> The Atlas Interfaces drift alone in the endless void. They are silent. They are unknowable fragments of an ancient whole. Yet their imprint on time and space molds our existence. They are the equation, and life is its answer. Through their Monoliths they give understanding to their boundless meaning, and that of our own.
I think the unfortunate reality is that lots of companies in our industry have suspiciously inflated employee counts in the first place. Even when removing AI and the pandemic over-hiring, I wouldn't have been surprised to see corrections sooner or later.
Employee count seems to correlate to stock market incentives - which is how GitLab is like 5x larger than Valve.
> I think the unfortunate reality is that lots of companies in our industry have suspiciously inflated employee counts in the first place. Even when removing AI and the pandemic over-hiring, I wouldn't have been surprised to see corrections sooner or later.
It's funny to me that everyone talks about pandemic over-hiring, as if this hadn't been a thing for a decade before that. Like, when I started at FB in 2013, they probably had about 75% of the engineers they needed (and about 10% of the sales people). But they grew engineering much faster than sales for some reason (engineering in the broad sense including data and product).
That being said, it's easy for people to look at other parts of a company and think they are over-staffed, because we don't see all the work necessary to keep the balls in the air.
AI aside, if we just look at other engineering disciplines in mature sectors, the future is not bright(er).
No fully paid golden cadillac benefits packages (for your dog too!), no twice daily uber eats comps, no $150k entry level, no unlimited PTO, no 6 months leave.
If you go to your uncles engineering department at the boiler company, these guy's engineering roles are about as pampered as the warehouse managers.
The upside is that it will cull those in it just for the money/lifestyle, and concentrate it down to those in it for the love of the craft.
Unlimited PTO works great for someone like me who regularly takes half day and full day PTO every quarter for various commitments/life stuff. But I’m also a guy and I have no problem asserting myself/spending my benefits. People who are a little less comfortable with that end up just parking on it and would do a lot better if they were given a set number of PTO days. The ambiguity hurts them in the same way not having set raise schedules and discussions hurts them.
Edit: folks, I can assure you I take a lot more than 10 days of PTO. I didn’t even say how many half/full days I take, you’re assuming a lot though I get my language could’ve been more clear there. I probably take between 20-30 days total annually. I am saying that because of unlimited PTO, it is very easy to just grab half days and full days as needed routinely and not worry about my overall count.
Yeah exactly. For some of us that is not a problem but for most people it seems to create issues. Company culture also plays a huge factor in that. Grey areas tend to hurt employees the most!
To be clear we still get federal holidays. So I net probably 30-40 days a year.
Again, the system works for me. But it is also a constant negotiation/workplace culture consideration, so I am not advocating for it writ large. I’m just lucky I work at a company that doesn’t put pressure on people not to utilize their PTO.
At my current job I get 25 days of PTO a year, with a cap of 30 days, plus company holidays (12 or so?), plus 5 days the entire company shuts down. I have to be careful not to hit my cap even taking 4 weeks of vacation because it’s still less than I earn. You have “unlimited” pto and take half of the vacation I do. If you ever change jobs you’ll get $0, while I’ll get a bonus.
You have it backwards. You’d benefit tremendously from fixed PTO that pays out because you’re taking fewer than 10 days a year. Its biggest benefactors are low-seniority employees who take 20-30 days annually.
It's more than that. Unlimited PTO in practice also results in people taking less time off.
Some people have to be reminded to take time off for themselves (like me in my early years), and without a predefined amount of PTO and an expiration mechanism, there's nothing stopping them from going extended periods without taking any time for themselves.
Most people grow out of that eventually, but their gems turn black before they fully advocate for their own work/life balance.
The other side to that coin is that when employees take varying amounts of time off, it becomes easy for managers to pressure anyone taking more time off to conform with those who take less (and sometimes none at all).
Yeah wtf? If programming becomes a generic white collar job, it's not really accurate to characterize the people as "those in it for the love of the craft". After all, do you think the average accountant or analyst is doing it "for the love of the craft"?
You don't need to look at other engineering jobs, just look at software engineers out side the US. We make decent money compared to the local market, but we've never had the royal treatment that US devs seem to get.
Exactly this - the late 00s and 10s we were riding the wave of the internet and smartphone adoption. Those markets have matured now and this is ultimately just the response. Covid was like some weird last hurrah for irrational spending. LLMs are just where all of that silly energy is going.
Well they keep breaking main FB page with additions or redesigns nobody asked for.
I am in Europe but I must be some test bunny that made some personal enemy there somehow, since beginning (cca 2008) FB was by far the buggiest site ever. Uploads of images didn't work, sometimes it uploaded twice. Uploading album of 30-50 pics took on average maybe 10 attempts in the past. Comments didn't work, or twice again. many similar behaviors. Ie since cca 2 weeks now if I click on some picture to have it full screen, clicking close after 1-2 mins ends up in FB error page instead of returning to main feed. I keep seeing FB error page a LOT considering infrequent use.
So, how they hired salespeople was based on revenue per head. If there wasn't enough revenue in a particular part of the business, no new hires for you. Obviously, that's gonna be a lagging indicator.
And they were honestly leaving a lot of money on the table there, but honestly not as much as they were leaving on the table by having a really, really, really bad and buggy self-serve interface that made it hard for small advertisers to spend money.
When they fixed that, the small businesses all got good results (e.g. hairdressers etc) and those people then convinced all the agency people to put more money into the platform.
Up till 2015 or so, it was a real struggle to convince people to even try FB ads, but after that it was like shooting fish in a barrel.
Also, up to 2013 or so (when feed ads became a thing), most of that money was coming from payments for f2p games, so they had lots of partner managers there, but not enough ads sales people.
I’d say it differently though. The company who hired all these people are finding they don’t have enough value added projects to keep them employed. Most of the FAANG and larger VC backed companies have something in common. They typically have 1-2 huge cash cows and they have a bunch of moonshots, R&D type stuff, etc. They are at a point where most of the moonshots have been disincentivized financially and just overall don’t see high enough ROI to continue pouring money into; exception is AI, which is a different skillset and this personnel can’t just pivot to it.
The fact is, a lot of tech products have matured and kind of entered maintenance mode. Feature additions are smaller incremental improvements than ever. If you cut out AI, I don’t know if any of the products we all use daily look much different than they did 5 years ago.
So my argument is this simply represents a natural business cycle where these companies are shifting from growth to mature. This shift always comes with a recalibration of expenses.
I keep seeing this over hiring argument everywhere. What's the evidence for that?
And weren't the layoffs of 2023-24 supposed to correct that?
In my experience, in every big tech company for the past two decades, almost everyone always complained that they were understaffed to meet their goals/okrs.
IMO, the real explanation is:
- Some unprofitable companies don't want to raise money in a high interest rate world, and are using AI as a partial excuse, partial ambition to do layoffs to cut costs. Eg: NET
- Other companies are investing heavily into data centers (eg: META) and need to cut costs elsewhere to offset that.
