whereas in practice you have to deal with all the nasty edge cases of reality.
Yes, AI is still not particularly prevalent in the customer-facing / consumer-facing things you described. But internally, I think these companies are up to their eyeballs in AI automations. At the company I work for, the customer-facing side of our business has not really been changed by AI (except for faster iteration of website design, and a somewhat limited customer service chatbot - not particularly good like your experience). But internally we are using it for so many things, for example we have brought 3rd party analytics tools in-house (cost reduction and ownership over our data) by writing our own internal tools, we have setup custom systems to give ourselves better visibility over the business to try catch issues earlier, we use it to produce internal reports, etc.
I don't have much experience in robotics but I've heard that machine vision models have absolutely been changing the industry. And I imagine they are also using AI to write a lot of software too. But not my area, just what I've heard / been reading.
> But internally we are using it for so many things, for example we have brought 3rd party analytics tools in-house (cost reduction and ownership over our data) by writing our own internal tools, we have setup custom systems to give ourselves better visibility over the business to try catch issues earlier, we use it to produce internal reports, etc.
How much of that is automation of "bullshit jobs"?
Fair enough. It's a mix. But for the most part it's a net positive.
We have replaced external tools with in-house tools written by largely AI. These are not one-prompt tools made in a day. We have built our own analytics stack which non-technical coworkers use everyday - they are not interested in "bullshit", they need real, accurate data and visualisations. And we had a cross-over period where our internal tooling was compared to our (now legacy) 3rd party analytics tools. It was done properly.
At least where I work, I don't think there is much automation of "bullshit jobs" - people don't typically just waste time chasing fairy tales (for the most part). But for things like report generation or tracing issues or generating presentations or infographics, there is certainly a much higher volume of these things getting pushed around which may or may not actually achieve anything, or be at all correct. So I reckon a large chunk of these would fall under into the "bullshit" bucket.
Because everyone can produce so much more "stuff" with AI assistance / automations, a good portion does end up spewed straight out of the slop factory. How much / what proportion kind of differs person-to-person tbh.
Yes, that's because the tech sucks, unlike what the AI supporters claim.
And my car should fly
The reason the payoff isn't visible yet is because outside of software engineering, it isn't systematized nor integrated into workflows. But believe me, people are using it and increasing their free time. I was at an HOA meeting in Reno. The board president works for a major US cable provider doing "contract approvals and negotiations", whatever that means. This last year, he said he's cut his 8 hour days down to a few hours using Claude (pretty sure he is just using the web chat interface). Employees of the property management firm are using it too -- our representative at the meeting said it had been a lifesaver for her in recent months. We haven't seen the full benefit from their end but they are also in a state of constant turmoil as a company so it's a wash. A tennis partner of mine who is the regional (entire west coast) manager for an automotive company's extended warranty program has also been able to use it to cut down time-consuming documentation and the firm encourages it; but again, no direct integration into her workflow, she is just winging it with ChatGPT online.
I currently work in E-commerce, local online marketplace in Kazakhstan. I scraped all millions of goods in the marketplace, then analyzed top 30k goods with AI. Goal here to find it's supplier in China accurately and estimate weight, similarity, etc. That index gave me huge competitive advantage compared to other sellers and extremely boosting my income. But that cost a lot, because, it's literally like running new Claude Code session for each good of tenth thousands, with custom tools like search in alibaba, viewing reviews, etc.
In marketplaces margin/ROI of goods are distributed like(just for example):
Top 1% goods (P01) - ROI 1000%+
Top 5% goods - ROI 300%+
Top 10% goods - ROI 100%+
Top 20% goods - ROI 50%+
And so on. So such automation helps me to find that Top 1% first
It’d be nice to stop calling it The Magnificent 7. Perhaps The Circular 7, or The Shady 7, something more realistic about how they are treating the stock market. Yeah number go up, but the means to those ends are shady as all hell.
I think it's the opposite. The stock market is doing this to big tech. In 2022 these stocks were sold mercilessly until they announced layoffs. CEOs fell in line because their ultimate boss is the stock market. Then in 2023+ it was all about how many times you can say AI. Until 2026 where it's a little bit more "show me some revenue growth". These companies are not the ones controlling their own stock price. It used to be the case they bought back a lot of stock, which you could say is a sort of "manipulation". But now the opposite: they spend over 100% of profits on the data center buildout.
But "magnificent 7" foreshadows that 3 will survive and 4 perish, keep it!
Microogletazon, here we come!
We’re heading into corporate budget season. Company after company is slashing their AI budget and the head cut wagons are circling around all the various “AI lab” teams that every company invested in two years ago with zero ROI to show for it. All while massive debt piles up to build data centers despite little evidence any of this is coming close to returning the ROI needed to justify any of it.
2027 is going to be very interesting. Buckle up.
Even in the M7, I doubt AI is having much impact other than benefiting from AI spending. Its hard to point to any non-AI products in the M7 that have meaningfully improved since 2023.
The problem isn't that "AI doesn't understand..." this or that, as concluded by the article though. Its more a lack of creativity and incorrect reasoning about what AI can and cannot do. For one, whole role replacement is pure fiction and likely will never happen with LLMs. I expect most businesses will look very different in the next 10 years due to AI and it will not be smooth as the needed creativity will mostly happen in startups (and perhaps companies that can truly operate like startups) and diffuse out from there. I'd expect a lot of churn in the S&P493 and even the M7 most likely.
One way they seem to have benefited from AI is that it reintroduced the idea of paying for new software features. For many years Google just gave away everything for "free" (modulo ads obviously). Now Google adds a new feature and adds a little sparkle and says "pay up!"
I think most people in the tech sector and especially driving the AI hype have absolutely no clue how difficult it is to implement change in complex organisations. There are a bunch of technical reasons but half the reasons are not even technical - social, legal, etc. You have to shift the practices of humans and they have incredible inertia. Like decade level change response period. And during the change period they are less productive. Expect a decade type period before "productivity" impact manifests and even then it won't really be measurable directly because people will be doing different things.
Especially people who are attracted to startups are the people who don't want to deal with steering a giant company in a new direction.
> I think most people in the tech sector and especially driving the AI hype have absolutely no clue how difficult it is to implement change in complex organisations.
Simple organizations too, and sometimes especially so. The amount of unrealized productivity many small business stand to gain just by adopting simple automations that computers have been able to perform for decades may be as large as (or larger than) the productivity gains ML model and agent adoption enables. And that adoption may just be enabling someone who never wanted to learn Excel macros or regular expressions to command an agent to go do those things which have been possible yet unrealized in a particular small business for decades.
The problem has largely been (and remains) humans set in their ways, of the sort who stopped reading books or seeking any new knowledge whatsoever the moment their mandated or socially expected schooling was complete. Incurious midwits are definitionally the bulk of any workforce, and getting them to change any process is like pulling teeth whether they're working in a small business or a megacorporation. Change often doesn't come until the midwit on some particular throne retires, quits, or is laid off.
When I tried to use CoPilot at my day job, I set up a prompt to process scans of checks and invoices (with the MICR account bar code line covered w/ a pen so that there was no PII), it would sort of work for 20 files at a time to write a .bat file command to rename the source image based on the content of the check and invoice --- this removed the need to open up scans in batches of 50, read the Invoice ID, save using that, then run a .bat file to rename them based on the balance of information in the database.
The thing is, the failure rate was quite high --- often only 15 (sometimes fewer) lines were returned, and after a while, Microsoft imposed a limit on uploading more files, even though I was deleting the conversations immediately and quitting CoPilot and re-starting (long sessions seemed to have a progressively worse failure rate).
Since then, access to CoPilot has been removed and we're waiting on a test program to finish. Kind of curious how many tokens this was using up (and if there's a model which could be run locally which would work well and quickly enough).
I recently had a somewhat similar task: My 80-year old friend who is on Medicare got a $30,000 hospital bill which the insurance company declined to cover. So he wrote by hand a letter to the insurance company to request a reversal of their decision. I scanned his hand-written letter and asked AI to turn it into text (to create a professional looking letter). Could not get that to work on Gemini nor Perplexity, but it worked on Co-Pilot.
It did save me a lot time to use AI, especially should next time since I now know which AI app can handle it. So, there was a bit of overhead before I found the AI-app and prompt that would do it.
I have another friend who works for SAP as a project manager. She said she hates the fact that management wants them to use AI for everything, giving them a bonus if they do. But one reason I think their employer insists on that is that when you try to use it for real tasks you discover what works and what doesn't. Gradually we should be getting better at it.
Currently there is still a penalty caused by
1. AI not being able to do a specific task
2. Users not knowing which AI-app can do the job and
3. How to prompt the AI to get it done.
Writing the prompt is what users must do, and writing a prompt that works is a skill they must learn for every task you want AI do for you. Therefore progres is is happening but incrementally.
Enterprise sales somehow convinced a lot of orgs that all AI is the same or operates on the same level. Even frontier labs have differences, and you have to use each for your own advantage.
LLMs are not the type of model to use for this task. There are more specialized models that have much higher accuracy and full results. You can probably find the latest by looking around the OCR ecosystem.
A co-worker in a different program whose invoicing was much simpler and more straight-forward tried to automate this using OCR and Python and had a higher error rate and still had to set hand-written checks aside to handle manually.
I couldn't find an OCR system which was:
- free/affordable
- able to cope with arbitrary check sizes/hand-written checks and my occasional need to arrange multiple invoices in a stack with the check rotate 90 degrees so that all the numbers are visible
- which would capture the data from a scan in such a way that it could be easily used to rename the file
If you know of a system which is flexible enough to do this which is essentially free, let me know.
llm/transformers may actually be better than prior methods (not my area of expertise)
I did see a talk recently of fine-tuning a small LLM specifically for receipt transcription: https://www.youtube.com/watch?v=vDyxCHl-p8s
https://github.com/fw-ai/cookbook/tree/main/training/case-st...
HF has OCR LLM based models too, eg: https://huggingface.co/baidu/Unlimited-OCR
Paradigm shifts happen when old minds retire or die. Something like that, but for companies. If you’re starting a company now, you are _definitely_ hiring less people much later, and different profiles.
For sure. The original quote by Max Planck in his "Scientific Autobiography":
"A new scientific truth does not triumph by convincing its opponents and making them see the light, but rather because its opponents eventually die, and a new generation grows up that is familiar with it."
Admirers of this idea shortened it to, "Science progresses one funeral at a time."
Thank you, I couldn’t remember where exactly that was from, and was too lazy to look it up!
In video production, 3D modelling, editing etc, I think a lot of jobs are disappearing, with cost dropping. Not data backed, just from what I can see.
Makes sense. My RTX 5090 can locally take a text prompt, generate a 2D image from that, feed it into a 3D generation model, rig up the animations, and make me assets that are better than a lot of video games on the market. Not Cyberpunk 2077 level m, sure, but pretty good and this is just what I can clobber together in my free time from open models.
When something on that scale is not data backed, it’s called vibes and feelings.
Are the results in video generation precise enough to remove the need for 3D modeling?
I'm a dev not a video or 3D guy, but I'm looking at people I used to know and also some youtube doom-scrolling, the blender MCP, cinema 4D MCP, and seeing that it's a big impact
No
Current evidence is compelling us towards the shared opinion that AI is changing coding and nothing else… but do remember that AI itself is code and we are in an early phase of recursive AI improvement.
We have no idea where this will take us but it’s like measuring the effects of COVID in 2020. The immediate outcomes (temporary unemployment, massive stimulus) created long waves that we are still encountering (inflation, global instability).
I am in a tech role that straddles traditional business. Yes, LLMs are very adept at software development and of course we are early adopters, but the amount of other things that can be automated or at least significantly augmented hasn't scratched the surface.
At the start of this all, execs were clamoring for "AI" and at that time the models were superficially impressive but fell short. Now I think they are all underestimating the amount of improvement in the last year. There's loads of work that can be automated
This is the bucket I fall into. Seeing this everywhere and it makes a lot of sense to me.
I have always wondered how this technology is going to actually effect the majority of the GDP in our country. Example: How is this LLM technology going to benefit me and make me more money in my laundry mat? How is it going to benefit the tourism industry? How is it going to benefit the Hotel industry? How is it going to effect the Bar and Liquor sales industry? I dont see LLMs bringing significant cost savings in the majority of the industries that make up our GDP... It brings me the benefit of being a first google type search before I dive into mechanic manuals to fix my equipment.
I dont see having a morgan freeman AI voice saying announcements and me asking it for how to fix my machines being much better than what already exists in my industry...
I have worked in multiple companies where the internally developed software was central to the business success, yet it was handled in an extremely bad way, preventing it from using even a small percentage of its potential. The (good) developers were not the main problem there, but the management. The management did constantly override technical decisions and/or switched the main goals/vison around all the time, preventing the development from finishing anything. In several companies, they replaced the developers with ticket-crunshing code-monkeys that have no interest at all in the business at and who bring no innovation, ideas, or even feedback into their work. Some of that replacement was active, but most was by just making (work-)life miserable for anyone who cared what happened. In my case, one of these companies even offered me a big payment (one that I will probably never get anywhere else without switching to another country) for coming back, because I understand the business good enough to always fix the fires in the background that the management caused, but they still rejected any good ideas and made life worse instead of better for the people that dependent on the software. It's been almost five years since I left there, and I still have several friends in that company. One of them is one of the users of the software, another one is a developer who cares, but he could not switch because he is stuck there for several reasons. One of the main reasons being that he has medical issues that make most companies reject him by default. They keep me updated about the companies progress, or the lack of it. The latter guy told me this week, that the company made less progress in the last four years, than it did within a few weeks while the original dev team worked there. And he really wasn't exaggerating. Oh, the company has a big AI budget and they made pretty much every model available to all employees. They just can't use it for good ideas, because they will be garbled up by the management and turned around to be bad. As long as they don't replace the management with AI, this technology will not improve their business.