And in both cases, there's some real productivity gains that they have already seen due to AI that will offset, plus they expect to see more gains as the remaining people will work harder and be more motivated to increase their productivity for job security.
It's largely also down to management hierachy. Valve and Jane street for example both have semi-anarchist managment structures. In normal companies like GitLab, your worth as a manager and compensation is usually tied directly to the hierarchiecal structure. This incentivises you to create a project to hire a team and so on.
Yup, it's certainly a mix. I think the stock market incentive is most obvious for the usual VC-style "this must look presentable to the stock market before IPO"-spiel. After that I'd expect management/corp politics incentives to play a larger part indeed.
But isn't tech a bit unique in how accepted this kind of self-serving corruption is at pretty much all levels? From the "This might require a few extra folks, but I want to pad my resume" IC level all the way up to "Let's make this look good for Wall Street" exec level.
My experience has been that this is not unique to tech, and is common in all large enough industries. I think it's just the natural emergence of reward hacking i.e. if you're an executive at Pepsi and your job is largely to increase the stock price, and you know that you can do something to change the way your numbers are presented such that Wall St will like it, you'll likely do it.
I do think tech certainly has its own flavor though, particularly because of how differently it is treated by investors.
I do think there's a certain level of vagueness and lack of rigor that permeate throughout (software) tech that enables self-serving to a greater extent than usual. I agree though that it's to be expected in most corps at a certain level where that vagueness also starts to appear.
The headline is backwards, the companies cutting headcount for AI have already lost, that's why they're cutting. Most of these companies are losing market share or are in dying markets and need to cut headcount. They're just blaming AI because it's convenient and forward-looking.
That’s the true answer. AI is just used as an excuse for “oh shit, we over-hired, and it would be nice if we had an excuse to shed these people”. Then suddenly AI popped up, and well, that’s a nice excuse.
Returning to office was the previous excuse, so now they have the next one ready to go.
I think it’s simplest to view these companies 2020s layoffs cycles through the lens of a political battle between executives and middle managers. Middle managers always want to grow ICs because it’s their main stake to higher title and more pay. Executives are the intermitten contra to this, with their incentive being tied to the stock (high growth at high margins). AI gave executives more leverage, productivity had gone up so they could fire more with compromising operations.
Its also notable that over the last few decades the business community has normalized layoffs and layoffs while highly profitable as net positive for the company (stock). While all the engineering driven companies, GE, Boeing, etc that go down this route end up in strategic decline.
There are companies of a certain size that hold on to thousands of people they don't really need just so they can but them at a moment's notice to appease the stock market if bad news hits. It keeps talent out of their competitors, and gives them a cheat code if earnings dip or bad news hits.
I agree with both the gist of the article and your comment.
In the case of the well known FAANG and other similarly structured large software companies, everyone working there or having approached them know how bloated and inefficient they are.
What is happening is likely a long-term plan to completely restructure themselves, progressively shedding headcount while using it as a disciplinary tactic.
I am sure they understand the cost and the effect on morale.
But they are foreseeing radical changes and they want to be ready and slim before the storm.
> which is how GitLab is like 5x larger than Valve.
Doesn't Valve hire tons of contractors? I don't think anyone except them knows how many people/agencies they're paying to work on stuff like proton, linux, mesa, vulkan, etc., not to mention internal projects.
Valve could not actually do all that they do with just the leaked headcount numbers.
I'd expect GitLab to have contractors as well like most companies. Unclear if there's a significantly larger amount at Valve. It's a large gap to fill with contractors though to reach GitLab numbers.
GitLab is also a smaller company than I'd expect Valve to be - given its products & reach - if it was public and went through the same kind of inflation process.
Cutting people because of AI makes no sense, you know these people are good without AI, you'd want to keep them! Freeze the constant over-hiring instead, and take care of the people you know aren't lobotomized yet, and train them if needed. I'm seeing so much shedding of knowledge workers though, even though AI clearly isn't ready to replace people, just ready to augment them currently, that it looks like looney-tunes currently.
The major issue, is that once people are more productive with AI, you are able to replace people because less of them are required.
You see this in enterprise consulting, wiht the increase in cloud, serverless, SaaS/iPaaS, low code/no code, content generation and translations, followed by AI agent orchestration, the teams can be reduced down to about 1/3 of what they used to be.
It isn't as if there are enough projects around to keep the other 2/3 busy, so eventually when there are enough of those people on bench they have to find something else.
This sounds like an attempt to rationalize the fact that your business isn't that effective, otherwise adding more people would result in making more money.
> It isn't as if there are enough projects around to keep the other 2/3 busy
I've never worked at any company where there was any limit to the work to be done. Sales people don't give a shit what your product can do, only what they can sell, and they never sleep.
You can always find things to do but how much of that work has a positive ROI, or contributes to your business? You always have some bottlenecks, let's say implementing features was your bottleneck before AI, and you had a team of 15 product engineers with a 1y roadmap. With Claude Code a team of 5 was able to get that done in let's say 3 months. You don't magically come up with 9 months of work for 15 people right away. And your bottleneck will now be sales or something else, but your engineers won't convert to becoming sales people (or at least not all of them), and you might only need 1-2 more sales people, not 15.
> This sounds like an attempt to rationalize the fact that your business isn't that effective, otherwise adding more people would result in making more money.
Yes, or that businesses are expecting a slow down in the economy that hinders their ability to sell (i.e. their customers are going to cutback on spending)
This was the case last year (or maybe it was the year before) where technology companies saw their customers reducing spend and tightening belts.
The current economy feels hard to figure out, in that the market keeps going up but so is inflation and the struggle of the everyday American at least.
Perhaps that is leading technology companies to be more conservative in how much they produce.
But if we're assuming that AI is highly effective, shouldn't that lead to short-term growth in the economy as inflation drops? Shouldn't people then be expected to spend even more?
>I've never worked at any company where there was any limit to the work to be done. Sales people don't give a shit what your product can do, only what they can sell, and they never sleep.
The issue is how much of that work is "valuable" in the sense = makes money.
I have both been in projects and seen projects which were canceled once it turned out they didn't make money (bad sales? bad product? bad market fit? a bit of everything?). This you can only afford when you have money to spare (= with debts? high profits...?).
With the interest rates so high, how can a company justify hiring dozens/hundreds of people more? It's a risk, and what I am seeing now is that companies are shrinking left and right to focus on the business that makes money and reduce headcount on what they believe doesn't make money at all, or it's a cost too high for their "long term strategy" or whatever. Right now the only metrics that they are caring about is EBIDTA. They don't even care anymore about ARR, they are becoming irrelevant as long as they stay within a range (we want 20% increase, but we're ok with 5%).
The AI will replace everything and everyone is working out pretty well for Anthropic/OpenAI, though.
Predicting which projects will be valuable, directing resources in their direction, and making sure the right people are doing the right work with as few distractions as possible is the definition of high quality management and leadership.