I wonder if this is going to be something like the Solow Paradox, from Econ Nobel Laureate Robert Solow: https://www.brookings.edu/articles/the-solow-productivity-pa...
>You can see the computer age everywhere but in the productivity statistics. Robert Solow (1987)
I dont even think theres a payoff in the software dev.... seems to be great to spin something up quick as a POC, then maintaining it requires people immediately who struggle with reading the amount of code it produces.
AI payoff in the tech sector so far is the loss of thousands of jobs, the disappearence of entry-level positions, a flood of low-quality software and more cybersecurity risks.
Thus far AI has been great at extracting value from every area its introduced in, but terrible at actually generating value.
There are three kinds of lies: lies, damned lies, and statistics.
I'm pretty sure most of the growth comes from companies laying off useless COVID-epoch engineers. If you lose the dead weight, you climb up without ever gaining more profit. And IT companies hired the most during COVID — the rollback is logical and it has nothing to deal with AI.
But this theory is harder to grasp than the sweet "growth is coming soon". Personally, I don't see any signs of long-term recovery after 2022, the US gov. debt growth is unsustainable, and the global market falls apart i.e. GDP growth while real markets collapsing will just collapse the currency.
'a.i' is unlike software before - where it was build then deploy with lack of customization except of a few high ticket items such as SAP, Workday etc.
for 'A.I' to work for the rest of industry outside of software - you need high touch customization.
I think we are still in a value discovery phase, as the token cost of one shotting someone's job is not cost effective, you have to constantly pay for tokens. Distilling someones role and breaking it up into AI responses and have AI build the deterministic workflow is where the efficiency is. It should be an upfront cost and then stable efficiency but I don't think many industries are setup to even begin this transformation. Someone else mentioned it's kind of like electricity, in itself productivity doesn't change, it's systems adopting it that compete and scale versus non adopting competition.
I think you could say the same thing of tech and the economy prior to AI.
No Signs of Tech Payoff Outside the Tech Sector. Manufacturing, basic science, education, etc languished while Big Tech reaped record profits and achieved stock heights for the last 15 years.
Let's not forget tech was in a major slowdown announcing layoffs left and right before AI picked up
End of ZIRP, tax law changes, post-COVID inflation, and general economic malaise from federal mismanagement.
There are too many confounders.
Hell, on top of that (i.e. the confounders you mention are very very apparent everywhere,) the challenge of folks understanding how to use AI to be more productive.
Even last week I had a case where, I was reviewing code that I know the dev was vibe coding, and honestly was good up to that review.
Buuuuuut, for the stuff I reviewed that day, there was some base-class inheritance use that was so ugly that my initial teams message to the dev was along the lines of 'What crack was Copilot smoking on that one'.
Turns out, they had asked some other random-ish questions in the chat they were doing their actual work in, and those questions more or less polluted the context.
Sure, I was able to guide them for how to not do that in future (i.e. keep unrelated questions around the work in question to a separate chat, etc.) but this was also a dev that is normally both very good, but also overworked semi-related to the general 'AI will save us on dev costs thing'.
I can't help but ask whether, for the even less-technical people that are told 'yeah you lost two colleagues but here is Copilot/Claude/etc' and they don't even get enough advice to have it actually increase productivity versus chasing down gremlins in the actual work.
What? Data please.
Manufacturing does not benefit from robots and modern machines? Sciene does not benefit from computers?
I don't follow.
Just talking about salaries, revenue, and stock IN AMERICA.
Doubtlessly Chinese manufacturing benefited tremendously from those advancements. But my point is the headline whinging about AI in America would also apply to Tech in America.
You do realize that the rest of spx also continuously achieved aths for the last $X years
It's almost like Big Tech is a net drain on everyone.
I think what people fail to appreciate, is that past a certain size, organizations and the people within them get so tightly wrapped up in working a certain way that it becomes nearly impossible to fix without rebuilding the whole thing from the ground up.
If you want to see if AI is paying off, look at new startups that build AI-native versus those that don't, controlling for the type of industry/product. This is a difficult causal inference comparison to make, but it's the one that I find most meaningful and interesting.
Maybe, but what the hell does "ai native" actually mean?
No signs of AI payoff outside the tech sector...yet. I think it's partially because our harnesses are generally much, much better. And engineers seem to be more deft at figuring out where AI provides real benefit and where it needs a watchful eye.
I'm seeing non-engineers use AI to send more emails, have more meetings, create more documents, and create more dashboards. So much of the busy work that happens in corporate environments was never that valuable in the first place. A firehose of busy work doesn't move any needles that matter.
There is nothing I dislike more than people using AI to generate tens of pages in text and then expect people to read them, when they did not read them themselves.
Well this is an actual effect. It is so easy now to generate presentations etc so as a consequence there are ten times as many that people seem to expect you to read. Productivity is higher but that which is produced sucks and is pointless. Actual value creation is harder than nice looking artifacts.
Normally productivity is measured in GDP per working hour, not powerpoint slide count.
Yes, my point is that when people say they are more productive, and that consists of more ppt slides, there is confusion and slop happening.
People use AI to have more meetings??
Oh yes. Copilot makes it easy!
Copilot schedules the meeting, takes notes during the meeting, chirps at you via teams during the meeting, provides a summary in Teams after the meeting parroting what you just heard, and then asks if you want it to schedule a follow-up. It's so helpful![/sarc]
I find it totally unsurprising and uninteresting that the tool makers themselves are the first to adopt the new tool they made.
this is literally the textbook historical process across every industry ever.
> the AI story will be decided in the S&P 493, not the Mag 7, and so far, there is no evidence that revenue growth or margins are improving outside the tech sector.
Lots of theories seeking to explain this reality, but that's all they are: theories. Maybe some will be proven right. Maybe all will be proven wrong.
As of right now, the economic reality lags far behind the hype. But down the road... I wouldn't place a bet either way.
1792936609 | Apollo's Slok Says AI Has Delivered Zero Payoff Outside of Tech | https://www.bloomberg.com/news/articles/2026-10-25/ai-provid... | https://news.ycombinator.com/item?id=50231661 | 1 comment
1792952284 | AI delivers 'zero financial payoff' to most companies, Apollo economist says | https://techcrunch.com/2026/10/25/ai-delivers-zero-financial... | https://news.ycombinator.com/item?id=50234408 | 1 comment
1792968546 | Apollo says AI has delivered 'zero financial payoff' outside the tech industry | https://www.businessinsider.com/apollo-ai-zero-financial-pay... | https://news.ycombinator.com/item?id=50236828 | 47 comments
And we’ve seen tech sector payoff?
I doubt it.
People are, on average, too stupid to make good use of AI to the point where you can actually gain massive productivity.