Most managers are mediocre, and many are poor. Some are lucky for one or two projects but can't keep it up consistently.
Companies with a lot of wealth often scattershot random projects - some of which are directly competitive - in the hope that one will stick.
The people who have the insight and intuition to skip this and hit the mark directly are incredibly rare.
A lot of business culture is a set of cargo cult "solutions" that pretend to address this problem.
Only if all roles in a company are bottlenecked in the same way.
Example: natural monopoly in some geographically-locked domain. Just to grab an example. You have 150 people in the field, can't let them go because the company still needs hands and eyes on the ground. You have 15 people in some other, paper-pushing department. Thanks to AI advances, you only need 10 of those now.
> This sounds like an attempt to rationalize the fact that your business isn't that effective,
That's actually it. The part that can be sped up with AI don't change how slow everything else still takes. If you need 2 weeks to see the results of a change before AI, you still need that after AI.
Basically, your business is not keeping pace with development.
I’ve never worked anywhere where lack of work or projects was the bottleneck. Time, headcount, and budget were always the constraints.
Even in cost centers like IT or ops, there’s usually an endless backlog of work, technical debt, support requests, and improvements that never get prioritized because resources are limited.
> It isn't as if there are enough projects around to keep the other 2/3 busy, so eventually when there are enough of those people on bench they have to find something else.
Eh, are they doing that though?
Anecdotally, firing people “on the bench” isn't whats happening.
Read the tweets. Listen to people still working at these big corps. They are gutting teams and pushing more work onto people because they believe they can be more productive, not because they are.
Lets that sink in.
People are being made redundant on the basis that leadership believes that in the future they will have an over capacity and theyre cutting early to avoid the bench scenario.
It is speculative.
What this article is arguing, is that, that is stupid.
If, in the future, you need to spin up new initiatives, youve screwed yourself by disposing of your excess capacity in the magical hope that your current capacity will magically increase itself by … spending more money on tokens.
Its just nonsense.
AI is just an excuse for poor historical decisions and unfortunate global economic conditions.
The “cuts due to AI” will be real. There will be people sitting idle as the models improve and people learn to use them better.
… but right now?
they're not. Im not. My friends arent. My former work mates arent. The people left at these companies arent. The people being cut werent (except perhaps, at meta)
In consulting firing people “on the bench” happens regularly if the sales pipeline cannot keep everyone busy, there is only so much agencies pay from their own pocket if there are no outsourcing deals landing.
In some of those well known offshoring companies, being on bench automatically means a downcut on the salary as cost measure.
If you're increasing productivity, you should be doing more and growing, yes? If you're cutting payroll costs and trying to have the same level of capacity...your business sucks and you deserve your stock tanking.
Layoffs are a strong signal that a business is not investing in growth and is just trying to wring more profit from the same thing. If investors were rational, they'd walk away.
Maybe replacing the expensive C-suite with an LLM would help make better, growth-oriented decisions.
> Layoffs are a strong signal that a business is not investing in growth and is just trying to wring more profit from the same thing. If investors were rational, they'd walk away.
Not always. A buggy whip maker in 1920 should be laying off people. No amount of investment in buggy whips will bring that market back.
A layoff is saying that the investment will not pay off. So long as the company is cutting the right things they are good. Many layoffs are not done with a proper cut of the work do be done and so are bad, but that doesn't mean they are always bad.
I think the issue is that at any given point the work required to fire a 100-200k engineer is less than making that engineer do work to make you that much.
If you're growing obviously it's stupid to fire but if you have plateau'd the easiest gain is to trim the fat, so to speak. Also once a company is acquired by a private fund, everyone becomes a title and a number in a spreadsheet, that's all.
Again, shortsighted. There's no reason a business has to have a single product. If you run out of customers for cars, you make HVAC and front end loaders.
A lot of companies simply have no direction and aren't looking to build new products. They had a success, rotated in some myopic execs, flipped into rent-seeking mode and are trying to wring more cash out of the same progressive enshittified product.
Generally you can only profitably expand into adjacent products. Making and selling cars and HVACs are completely different, meaning that your expansion will be starting from zero knowledge. Furthermore, the sales channel of HVACs, sales strategies, etc. is not likely to have much in common with that for cars, so you are essentially creating a whole new startup company. (In the case of HVACs, it would be a startup in a commodity market, which would make no sense, because commodity markets have no real profits.) Doing one thing well is not just Unix philosophy, it is also a sound business strategy. Of course, usually adjacent things that could be done well suggest themselves, but often there is a limit to these.
Perhaps you've heard of Mitsubishi, Toyota Group or Samsung? Notably, Mitsubishi is involved in making cars, HVAC and heavy equipment.
Or, for that matter, Apple. If they subscribed to that philosophy, the iPhone wouldn't exist. Honestly, they would have kept trying to make the Apple II.
The rational thing is to analyze the opportunity cost of the investment, which is dependent on the always fluctuating prices.
Some businesses can grow, some cannot grow, some grow at different rates. The risk adjusted (subjective) prices determine whether or not an investor should walk away.
we're not taking into consideration the new bottleneck: product growth. imo it's the growth/expansion of the product that creates new work and opportunities for more work. at the moment, product growth is human-driven--because it's a risk to take a product in any direction, and whoever makes it has to bear the full responsibility. ideas that cause product growth are still generated by humans (and rightly so), perhaps also augmented by ai.
there's also what the market is ready for. we've said that some products failed because they were ahead of their time. it's even more true now where the power of ai, ai productivity, etc could take the product far beyond the markets expectation. what does that lead to? the deliberate slow rate of growth means power/potentials have to be controlled somewhat. so even without hiring, if ai is adopted as a first-class tool within the organization, there will be surplus resources that need to be shed somehow.
On the scale of a company, augmenting is replacing. If a worker plus AI can do the work of two workers without AI (but cheaper), you go for that; and it doesn't matter how good or bad AI is without the human.
The point is if a worker plus AI can do the work of two workers without AI, then why not keep both workers and have them both use AI to have the equivalent of four non-assisted workers?
Because you don’t have enough work that really needs doing, at least in that particular area. You cut engineers because the bottleneck to increased revenue isn’t software features or bugs, it’s marketing/sales; human beings’ limited attention for which there is now more competition than ever; and customers’ available funds.
ETA: this is sometimes (though not always) very different for a mature company than an early stage startup.
Ha! I’d never thought about it like that but…yeah.
I suspect another big part of it is that marketing and sales are relatively easy to measure and to scale.
You can hire one, two, or three new salespeople and expect that revenue will change more or less proportionately. Fixing (or ignoring) a handful issues doesn’t scale so smoothly—-there are jumps where the product suddenly seems much better/worse.
Because the entire structure of the business is designed for approximately the amount of work it currently does and likely has no particular immediate use for twice as much work in most departments.
Your experience of how the world works is usually because of what work you have done. You can't grow twice as many crops, sell twice as many groceries, drive twice as many busses, because of AI - fundamentally there's a consumption problem as well.