The pointy haired bosses are going to do their usual approach: force people to work more, claim success, move on.
Don't you mean that the pointy-haired bosses will force people to "adopt AI" because not adopting it is "stupid", then claim success?
They will do that, but then it won't work, so they'll blame the underlings and force them to work harder.
all I've seen are large tech sector bills for token usage and investor capital trying to make ai happen
Pretty much the same thing inside tech, I love AI for my personal projects but the productivity return in my dayjob has been slightly negative so far.
I would say the driving industry will get up-ended, but we'll see. It may just be workers get paid less because they only babysit the truck during very complicated times.
Robots still seem far off. Its just like war, air superiority might make the war winnable but its the boots on the ground that are needed to actually take the territory.
The problem with this analysis is twofold.
1. Even before AI, the U.S. economy was growing because of a handful of tech companies. The rest of the companies were collapsing.
2. If revenue per worker is stagnant despite AI investment, we don’t know what the revenue per worker would be without AI investment.
3. There’s no reason to believe AI should improve revenue. AI doesn’t just make companies more efficient. It makes their customers more capable of replacing their products. This continues all the way down to the individual. I might have thought about buying a calories tracking app in the past, but now, I can just use one of my AI Chat apps to do that for me.
AI might be reducing overall demand for services and therefore naturally contracting the exchange based economic system.
> The rest of the companies were collapsing.
5 year chart doesn't look like collapse to me
https://finance.yahoo.com/quote/VTV/
> AI may be a great calculator, but in business, intelligence is mostly judgment. Judgment depends on context, and the answer varies with which human or which model you ask. Perhaps most importantly, AI does not understand accountability, trust, culture, unwritten rules, ethical trade-offs or fairness, all of which shape reputational risk and ultimately drive revenue growth and profit margins.
All of this is true, but the reality is that LLM agents free up time for people to make judgement by taking care of the grunt work, which means each person can do more. That increased efficiency can mean higher revenues in growth industries and higher profits (through reduced labor force) in stable or declining industries.
Assuming the grunt work is not a component of good decision-making.
Yes, I think about problem as I write code.
So getting an LLM to write code is kind of self defeating.
Anecdotally this is not true. At $JOB I'm responsible for running the accounting function of a small US public company. With tools like OMP and shortcut my team has been able to significantly reduce the time it takes to prepare internal and public reporting documents. Not only are we getting more done within tighter deadlines, I have also observed a dramatic increase in the quality of the work product (not least in the form of fewer audit findings).
Maybe it takes a while to show up on the S&P 493's bottom line, but our token spend has had one of the best ROI as far as I'm concerned.
Is "time to prepare report" something that will actually increase overall productivity? Or does it really not matter that much?
Article is spot on!
The only issue is that the benefits would be more enmeshed with other gains in other industries.,
Heck, even in Software Development, the real benefits are not with mass labor elimination resulting in cost savings but with increasing available throughput. Technology if used to the fullest potential should be able to expand the proverbial pie., which is where the gains are!
That is not to say people won't try to save cost by eliminating labor due to AI., which is super shortsighted and misses the forest for the trees, but the overarching goal should be to increase the quality and quantity of deliverables, as demanded by the market.
> the AI story will be decided in the S&P 493, not the Mag 7
Maybe? But aren’t they the Mag 7 for a reason? Don’t they control a substantial part of GDP? Don’t they have a significant impact on the global economy that the s&p 493 may not - especially with large scale adoption of these tools.
Not saying that this means AI spending at current rates will be profitable or not crash - but only that there may only be 7-10 big companies at the end of all of this and the S&P 493 will be irrelevant (it’s one potential scenario in my book at any rate)
Have you heard of the SEC
Yes? How are they going to prevent consolidation of mega corps?
That is literally their entire charter?
And they have guns?
Do you mean the FTC? Typically the SEC is just concerned with trading regulation which would not prevent a merger.
Also I no longer assume any of these agencies will enforce their charters if the executive is sufficiently captured by companies as to allow mergers. For evidence I’d just gesture around at the latest batch of mergers.
A winner of the Nobel Prize in Economics, Paul Krugman wrote in 1998, “The growth of the Internet will slow drastically, as the flaw in ‘Metcalfe’s law’—which states that the number of potential connections in a network is proportional to the square of the number of participants—becomes apparent: most people have nothing to say to each other! By 2005, it will become clear that the Internet’s impact on the economy has been no greater than the fax machine’s.”
"most people have nothing to say to each other!"
He wasnt wrong actually.
Most people just watch and observe. As we see on Tikky Tokky and Instagram.
You can't use one successful technology in the past to justify every hyped technology from there on out
Wow this one specific guy was wrong thirty years ago!
Related:
Only 13% of companies see real returns from AI, survey finds - https://news.ycombinator.com/item?id=50207881 - October 2026
13% of all companies, or only those that invested in AI. And if invested, by how much?
is this 13% of leprechauns become elves?
Because AI is a tech invention and it's expected to be applied in a tech environment. As it cannot replace hand working, it cannot replace travel agents, doctors or lawyers because the latter put in human mental work that cannot be substituted yet and there is hardly any obvious chance in the near future.
There's a lot more to business in America than the corporate S&P. If you think AI isn't paying off for businesses, then you probably aren't a consultant right now. I have small/mid-size clients who, in pre-AI times, would have never been able to afford the work they're currently scoping out for me. Now I can quote them cheaper rates to build more custom solutions that they didn't even imagine was possible, and they are loving it.
This is a use case issue.
Right now, the most effective use case for AI is coding agents. Software development is a profit center for the tech sector, so it's not surprising that they're seeing big gains. In Corporate America, software development is part of IT which is a cost center, and the software landscape is 30+ years of bandaid-ed legacy systems, maintained by offshore contractors who submitted the lowest cost bid, so attempts to gain productivity via AI have not been as successful.
Where Corporate America has seen gains from AI in my experience is with customer support chatbots, but the customer service function overall in Corporate America is an unmitigated disaster, so the gains are hard to measure.
Beyond these two use cases, most people in Corporate America are just using ChatGPT to create slop and copy/pasting it wherever, and then goofing off in their newly freed spare time.
We need to see new AI use cases in Operations, Finance, Marketing, etc. that can handle business processes end to end before the productivity gains start showing up in the financials.
This kind of seems to me like exactly what a reasonable person would expect, based on the nature of the technology and the dynamics of the market bubble: getting plausible-looking-text-generators to do text munging on web apps and servers doesn't seem likely to increase base productivity, but the companies who can collect on the hype will make a killing.
This is not surprising. AI in its current form is more obviously useful at generating software than it is at assisting with other kinds of knowledge work. Software can be automatically validated. Validation loops are what let these systems iterate toward acceptable solutions. If you want AI to produce a document for humans, you need to carefully review and revise the output. In that case there's little to be gained, as reviewing plausible but subtly incorrect information sentence by sentence is in some ways harder than authoring it yourself (a quote from someone who worked in like city planning). Depending on the company/industry, some of the non-tech sector workloads still involve phone calls, parcels, faxes, etc -- things AI can't really do.