Many businesses are not bottlenecked by processes that are computer based.
But the firms in the headlines doing layoffs after layoffs aren't growing crops, selling groceries, or driving busses... They're knowledge work roles in companies selling intangible products and services. It's large corporations doing this much more than SMBs.
They’re also the ones constantly hiring and recruiting because internally nearly everyone benefits to having more people “under them”, and there’s a massive HR/Talent team that doesn’t go into hibernation after a 20% workforce reduction. Organizations want to grow, not because they need to but because it’s in the best interest of nearly all individuals still on the inside.
And I’m sure those companies also have “backlogs" due to limited labor/labor costs. There are always shelves to face, vehicles with deferred maintenance, and so on.
Obviously, there are limits: I’m not sure what my local grocery store or bus line would do with 100 new workers, but I have no doubt they could put a few people to work right away.
I think you're viewing this from the perspective of someone who has a functioning brain and plentiful concepts and ideas that aren't being built because you're labour-constrained. Companies like Meta simply don't have productive uses for all of that human + AI labour. Meta spends tens of billions a year paying people to throw shit at the wall and see what sticks. If the idea well you're going to is running dry, AI with a smaller number of humans can slop out the stuff you do want to build more efficiently is their implicit argument, especially when you don't care about quality (as is the case with Meta). Layoffs are also being used to tell a story around efficiency to investors while companies wait for the billions they're plowing in AI actually show profit.
Except that comforting C-suite narrative does not reflect reality. 2026 agents both increase productivity by knocking clearly specified but error-prone and tedious tasks out of the park whilst simultaneously vexing and annoying their users with hallucinations and downright lies on tasks with intrinsic ambiguity. This is made worse by the token providers with their constant tweaks to their deployments to cut costs w/o losing accuracy which flat doesn't work out well for the end user.
- AI pricing is variable, probably the cheapest it will ever be right now
- AI produces a lot more shit for humans to review, and you will always need humans. If you don’t focus on keeping things simple you will probably play yourself unless you’re good at separating out blast radiuses.
- I see a lot of super low quality work that doesn’t solve the problem but it’s like look that guy solved the problem in one day! Promote him! Everyone is happy except for the end users who for whatever reason are being totally ignored (whose problem it fails to appropriately solve) and I saw this in accounting software so…hello eventual lawsuits?
AI is definitely not the cheapest it will ever be right now. The frontier is getting more expensive, but the same capability will get cheaper over time.
The typical playbook for a VC funded startup is to race to a monopoly where the company can have higher margins. Prices continuing to go down for the consumer over time would require competition to stay high in the long term, and even then it’s not clear if even current prices are profitable.
The current level of AI has plenty of inherent competition from local models. In the long term, most of the profit will probably be from very smart models that run at something closer to datacenter scale over long inference loops - where local inference can't do much and even third-party inference/small neoclouds will be severely challenged. That is a very natural "moat" and has natural cross-efficiencies with AI model training, which requires a similar scale.
That's not always the case. Augmenting certainly can mean that, but it can also mean doing something that people couldn't do before.
For example, looking through meta data in a SQL environment that you didn't know existed to troubleshoot an issue. And a million other things. The odds of any employee not knowing everything are very good, even when humanity as a whole had already discovered that thing.
The AI story is a cover for the companies becoming less profitable in a bad economy, less able to invest because of the end of ZIRP, not capable of finding new markets, and developers becoming more expensive due to changes in accounting laws on the US.
People keep trying to make sense from the official story told by professional liars.
Because they aren't getting cut cause of AI. They are getting cut cause of the uncertainty re: future revenue streams AI is bringing. These layoffs are driven by fear and leadership lacking any vision for how the post AI world looks for their company. Their solution is to tighten the belts, which of course opens the window further for the AI companies to actually write software and make them obsolete.
I disagree. Lots of people are so laser focussed at only the close to the code aspects of programming they are unable to leverage the enormous increase in scope it can offer them.
There is absolutely room for head count reduction while companies restructure around this.
In reality there are two kinds of layoff going on.
Idiots who truly believe that AI will actually replace a significant amount of the workforce (Long term, it will not but some jobs displaced) and those who have internalised that ZIRP is over, they need to be more lean as VCs have closed their wallets and saying "We've revolutionised our workflow with AI" than "We haven't turned huge investment debt into profit in 5 years and we need to reduce headcount for even the slightest chance of that happening, or at least raise again by saying 'AI' over and over" has far better optics.
I would say that just the old "stupid" corporate practice of overhiring - for stock manipulation, reduce the talent pool for competitors etc. - Why Do Companies Overhire Just To Lay Off? - https://www.youtube.com/watch?v=KCF78a2wTHM
I think my read of it would be something like extreme short term thinking, and essentially like a push your luck game. If we can make it to the next model, maybe we can get enough productivity to reverse course.
As I understand it there are some underlying problems in the industry. Companies aren’t exiting which means VC’s aren’t getting returns and investment is slowing. In order for NVIDIA to maintain its current valuation, it has to keep growing at the rate it’s been growing which is the time during which companies have been piling on debt to perform extreme investments in AI. On top of this the growth rate is also based on extreme circular dealing.
So the whole thing isn’t sustainable, but the crash doesn’t happen until some big player blinks and their forecast / valuation goes down. And the only path to the kind of growth / gains people saw during the peak of the internet era is through AI.
Probably none of these big companies will fail, but it will be deeply uncomfortable for all of them when the music stops. And I wonder if it bankrupts a swath of the current generation of VCs.
That's the less common third kind. late 10s tech spending was absurd, it's the same excuse as the first but again, an easier way to reduce Capex for staff they don't need.
Preaching to the HN choir - great for upvotes but does anyone read anything new in this? Feels like I’ve seen the same article about this every month and I don’t think business people care.
more articles there will be about this - bigger the probability of this idea it will be in the dataset for chatgpt 6.9 thinking so the c-level of the companies would stop firing people
Even if you hire more people into jobs where AI acts a multiplier, there are still going to have to be layoffs where AI basically makes certain jobs _redundant_. I think both things are happening and will continue to happen, and happened anytime new technology came out.
My pet theory is that AI enables programmers to be relatively more productive than other roles. So, if I want to grow my company, shouldn't I hire MORE programmers? Anyone know a good counterargument?
Along another vein, I guess I wonder with my limited knowledge of economics if the demand for programmers is elastic or inelastic.
Most businesses don’t grow just by churning out more units of software. At some point, it doesn’t matter how quickly you can churn out features if you’re not solving customer problems and convincing customers that they should pay for those solutions.
Once software becomes cheap, the bottleneck to growth shift to product design, infrastructure/manufacturing, sales and support.
if constructing houses or buildings are quick, the bottleneck will be moved to other things. Like city planner, regulations / legals, material procurements, furnishing / equiments, etc.
for software it will be requirement gathering, product planning, looking for buyer / customer, even brainstorming on finding what to make.