Yes there are technical solutions to such problems; I'm sure there's probably a startup to make phone calls that's hooked into your company's knowledgebase, multi-LLM workflows for document review, etc. But non-tech companies aren't adopting those. Tech companies came for the software generation and stayed for the general knowledge worker productivity. Since 2023 they made the investments to procure services, change culture, and generally re-tool. Software companies are also uniquely positioned to deal with the limitations of AI in-house, like by building custom glue components to link systems, building controls to mitigate risk, etc.
So basically the tech sector leverages AI more than other sectors. At a company level, non-tech has less investment, less training, less integration of AI with business processes. At an individual level this looks like workers having fewer tools, not knowing how to use the tools they do have, not knowing even what the tools' capabilities are, and incorporating them into their work only superficially if at all.
Someone asked me for help using AI at their job (which is tech adjacent). They were trying to create a process document using ChatGPT. Their selected model was GPT Luna Light (!), and they were copying and pasting data between ChatGPT and SharePoint. There _was_ a SharePoint integration but it lacked write access. The model was not aligning the doc images. Their prompts were not specific. It was a bad experience. Their conclusion was that AI was stupid and not worth the trouble.
So even though the non-tech sector can't take advantage of AI to the same degree that the tech sector can, they're utilizing it even less than they could be.
At a fundamental level LLMs reshape how we gather, shape, process and interact with information. Information is that which allows you to make a predictions with a probability better than chance. This will slowly change any industry where decisions are being made. So, it will touch all industries. And if we let it, all of life eventually. Better decisions, better survival.
"Perhaps most importantly, AI does not understand accountability, trust, culture, unwritten rules, ethical trade-offs or fairness, all of which shape reputational risk and ultimately drive revenue growth and profit margins" - This is perhaps the most consequential sentence in the entire article, and yet was almost mentioned as an after-thought.
When IBM PC compatibles first became widely available, most economic reports I saw said that there was no economic impact from computers becoming widely available in offices. Word processors, spreadsheets, and email didnt have any productivity improvements... and I think we know better.
It's too early and we are probably not measuring the right things.
Startups are where you can always look to see the future. The fact that startups are getting so many benefits from AI, I think, is suggestive that the rest of the economy will
Hard to tell though because:
a) So many startups are reselling AI, and
b) Many startups are VC-funded, and with VC investing significant amounts of capital into the space, forced collusion is definitely a possibility
I hear that a lot of language translators lost their jobs to AI ... I can't imagine this hasn't lead to cost reductions.
The problem is that very few people exist in both worlds, so most don't know how to use AI to streamline their workflows. They waste time repeatedly asking AI the same questions instead of having it write scripts to automate their repetitive tasks
I doubt Excel'ification happens this fast. To make it "payoff" it needs to become the next Excel-levels of usage.
This is like saying that the internet hasn’t created positive value for all but the biggest firms in 1998. Give it a second.
You people drunk on LLM kool aid forget that some of us are old enough to remember those years.
I do.
I worked as an installer for an ISP in 1995. Yes, it did exist. I remember the excitement of small business owners for getting it. No one needed explanation that instantly sending editable files over email is a game changer for many small firms. Previously you could either fax which was not editable or you needed to courier which was expensive and while overnight was possible it was even more expensive and still much slower than just email.
So yes, it created positive value the very moment you would open Eudora or Pegasus and press Send.
It didn't require hundreds of billions of capex. It required someone to set up shop with a PC, a serial card and a couple modems.
Ironic statements, given that you yourself are an LLM.
>> You people drunk on LLM kool aid forget that some of us are old enough to remember those years. I do.
>> It didn't require hundreds of billions of capex. It required someone to set up shop with a PC, a serial card and a couple modems.
This is a bad faith comparison. Those small business owners who gained access to the internet benefited from decades of research by military and research institutions, as well as massive amounts of infrastructure spending by telcos. That spending really started to pick up near the end of 1995. Between 1996 and 2001, North America collectively spent $1 trillion on the Internet buildout.
Many will fall for the road to be paved. This is the most advanced discovery for humans to date. There is no doubt it will succeed and is here to stay. It is still new (gold rush) and deregulated. A mine field for govs and businesses. We may see to big to fail crashing (.com bubble style or worse) but no one will ever stop AI.
Let's not forget that the gold rush saw a very small percentage if people who got rich, to those who lost money. It was the people selling a pickaxe and a dream who did well.
Indeed well said, but gold is adopted worldwide now and many commerce it daily. I see it as a rewardive opportunity, some expect economical benefits, some others simply health and general knowledge, Me... breakthroughs :)
That's fair. I think over time (assuming no Terminator scenarios) it will be a net positive, but that is also assuming it doesn't crash the economy when demand doesn't reach expectations within their timetables, or it gets cheap enough that it makes more sense to run locally. I fear it may end up closer to what happened with fiber laying in the 90s. Lots of players building lots of expensive infrastructure, only to see tech improvements make it worth pennies on the dollar. I'll be keeping my eye on eBay for new gear in a couple of years.
It was a gold rush because gold was already "adopted" in the world for a mere couple millenia.
Thats very muchso “ends justify the means” thinking, except im not seeing an ends worth working towards.
Also, the woman who has an AI child with Gemini said not to support artists who hate on AI, and I agree. We must crush the haters before it's too late, for they are irrational and will be left behind as we go vertical any second now.
seriously though: nothing will stop the sky from being blue, anytime soon. How I know this? Because we don't constantly talk about how the sky will always be blue. If you got the goods, you don't need the sales pitch.
fascist thinking
Love it when some write nonsense just to be heard or had a bad day, maybe simply ignorance! Fascism comes from the latin word Fascia, which has nothing to with this topic. It was implemented later on as political philosophy, which still has nothing to do with this topic. The author of the article is trying to make a point which is correct in micro but incorrect in macro/long run. My claim is that AI is working and is here to stay (but always as a tool in our hands). Do you understand now how ignorant you sound? You got this champ, accept this critic.
> but incorrect in macro/long run
That you don't know. Nobody knows what AI is going to be capable of in 2 years. It could continue improving, or not. There's no strong evidence in any direction. You're just ignoring what doesn't fit your narrative.
here's an idea: take your stupid hand-wringing over definitions and shove it directly up your ass.
when you say "Many will fall for the road to be paved", you are saying your idea is more important than human fucking lives. this is a recurring theme in every fascist regime throughout human history.
Fascism is an phenomenon that arose in early twentieth century Europe. Historians can tell you exactly how. By trying to describe regimes earlier than that, "throughout human history", as fascist, you are only engaging in that very semantic bleaching of the term that makes it useless.
Of course there have been regimes throughout human history that were unfree, even brutal, but no need to apply a term so linked to specific early twentieth century developments.
nerd
This is the wrong read. Industries outside of tech are weaker in tech, duh. But, they are currently experimenting with the technology, and it’s also less good at non coding tasks, but getting better.
Also adoption is less pervasive.
But wait a few years and you will see a big payoff.
If debt doesn't come from useful productivity, how does it get paid?
"Tech" has been using LLMs since inception. Let's say 2022 for ChatGPT. Bank of America only just four years later, in October of 2026, managed to get some sort of AI (Copilot) to their employees to use. It's not showing up because it hasn't happened yet.
There's no payoff for tech either. AI has accelerated the accumulation of tech and cognitive debt that, once the technology is priced appropriately, will come due when it gets too expensive to continue to use.
„In SNP500 outside of tech”. This says little about general economy - and small companies are way faster in benefitting from CC.
Also, it’s been just 9 months of modern claude code, and for many companies less than 6 months since first employees outside of IT got access to Claude Code. What do they expect?
The rhetoric has basically been "replace 75% of white collar work in two years" so zero impact after 9 months surely seems.. disappointing?
Jevon’s paradox. If you make work more efficient you may actually end up with more of it.
I thought Meta Muse, OpenAI's Dot are solving consumer problems. Isn't it? For me most of Google searches are powered by Gemini & I interact with Gemini with follow up questions - completely non-tech.
What problems?
I have to wonder in this: how much of the revenue per worker boost in tech due to massive layoffs and coasting on momentum? Most tech workers I know do not describe a sustainable new environment with AI, but rather over-worked skeleton teams
Small sample size, but I've heard from multiple physicians that AI note taking allows them to see more patients or spend more time with them and they're very happy with those kinds of capabilities that take over toil
I've been doing workflow automations for many organizations for the last two years, and here is what I see:
1. every business unit and department was neglecting a best practice due to lack of investment in their organization. it has always been that way for decades and always will be that way because it isn't economically viable for it to be any other way. the workflow automations and agents we create act as digital workers that either do that neglected part, or let the rest of the organization do that neglected part.
2. the whole company always starts the year off with a bunch of goals, and by Q3 most of them are cancelled. now less of them are cancelled.
both factors mean that anyone just looking at "worker productivity" or "did we save money" is going to get the answer of no. everything is costing you the same, or the AI itself is costing you more. whatever contract I'm working under is definitely going to cost you more, if you don't cancel all those enterprise SaaS seats we told you to cancel.
I see a path to savings, but I don't think that's necessarily the metric to focus on. you need to look at the whole corporate structure, why are you still doing certain things, because those pressures will just make the org do more things because nobody is empowered to say "wait, maybe this is dumb"
“Tech sector is the first to adopt new tech” isn’t a groundbreaking revelation. I’m sure you could make similar charts for the computer, the internet and all other similar innovations.
Are people on here still arguing that these tools are not a good ROI for SW engineering? IF someone has a good argument for it I would love to hear it
Well the I is like $4T so the R has to be pretty damn good
I think honestly, tech really doesn't help the core business case in a lot of instances. Imagine the local restaurant scene. Would sales really decline if you couldn't order takeout on grubhub? I suspect if a switch were flipped and there was no grub hub et al, sales would be the exact same. Demand for takeout has nothing to do with the manner it is ordered; you are hungry and don't want to cook and that's it. It's a ten thousand year old industry just from that.
Businesses end up buying into all these services because thats just what is done. It's keeping up with the joneses. It isn't clear that they really lead to more sales revenue and improved bottom line. Even taking on orders using the system doesn't imply that; your customers might have just been shunted through a different ordering system instead.
And that is with an actual tangible service being offered. What does AI offer a restaurant? Outside code gen it's basically fancy google. What are they going to do, prompt it 'how can I make more money'? That would probably offer the same utility as just searching the internet on similar keywords relevant to your small business. Not really a value add that isn't there.
This is a pretty pathetic analysis especially for an analysis from a firm like Apollo. A few notes:
1. Profit margins from the mag 7 are on an upward trend starting in 2020. This is pretty consistent with the insane topline growth had by covid and the most recent layoffs occurring over the last 6 years. To unilaterally attribute this to AI implementation is lazy.
2. Most non-tech firms are only implementing AI just now because we finally have better enterprise tooling required. We won't see the effects for another 12-24 months at least.
3. As a former management consultant, this might be the worst use of a stacked bar chart I've ever seen: https://www.apollo.com/content/dam/apolloaem/images/daily-sp...
4. I don't know how you actually isolate AI productivity gains on a macro level. It seems pretty much impossible.
This whole thing is laughable.
The first giant web market collapse was because no-one could work out how to make money from it or do anything more useful than put the company brochure online.
no signs of mega tsunami damage outside of the shore - "News reporter says in the 50km radius shadow of a tsunami large enough to block out the sun"
Plumbers, electricians, car mechanics ... are probably making good money fixing messes that people got themselves into by following AI.
The best training comes from LLM users. Loads in software to train on.
Not so much in accounting, which is a much easier problem.
Software devs eagerly trained their replacement.
Well, writers can use it to speed up the draft review process. Not a big financial deal, but it's a 'payoff' for the craft.
My sector used it to make all the tools we wanted, then put it back in its box awhile ago. Just enjoying the fruits rent free.
Tell that to legal firms and patent guys
Ye, it was easiest to steal all of stackoverflow and sell it back to you.
Payoff doesn’t matter. The future potential (payoff) is what matters.
This is not just "Outside the Tech Sector", but "outside the magnificent 7"
"Tech sector" is supposed to be much bigger than just 7 companies, no matter how big those 7 are.
This looks to me more like just a huge centralization of wealth.
You won't get magical productivity gains from a workforce heavily conditioned to following orders and a pre-packaged process (the Henry Ford model and its various bastardizations). In order to extract value from an LLM, you have to be able to think and ideate so that the inputs are of higher quality.
It should be of zero surprise that the rush into AI has yielded limited results (if we're honest, both in and outside of tech)—you won't get a bunch of artists or creative thinkers out of a pliant, ironically agentic labor force. That's why most people say "it doesn't work": it's because they lack the skills and knowledge to use it properly (but corporations want to delude themselves into thinking they have a genie in a lamp).
Just a thought, but I kind of think this undersells just how important technology is or eventually becomes in almost every other economic sector.
"Productivity growth" just means same revenue for fewer employees. It would be ironic if that were the measure. (And no one argues that every business would benefit as expected when using the entire index as a signal.)
AI in medicine has become pervasive, resulting in less time spent fighting IT or staying up-to-date and more time with patients; no one need be fired because you can always do better work.
It's the opposite of enshittification: quality has gotten better without any visible economic difference.
I think this is misreading the data.
We are functionally in a recession. The only way this isn't true is because of AI investment [1]. Additionally, GDP itself is flawed because it is propped up by housing, specifically imputed rent [2]. Another ticking time bomb is the $4T private equity backlog [3].
The real product of AI of labor displacement and wage suppression.