The companies cutting headcount because of AI often just aren’t creative enough to expand their portfolio. There, I said it.
Today, even someone with minimal context can ship end-to-end prototypes. That means existing employees, the people with actual domain knowledge should be able to innovate faster than ever. Anthropic has shipped one of the best coding agents on the market and still has hundreds of engineering roles open.
High-agency people with company context and ideas within each company should be getting amplified right now, not laid off. Instead they're stuck figuring out how to "tokenmaxx" because leadership mandated it and they are forced to ship software no one will use in order to keep their jobs.
This right here. Big companies laying off engineers are basically saying our backlogs are empty and we're out of ideas. This should represent opportunity for small/mid companies to finally compete with the large competitors as some of the advantages of scale are normalized.
I think we're forgetting the companies somewhere in the middle: letting employees use AI, and refine their internal process, and not wasting time on "tokenmaxxing" and just building better products, the companies you rarely hear about, because if you're doing things right people don't notice as much as if you mess things up.
I agree with the article in its entirety, but its just so easy to play devil's advocate here. There might actually be a large value add to companies who replace poor performing employees with AI. So, who are the low versus high performing employees?
The largest problem with software employment is defining software developer performance. There is no industry baseline for defining or measuring this. Its so easy for a person to be that 10x (or much greater) developer in a compatible team or be a complete failure on the wrong team.
The problem for high-performing developers (even the elitest) is that over a long enough time scale, they won't be remotely competitive with high performing AI. At the moment we need human operators to guide and supervise these AIs, but unless the tech's progress grinds to a halt for some unforeseeable reason, it won't be this way for long.
I have been in multiple meetings where corporate folks parrot AI and productivity but if I ask for a business case and/or productivity savings, they go silent.
The funniest meeting has to be where someone said - "I have been using LLMs and it is terrible for our use case. Can we have an AI demo for Microsoft Copilot?" Imagine their surprise when they were told Copilot is LLM.
AI is just code for the actual economic reasons that companies are struggling with but it sounds better if you blame innovation instead ( Think AI = Actually Inflation, Actually Interest Rates, Actually Iran as the causes). The premise of this article is true because the companies who didn't shed jobs probably have a healthier balance sheet.
One datacenter rack of NVIDIA Vera Rubin GPUs costs $8 million. That's about the same cost as a team of software engineers. These data centers are not going to pay for themselves.
Europe. 100k USD is about 75k GBP, which deducting National Insurance (our social security program) and pension brings you to a salary of £63k roughly. You can easily hire good devs for this amount (especially outside of london).
To take advantage of AI orgs needs to optimize for giving humans broader agency, not AI usage or skills.
These large companies aren’t well positioned for this. There’s too many human bottlenecks to every decision. You get in trouble for going outside your role. Thats why AI is drudgery at many of these places, often even negative productive.
Smaller companies though started assuming AI will let their employees stretch to broader roles, wear many hats, and given larger agency
The article makes a good point, but it's based on a polarized view that people make good judgement and AI is incapable of making good judgement.
I've worked with people who demonstrated below-average judgement, and I've seen cases of good judgement from AI. I think if a company can identify the poor performers and part wit them, there is a decent chance that the remaining people with AI in hand can more than make up the difference.
What's great about this article is it's a falsifiable hypothesis. Is anyone keeping a list of companies in different sectors not firing due to AI? We can revisit this in 2 years and see how we did.
> AI Does Not Replace Judgement. It Multiplies It.
And that is a problem. If you employ people whomlove to comlkain are basically minus 0.1 of regular employe (tolerable annoyance), they get power multiplier and instead of small rants at water cooler, will be able to file federal law suits under employers ass.
And they will also get way more ways to harrass productive employes.
> It is the knowledge they carry. How the business actually operates. Where the edge cases live. Why certain decisions get made the way they do. What customers really mean when they complain about a specific issue. The context that never makes it into a process document because it does not need to — because the right person already knows.
I get this is about the immediate firing of people and losing knowledge quickly but there’s no reason this this knowledge can’t be captured actively and passively over the next year or so now there’s a massive incentive to do so. In fact using AI to capture this knowledge is a huge opportunity tons of startups are working on.
Relatively speaking, a company could obtain AGI first merely by just by keeping their critical thinkers while the rest of the industry offloads their thinking to LLMs. It wouldn’t be AGI, just GI, like we’ve always had before before GPT arrived
There are clearly defined tasks these things already do that are tedious and error-prone for a human. Using them correctly amplifies the abilities of humans. The problem is with the C-suite mentality towards cutting costs by replacing humans. It isn't going to work, but the only way they learn lessons like that is, as always, the hard way.
I don't think any company is cutting headcount because of AI. I think that's just what they say. The reality is that they're facing a cash crunch due to a bad economy. And if you state a cash crunch as the reason, your stock will fall or VCs will run away. But surprisingly, if you state AI as the reason for layoffs, your stock will rise or VCs will get excited. Same thing, just different clothes, and it works.
I still think AI is just a cover story for these job cuts. Tech companies are still "rightsizing" after unsustainable growth during the pandemic. At the same time we're clearly headed into a recession.
Investors like growth, not shrinkage. Claiming AI is replacing those jobs helps avoid the appearence of shrinkage, while also feeding the AI hype machine that many of these companies have invested heavily into.
Nobody ever knew what the right size was. There was a valve that was opened to allow hiring. And another "bung" that could be unplugged to facilitate firing. The algorithm was to open the valve when times are good/money cheap/greed is good, and opened the bung in other times. Totally decoupled from how many people (and largely what sort of people) are required. AI has changed that mechanism.
A friend of mine works at big corporate. He is in meetings all the time and barely codes. For him a 20$ AI subscription is more than enough.
I'm personally at a fintech startup and have a 100$ subscription. Our org is much leaner and its much faster to ship stuff. What used to be entire departments or teams can now be done by two or three people. Three people to ensure we share context and have a decent tram factor.
Its a little odd that you think these two examples are comparable. How many customers does his product have and how many does yours?
Of course a leaner team can ship faster. But your stakes are much lower. A production bug has less visibility and impact. Limited revenue means less stakeholders, and less opinions about how and what to do in a startup. Less customers means less customer support and feedback. The differences go on and on. I'm not sure you can attribute all of that to AI assistance.
Good luck with your fintech startup. I'm in that space too and its a bear.
The problem that software engineers and product teams often face is that the time from roadmap to feature is quite long, and AI has offered a clear, meaningful speedup to some parts, like writing tests and boilerplate. In many cases, the same ticket flow can be managed much faster with AI. So, what? We're not quite at the point where people have transitioned this into demand for more product deliveries faster. As soon as that occurs, it doesn't matter how great the AI is, because the current pace will be slow. Why stop at what the current roadmap is? Why not ask for double the features? When we gave everyone power tools that didn't dramatically make construction easier, it just enabled more complex buildings.