So while this article sees profits flat outside of the Mag 7, it should really be in decline to being flat is actually an improvement (IMHO).
We're not producing anything anymore. All we're doing is driving up revenues on inelastic demand on things like food, housing, electricity and health care. That's all PE is. And that's all our economy does now.
I really think people are underestimating just how close to breaking point our society is.
[1]: https://uhero.hawaii.edu/would-we-be-in-a-recession-without-...
[2]: https://www.taxresearch.org.uk/Blog/2023/02/14/10-or-gdp-is-...
[3]: https://www.notveryprivateequity.com/eight-year-hold/
> AI does not understand accountability, trust, culture, unwritten rules, ethical trade-offs or fairness, all of which shape reputational risk and ultimately drive revenue growth and profit margins.
This reminds me of the book The Redemption of Time. An alien society incapable of lying has no other way to understand deception than to study an actual human mind. The different ways that we are capable of thinking can be so hard to capture in an artificial system.
It's too early. There was a time where economist could not find productivity gains from PC products. I think I remember reading about it in the early 2000s over a decade after computers were common in business. Part of it is that companies aren't quick to throw what's work for tech that hasn't fully proven itself.
AI productivity gains are coming make no mistake about it. My guess is that we'll see them within the next 5 years.
It's almost as if profit isn't directly related to your jira board.
I would expect new disruptors to eat the lunch of the old, but that's slow.
For AI to be properly integrated into businesses it would need to be enabled by people who would shortly after become irrelevant.
This may well remain a "true communism/socialism has never been tried" situation.
There is two ways of looking at this - one is that things are overhyped.
The other is that there is so much more value to be extracted before this bubble burst.
Not sure which I believe, objectively. Subjectively, inside of tech, it certainly feels like the former.
“Despite massive investments and widespread hype around [THE INTERNET], the technology has delivered zero financial payoff to corporate America outside of [THE PORN INDUSTRY]”
(1998 version of the story)
Honestly, is there really a good return in the tect industry?
The economics need to be rethought. What even is capital anymore? A ledger entry, compute, a license, IP? We are getting returns on what, setting the initial conditions?
All that’s left is offence surveillance logistics and insurance for a breakaway civilization.
The problem with profit in the tech sector is that they reduced payroll to boost the stock and enrich the single CEO.
They don't have the people. That is why all the big players keep using financing nonsense to buy everyone up. Killing competition is all the monopolists have left.
That's becuase eventually tech is going to eat the world. Everything will be augmented or replaced. Farming, construction, transportation, health, war, labor, all of it will change. So it has to start in tech and then it spreads from there.
The entire economy will become part of the tech sector so this is kind of self-defeating.
Payoffs will come, but most likely it will be when new companies start up from scratch in other industries, not from current incumbents somehow adapting quickly to advances in AI. Give it a generation, and all will be transformed by then.
The car dealership I bought a vehicle from is using AI for all appointments, and it is quite highly functional, voice operated.
Suggest Apollo is not doing a good job of measuring improvements since these are soft costs and harder to measure. But, instead of saying that, they are just repeating a low IQ take.
Let's be clear: there are no signs of payoff within the tech sector, either. The supposed productivity gains are minimal to non-existent, unless you completely give up on quality. And that's the best case, the one they've put the most effort into! LLMs do not deliver what they promise, it is very clear based on real world experience.
Which is to be expected. Anyone who makes a tool without targeting specific use cases is really just making it for themselves.
What's embarrassing is how many people are clueless about where AI is going. The tech sector is one giant echo chamber, so it thinks whatever it makes will be a miracle cure. Everyone outside the echo chamber has no idea what it is/how it works, so they're just hoping it's the miracle cure the echo chamber keeps saying it is.
In reality, AI is just a minor improvement in productivity that doesn't solve hard problems. It does what we could already do before, a little faster & cheaper, without determinism/trust. It's a bigger, faster truck, but somebody still has to drive it, find a place to park it, spend more on gas, etc. It may be able to "think", but it won't replace humans until it can accept liability.
> the technology has delivered zero financial payoff to corporate America outside of tech.
I wonder if the argument cannot be made that it has delivered little payoff in tech as well - in terms of actual stable revenue vis-a-vis the immense capital investments, the deep loans, the circular commitments etc.
It's such a trashy deceptive blanket statement "has delivered zero financial payoff to corporate America outside of tech." - literally I know at least a few companies myself that drastically benefit from it and have lowered their operational cost by a lot.
That's just a false statement.
Those with the payoff might not be sharing.
ctrl+f: rsi, recursive self improvement - 0 hits. Really?
AI labs are focusing on automating and accelerating AI research, because that will accelerate AI development in other fields. Claims that non-tech topics are harder to train for are wildly over stated in this thread. Labs are focused on beating each other to RSI, and training on tasks that do not support that are a lost opportunity cost.
Or inside for that matter
“Internet may be just a passing fad as millions give up on it”
You are missing the point: GenAI is here to stay. When and if it will bring a positive RoI for anyone other than those in related businesses, remains to be seen.
Internet has allowed a new type of businesses to appear and prosper. Some old-style businesses got the hang of it, some did not (look at all the newspaper businesses trying to figure out how to stay funded without falling under oligarch's rule).
the amount of copium here, things are slow to move we get it, but thinking that AI is not going to disrupt everything is... some next level denial
I think the question discussed here is more: it didn't disrupt everything in ~four years, will it take another two, five or twenty?
For now, that is...
lol, you are, sooooooo, blind, and iditic. Enjoy each day, then we, you are dead, vaporised, or some life shit. What a load of bollocks. Full of shit. POinltess drivel.
great. let's pop this bubble so everyone panicking can leave us devs alone with our tools. GLM 5.3 is good enough if anthropic and openai go under (which they wont).
What about the flood of LLM generated books on Amazon? Maybe sales of slop aren't being systematically tagged as such?
Funny to see some people compare the AI rollout to the industrial revolution.
The industrial revolution produced physical goods at scale, which eventually gave regular people access to things they could not have before.
AI produces intellectual labor at scale. Regular people do not need mass-produced content in their lives: there's already too much music, too much TV, too much software that only seems to get worse year after year.
Who needs mass-produced intellectual labor? The owning class. This is the only population segment that stands to benefit from this revolution. It's the reason every tech CEO is trying to shove AI down everyone's throats. It's the reason they're resting their feet on the table while everyone below them is starting to panic about the future of their careers.
I don't believe we will see compounding acceleration of anything other than spam and enshittification in this AI-laden world. This is not technology for building futures we can barely dream of: it's a gold-plated gift to the oligarchy and a fuck-you to everyone else.
I think we are struggling to achieve payoff within the tech sector.
There have been wins with the customer facing side of the business, but the development team seems to be the biggest problem right now despite the direct & obvious applicability of these tools. I know they aren't a special unicorn edge case because I was very rapidly solving their issues before they stopped handing me things that revealed the lethargy.