The misunderstanding is that people equate what these companies are saying and what they are doing. In reality this is the perfect time to fire people under cover. This happens every X years where as leadership you can say this tech is displacing you, when in truth its just that you arent needed anymore.
This is a capitalism functioning correctly as labor as in then moved to places it is needed more. Its sort of shocking to see devs complain about getting laid off. Thats the point of a high variance career. Great and terrible outcomes go hand in hand, if you want one you must expect the other.
I don't think software was ever sold as a high variance career. Start a startup! was sold that way, particularly by PG's early essays, but not not W2 employment. You do W2 because you want low variance, and you either love software or it pays well. Either way, I don't think most people were okay with terrible outcomes. (I don't think getting laid off is terrible, I think it should be an expected possibility. I've found that contracting/freelance has helped reduce my fear of losing my job, since you expect the job to end when you start it, and since you're regularly looking for a new job, you get a sense of how long it takes.)
Simply reality is very few companies are cutting because of AI. They’re just trimming bloat that’s been there the whole time. The AI narrative is broadly just PR spin and CEOs trying to look cool.
These two ideas, that (1) AI is making people more productive and (2) we are laying off employees because AI replaces them. Those two ideas are mutually incompatible. When employees become more productive, you always always always want to hire more of them, because it’s a multiplier on your outputs and therefore an efficient use of resources.
That is why one of these narratives is total bullshit. Either “layoffs from AI” are layoffs for some other structural reason in the tech sector, or “AI makes employees way more productive” is total bullshit.
This assumes that demand is elastic and unbounded. It might be true that there's a lot of untapped demand for software, but especially considering individual firms, it's plausible that demand for their offerings might be capped. In this case it does make sense for firms to shrink.
Analogously, automation and productivity improvements in farming drastically decreased employment in the farming sector, from basically everyone to basically no one.
This is a simple economic fallacy. Your idea holds true if your business has room for growth, but you always need a market for those multiplied outputs. In an economy where every business has multiplied outputs, but demand doesn't scale with those outputs, what do you think happens?
If the article is about the craft of writing, then no point. But if the article is about communicating ideas, AI is just the means the person with the ideas chose.
However, in this case I think both were fairly vacuous. Also, sentence fragments are a hallmark of bad writing that high school English teachers try to train you out of, apparently unsuccessfully if the training data makes them so likely.
I don't know anyone who has been hiring in any kind of significant numbers. Everyone fires either because of alleged AI efficiencies or to pay for AI token shit or to just bla bla to investors how magically AI AI AI they are.
End result all the same. Massive job losses due to AI.
That's absolutely correct. I use LLMs heavily in my work and open source projects. These tools are effective, they can solve hard problems and they allow me to work on a wider range of tasks than could before.
But, they're also jagged in terms of functionality. When you work with a human, you can learn what their core competencies are, and then if you give them a task that falls within that domain, you can be reasonably sure they'll finish it correctly. That's not the case with LLMs. It might do one task brilliantly, and a next similar task, it just shits the bed on.
And since it has no understanding of the task in a human sense, it can't self correct, learn or improve. All its doing is stringing tokens together based on probability. So, you need a human in the loop to review what it's doing, to correct it when it makes mistakes, and to define the actual goals. My experience is that doing all that properly ends up taking up a lot of time, so your actual productivity gain per person aren't all that significant.
Companies that try to replace humans with LLMs will soon find that they end up with a whole bunch of code that doesn't actually work, and they have no hope of fixing. The double edge of LLMs is that they're really good at generating a lot of wrong code really fast.
This is a pro-AI piece. They realized they overplayed their hands when college graduates started booing them and their new line is "AI won't replace workers, it will make them more valuable!"
Another bury-your-head-in-the-sand article. It sucks but AI replacing coders is very real. I haven’t coded since last year but I’ve pushed more features at a rate faster than I ever could in my entire career. I pushed new code while chatting on WhatsApp with my friends yesterday using 2 prompts.
“Remove this old feature.”
“Are you sure you didn’t break anything?”
That was it. Then I manually tested it to make sure nothing was broken. Then I did a brief review before posting it for code review and then pushed it. What would have taken me probably about 1 day to go through and figure out code changes and then actually change the code took me about 30 seconds.
To think that we need to maintain the post-Covid hiring bloat is nonsense. I’m not so arrogant to think that someone with an llm can’t replace me, if I survive a few more years in this industry I’ll be amazed and grateful.
Capitalists usually recoup gains rather than re-investing in people or working in new verticals. This gels with companies who would rather do anything else than grow.
We need stronger goverment and socialist revokution! Strong socialist goverment invest it to people! Socialist goverment would not starve people, torture, or just send them somwehere else (those examples do not count)
Wasn't really trying to take the socialist view. Merely stating what usually happens. Grow the business usually means make more money in the easiest way possible (cost cutting usually) rather than invest in future gains.
I once was the final step in an interview process. My flag to the team was that this person had been laid off something like four times in a row. I said, layoffs don’t cut the great people, so it says something about the applicant.
The team insisted this person had a big heart, a lot of passion, and they’d make a difference. We extended an offer and they accepted.
They ended up being unremarkable in every way. It was a drag on the company. We let them go a year later and set ourselves back on a topic our customers cared about.
Anyway, some people are great. Some are not. The idea that there’s this hidden special knowledge trapped inside teams, if only the executives would unleash it, is nonsense. I personally feel for anyone who loses their job, but companies will always pick the most productive and efficient path to a result.
Also so does the labor market. How many people, when given the option for more pay and benefits, decline out of loyalty to their employer? At least in tech, this is almost never the case.
I think we’re a bit lost. On one hand, certain people don’t trust these big employers and think they’re evil, and on the other hand really really want their job at said evil megacorp.
Companies are cutting headcount to fund AI projects, not necessarily because of the gains they are getting from it. And frankly in many companies, cutting layers and reducing team handoffs will be a net gain - even if AI didn't exist.
Does anybody else find this article reads really AI written? The way the argument is built, the emdashes, the dramatic one sentence paragraph:
"That assumption is wrong."
"The human is the equation. AI is what makes that equation run faster." -- what?
> The Atlas Interfaces drift alone in the endless void. They are silent. They are unknowable fragments of an ancient whole. Yet their imprint on time and space molds our existence. They are the equation, and life is its answer. Through their Monoliths they give understanding to their boundless meaning, and that of our own.
Testament of the Korvax, No Man's Sky.
It is a little ironic that this article reads so much like an AI-generated one.
It is. Just count how many em dashes in it.
And the article images with the signature AI-face
I think the unfortunate reality is that lots of companies in our industry have suspiciously inflated employee counts in the first place. Even when removing AI and the pandemic over-hiring, I wouldn't have been surprised to see corrections sooner or later.
Employee count seems to correlate to stock market incentives - which is how GitLab is like 5x larger than Valve.
> I think the unfortunate reality is that lots of companies in our industry have suspiciously inflated employee counts in the first place. Even when removing AI and the pandemic over-hiring, I wouldn't have been surprised to see corrections sooner or later.
It's funny to me that everyone talks about pandemic over-hiring, as if this hadn't been a thing for a decade before that. Like, when I started at FB in 2013, they probably had about 75% of the engineers they needed (and about 10% of the sales people). But they grew engineering much faster than sales for some reason (engineering in the broad sense including data and product).
That being said, it's easy for people to look at other parts of a company and think they are over-staffed, because we don't see all the work necessary to keep the balls in the air.
AI aside, if we just look at other engineering disciplines in mature sectors, the future is not bright(er).
No fully paid golden cadillac benefits packages (for your dog too!), no twice daily uber eats comps, no $150k entry level, no unlimited PTO, no 6 months leave.
If you go to your uncles engineering department at the boiler company, these guy's engineering roles are about as pampered as the warehouse managers.
The upside is that it will cull those in it just for the money/lifestyle, and concentrate it down to those in it for the love of the craft.
Two things:
1) Unlimited PTO is a scam. Ask anyone who ever got paid out six weeks salary when they changed jobs.
2) “the craft” is doing some heavy lifting here. I happen to enjoy AI-assisted dev, but it is nothing like the work that drew me to the industry.
Otherwise agreed on all counts.
Unlimited PTO works great for someone like me who regularly takes half day and full day PTO every quarter for various commitments/life stuff. But I’m also a guy and I have no problem asserting myself/spending my benefits. People who are a little less comfortable with that end up just parking on it and would do a lot better if they were given a set number of PTO days. The ambiguity hurts them in the same way not having set raise schedules and discussions hurts them.
Edit: folks, I can assure you I take a lot more than 10 days of PTO. I didn’t even say how many half/full days I take, you’re assuming a lot though I get my language could’ve been more clear there. I probably take between 20-30 days total annually. I am saying that because of unlimited PTO, it is very easy to just grab half days and full days as needed routinely and not worry about my overall count.
Before unlimited PTO took off, standard in FAANG-like (US) industry was ~12 corp holidays + 15-25 days annual PTO, depending on seniority.
I don’t know anyone who takes the high end of that anymore, especially senior/staff folks.
Unlimited PTO takes a liability off the company’s books, and makes every time off request a negotiation.
> Unlimited PTO takes a liability off the company’s books, and makes every time off request a negotiation.
This is 100% it.
Yeah exactly. For some of us that is not a problem but for most people it seems to create issues. Company culture also plays a huge factor in that. Grey areas tend to hurt employees the most!
To be clear we still get federal holidays. So I net probably 30-40 days a year.
Again, the system works for me. But it is also a constant negotiation/workplace culture consideration, so I am not advocating for it writ large. I’m just lucky I work at a company that doesn’t put pressure on people not to utilize their PTO.
At my current job I get 25 days of PTO a year, with a cap of 30 days, plus company holidays (12 or so?), plus 5 days the entire company shuts down. I have to be careful not to hit my cap even taking 4 weeks of vacation because it’s still less than I earn. You have “unlimited” pto and take half of the vacation I do. If you ever change jobs you’ll get $0, while I’ll get a bonus.
You have it backwards. You’d benefit tremendously from fixed PTO that pays out because you’re taking fewer than 10 days a year. Its biggest benefactors are low-seniority employees who take 20-30 days annually.
> someone like me who regularly half day and full day PTO every quarter
So you take something less than 10 days off a year?
In all developed countries bar one that is against the law for being so few days off.
You are being abused.
No, that is incorrect. See my edit
You take 1.5x4 = 6 days of vacation per YEAR and use this as an argument for... wait... what exactly?
Huh?
Please read my edit
> Unlimited PTO is a scam
Well, no. It's just the target beneficiary is not you (the employees) but the employer.
Unlimited PTO resolves the employer from paying out PTO when they make you "redundant".
It's more than that. Unlimited PTO in practice also results in people taking less time off.
Some people have to be reminded to take time off for themselves (like me in my early years), and without a predefined amount of PTO and an expiration mechanism, there's nothing stopping them from going extended periods without taking any time for themselves.
Most people grow out of that eventually, but their gems turn black before they fully advocate for their own work/life balance.
The other side to that coin is that when employees take varying amounts of time off, it becomes easy for managers to pressure anyone taking more time off to conform with those who take less (and sometimes none at all).
> The upside is that it will cull those in it just for the money/lifestyle, and concentrate it down to those in it for the love of the craft.
Is that what you see at your uncle's boiler company? People who are truly in it "for the love of the craft"?
Yeah wtf? If programming becomes a generic white collar job, it's not really accurate to characterize the people as "those in it for the love of the craft". After all, do you think the average accountant or analyst is doing it "for the love of the craft"?
That is a SV thing, in many places it is a regular office job.
You are already in a treat if having free coffee and fruit.
You don't need to look at other engineering jobs, just look at software engineers out side the US. We make decent money compared to the local market, but we've never had the royal treatment that US devs seem to get.
Exactly this - the late 00s and 10s we were riding the wave of the internet and smartphone adoption. Those markets have matured now and this is ultimately just the response. Covid was like some weird last hurrah for irrational spending. LLMs are just where all of that silly energy is going.
FB is the weirdest company though. For a lot of companies like gitlab you know they have sales for customers, roadmap for their core product etc.
FB has what?
Intel has around the same number of people and they make chips, laptops, different types of chips. Hardware, software, supply chain etc.
FB does what? And what do they do with all these very well paid people?!
algorithm, tracking, bypassing adblocker, reports, moderation and backoffice tools I think
They do A/B tests all day
Well they keep breaking main FB page with additions or redesigns nobody asked for.
I am in Europe but I must be some test bunny that made some personal enemy there somehow, since beginning (cca 2008) FB was by far the buggiest site ever. Uploads of images didn't work, sometimes it uploaded twice. Uploading album of 30-50 pics took on average maybe 10 attempts in the past. Comments didn't work, or twice again. many similar behaviors. Ie since cca 2 weeks now if I click on some picture to have it full screen, clicking close after 1-2 mins ends up in FB error page instead of returning to main feed. I keep seeing FB error page a LOT considering infrequent use.
Horrible, terrible engineering.
> when I started at FB in 2013, they probably had […] about 10% of the sales people [they needed]
I find that hard to believe. Looking at https://www.macrotrends.net/stocks/charts/META/meta-platform..., they had a good revenue growth in the early ‘10s.
What did they leave on the table by not having more ten times as many sales people as they did? Revenues? Profits?
> I find that hard to believe. Looking at https://www.macrotrends.net/stocks/charts/META/meta-platform..., they had a good revenue growth in the early ‘10s.
So, how they hired salespeople was based on revenue per head. If there wasn't enough revenue in a particular part of the business, no new hires for you. Obviously, that's gonna be a lagging indicator.