Within the sector where these tools definitely work, the biggest barrier to effective adoption is figuring out a way to undermine the various ideological high grounds. I have found the most effective argument to untangle this knot to be something approximating:
Challenge:
"Why should I adopt or help to implement technology that might eventually replace my job?"
Response:
"Because refusing to won't protect your job. It puts everyone's job at risk."
It doesn't matter if you protect your job (not using AI) if the entire company goes to the bottom of the ocean as a consequence. The neat thing about this is that your job is part of the set known as "The Company" and if that ceases to exist, so does your job. Good luck suing a dissolved entity for unemployment or whatever. There is enough competition in most markets to punish anyone who thinks they can win on virtue alone.
I personally have no problem per se with AI rendering certain tasks or even jobs economically redundant (although I am sceptical about many of the grander claims). What I do find tough to take is that the benefits will accrue to companies who have effectively stolen (and continue to steal) the intellectual output of everyone based on some bullshit fair use ruling for AI training. Hopefully the latter will get walked back and mechanisms and systems will be out in place to more fairly reward human intelligence.
> Hopefully the latter will get walked back and mechanisms and systems will be out in place to more fairly reward human intelligence.
Is there even any actual basis to walk this back at all anymore? It kinda seems like its a done deal and this is the outcome.
If the options are 1) watch the economy become more k-shaped as my status erodes and I'm made precarious 2) drag down aggregate output but be better off relative to everyone else than in the first scenario, why would I not go with option 2 from a purely self-interested, rational perspective? I jus think this is assuming a level of selflessness most people are not capable of but maybe that's too cynical of me
> It doesn't matter if you protect your job (not using AI) if the entire company goes to the bottom of the ocean as a consequence.
The vast majority of entire companies are going to the bottom of the ocean regardless of what individual contributors do.
Once you can replace nearly every job at a company, the company doesn't need to exist. Companies are just collections of experts, if AI replaces all the experts then it replaces the whole company.
So I've adopted the accelerationist mindset of: yes, this is absolutely going to replace my job, but it is going to replace everyone else's job too so fuck it I'll use it and we'll all go down to the bottom of the ocean together.
tl;dr you are economically worthless and should give up
What a grim outlook.
Trickle-down AI economics here we go! Suuurely it will work this time, any minute now don't worry about it.
We're a little early to be calling the game yet.
The incredible improvements have only been over the last 9 months or so. Business, especially disrupting entire job functions, takes much longer than that. As is typical with this type of change, it'll be slow until a tipping point, then FOMO will hit.
If I were to count the number of times you all have said "incredible improvements" here on this site over the last 7 years, what number do you think I'd come up with, and would that change your tune? I'll do it, if I must, to show you all how un-serious is this claim.
If anything the expectation of incredible improvements just over the horizon makes an argument for not adopting AI or at least doing so very judiciously.
Why would you spend dollars on tokens today if you expect the same tokens will get you something vastly superior in a few months, or that they'll do the same job for nickels on the dollar? Do you just hate having money or what is the problem?
What other changes have been like that? Tipping points.
Weird, AI enjoyers have been calling the game for years. We're already supposed to have achieved AGI and already be in the midst of the technological singularity. The tipping point was when AI solved the Navier-Stokes problem. All human created art, media and software development has been declared dead a dozen times over.
Yes yes, AI has only become really good yesterday 4:30pm, you guys don't get it, it's literally over for the haters, they're not gonna make it.
Yeah because vibe coding has only really worked for 3 months. Give it a chance!
It'll probably be a couple of years before most companies even realise it's an option.
Vibe coding as in "Make me a postgres" is poor but if you think it hasn't affected CRUD feature work you would be wrong.
Most developers don't work building low level high performance scalable systems for a large tech firms, they are making CRUD back-ends for phone apps.
This section of the industry is definitely contracting. Even more obvious is the dev tooling part of it. Dev tooling companies are disappearing or being "bought" all over the place because what might have been a product before is now an hour or two with someone with CC.
As someone in what is relatively the equivalent of Raven's Call of Duty mines compared to past work...
Yeah, 100%. Even for relatively complex scenarios (i.e. over-engineered for purpose) as long as the engineer can prompt coherently and tweak it's a big increase in productivity, at least for those people.
The gap is in things like contract houses where you might see better code on code review from the people submitting but if you actually use the tools, you know that the hours billed are even more inflated than they might have been before.
>but if you think it hasn't affected CRUD feature work you would be wrong.
the mythical man month lesson holds for CRUD apps as well. I'd be careful to equate churning out lines of code with the health of a piece of software down the road, even a relatively straight forward one.
Keep in mind in any real world software what matters is maintenance of a system, not just how many features you can add, regardless of human or machine generated code, some badly glued together codebase that's rapidly growing is going to haunt you.
Which is also why replacing sound dev tooling and with a magical prompt genie is incredibly irresponsible and short-sighted. You could cheaply produce code by outsourcing it years ago.
The problem with this argument is that people have been saying "it just recently got good" for about two years now.
Have they? AI agents only really got usable around last fall. LLMs were pretty useful starting basically at the beginning of 2025, but not in the fully autonomous way that they are now.
Yes, and each iteration of your comment picks a different date and a different 'killer feature' that 'made them usable'
Perhaps it is different dates for different people then
I keep hearing this. People being premature before does not mean they are premature now.
How long have people been warning about global warming? Do you still think it isn't real?
> People being premature before does not mean they are premature now.
It does not. But it does mean they are bad at predictions and you shouldn't give their latest prediction much credence.
I disagree. I was totally on the "vibe coding doesn't work and it's stupid" side a few months ago, when that was definitely true... So technically I was right, but I would say the "coding is solved!" people were more usefully right even if they were 6 months early. That's really not very early and it gave them a head start in preparing for today's reality.
Six months before the start of WW2 would your rather be "we're not at war, don't panic" or "peace is over, stockpile food"?
I know of a brake parts factory outside dayton ohio who had the Microsoft folks come through with promises well over a year ago now. They've already overspent and had to pull back their AI budgets when it made more work than it solved.
> before most companies even realize it's an option Couldn't be further from the truth
Seven years ago (*!!SEVEN!!*) GPT was too dangerous to release and yet everything that AI has touched since then is WORSE now than it was then. "Three months" is entirely nonsense. Seven years. Seven. S. E. V. E. N. Y. E. A. R. S.
If you moved the goal posts any farther, they'd have traveled around the world and would be sitting at your back right now.
Vibe coding still doesn't work today.
It did though. I've vibe coded three programs now. Two are replacements for existing programs that didn't work as well as I wanted and were missing features. The vibe coded ones work much better.
The third is definitely not as good as the commercial version, but it also doesn't cost $100k/seat/year.
Unfortunately work paid for the tokens so I can't share them, but I've seen it with my own eyes. It really works now. The code is not slop (and I am a perfectionist; it's not just that I have low standards).
It doesn't work with pre-Astra models for sure though.