And they were honestly leaving a lot of money on the table there, but honestly not as much as they were leaving on the table by having a really, really, really bad and buggy self-serve interface that made it hard for small advertisers to spend money.
When they fixed that, the small businesses all got good results (e.g. hairdressers etc) and those people then convinced all the agency people to put more money into the platform.
Up till 2015 or so, it was a real struggle to convince people to even try FB ads, but after that it was like shooting fish in a barrel.
> I find that hard to believe. Looking at https://www.macrotrends.net/stocks/charts/META/meta-platform..., they had a good revenue growth in the early ‘10s.
Also, up to 2013 or so (when feed ads became a thing), most of that money was coming from payments for f2p games, so they had lots of partner managers there, but not enough ads sales people.
I’d say it differently though. The company who hired all these people are finding they don’t have enough value added projects to keep them employed. Most of the FAANG and larger VC backed companies have something in common. They typically have 1-2 huge cash cows and they have a bunch of moonshots, R&D type stuff, etc. They are at a point where most of the moonshots have been disincentivized financially and just overall don’t see high enough ROI to continue pouring money into; exception is AI, which is a different skillset and this personnel can’t just pivot to it.
The fact is, a lot of tech products have matured and kind of entered maintenance mode. Feature additions are smaller incremental improvements than ever. If you cut out AI, I don’t know if any of the products we all use daily look much different than they did 5 years ago.
So my argument is this simply represents a natural business cycle where these companies are shifting from growth to mature. This shift always comes with a recalibration of expenses.
I keep seeing this over hiring argument everywhere. What's the evidence for that?
And weren't the layoffs of 2023-24 supposed to correct that?
In my experience, in every big tech company for the past two decades, almost everyone always complained that they were understaffed to meet their goals/okrs.
IMO, the real explanation is:
- Some unprofitable companies don't want to raise money in a high interest rate world, and are using AI as a partial excuse, partial ambition to do layoffs to cut costs. Eg: NET
- Other companies are investing heavily into data centers (eg: META) and need to cut costs elsewhere to offset that.
And in both cases, there's some real productivity gains that they have already seen due to AI that will offset, plus they expect to see more gains as the remaining people will work harder and be more motivated to increase their productivity for job security.
It's largely also down to management hierachy. Valve and Jane street for example both have semi-anarchist managment structures. In normal companies like GitLab, your worth as a manager and compensation is usually tied directly to the hierarchiecal structure. This incentivises you to create a project to hire a team and so on.
Yup, it's certainly a mix. I think the stock market incentive is most obvious for the usual VC-style "this must look presentable to the stock market before IPO"-spiel. After that I'd expect management/corp politics incentives to play a larger part indeed.
But isn't tech a bit unique in how accepted this kind of self-serving corruption is at pretty much all levels? From the "This might require a few extra folks, but I want to pad my resume" IC level all the way up to "Let's make this look good for Wall Street" exec level.
My experience has been that this is not unique to tech, and is common in all large enough industries. I think it's just the natural emergence of reward hacking i.e. if you're an executive at Pepsi and your job is largely to increase the stock price, and you know that you can do something to change the way your numbers are presented such that Wall St will like it, you'll likely do it.
I do think tech certainly has its own flavor though, particularly because of how differently it is treated by investors.
I do think there's a certain level of vagueness and lack of rigor that permeate throughout (software) tech that enables self-serving to a greater extent than usual. I agree though that it's to be expected in most corps at a certain level where that vagueness also starts to appear.
I don't think it's that unique but Engineering type people might be more optimisation minded and see total comp as something to optimise.
And moreover,
The headline is backwards, the companies cutting headcount for AI have already lost, that's why they're cutting. Most of these companies are losing market share or are in dying markets and need to cut headcount. They're just blaming AI because it's convenient and forward-looking.
That’s the true answer. AI is just used as an excuse for “oh shit, we over-hired, and it would be nice if we had an excuse to shed these people”. Then suddenly AI popped up, and well, that’s a nice excuse.
Returning to office was the previous excuse, so now they have the next one ready to go.
I think it’s simplest to view these companies 2020s layoffs cycles through the lens of a political battle between executives and middle managers. Middle managers always want to grow ICs because it’s their main stake to higher title and more pay. Executives are the intermitten contra to this, with their incentive being tied to the stock (high growth at high margins). AI gave executives more leverage, productivity had gone up so they could fire more with compromising operations.
Its also notable that over the last few decades the business community has normalized layoffs and layoffs while highly profitable as net positive for the company (stock). While all the engineering driven companies, GE, Boeing, etc that go down this route end up in strategic decline.
There are companies of a certain size that hold on to thousands of people they don't really need just so they can but them at a moment's notice to appease the stock market if bad news hits. It keeps talent out of their competitors, and gives them a cheat code if earnings dip or bad news hits.
I agree with both the gist of the article and your comment.
In the case of the well known FAANG and other similarly structured large software companies, everyone working there or having approached them know how bloated and inefficient they are.
What is happening is likely a long-term plan to completely restructure themselves, progressively shedding headcount while using it as a disciplinary tactic.
I am sure they understand the cost and the effect on morale.
But they are foreseeing radical changes and they want to be ready and slim before the storm.
> which is how GitLab is like 5x larger than Valve.
Doesn't Valve hire tons of contractors? I don't think anyone except them knows how many people/agencies they're paying to work on stuff like proton, linux, mesa, vulkan, etc., not to mention internal projects.
Valve could not actually do all that they do with just the leaked headcount numbers.
I'd expect GitLab to have contractors as well like most companies. Unclear if there's a significantly larger amount at Valve. It's a large gap to fill with contractors though to reach GitLab numbers.
GitLab is also a smaller company than I'd expect Valve to be - given its products & reach - if it was public and went through the same kind of inflation process.
>which is how GitLab is like 5x larger than Valve.
Seems questionable to compare the valuation of a public company vs a private one, with the latter being marked to market rarely, if ever.
This was in relation to employee count, not valuation.
A lot of places did over-hire, but I'm suspicious that those places will successfully cut the fat rather than the muscle.
Cutting people because of AI makes no sense, you know these people are good without AI, you'd want to keep them! Freeze the constant over-hiring instead, and take care of the people you know aren't lobotomized yet, and train them if needed. I'm seeing so much shedding of knowledge workers though, even though AI clearly isn't ready to replace people, just ready to augment them currently, that it looks like looney-tunes currently.
The major issue, is that once people are more productive with AI, you are able to replace people because less of them are required.
You see this in enterprise consulting, wiht the increase in cloud, serverless, SaaS/iPaaS, low code/no code, content generation and translations, followed by AI agent orchestration, the teams can be reduced down to about 1/3 of what they used to be.
It isn't as if there are enough projects around to keep the other 2/3 busy, so eventually when there are enough of those people on bench they have to find something else.
This sounds like an attempt to rationalize the fact that your business isn't that effective, otherwise adding more people would result in making more money.
> It isn't as if there are enough projects around to keep the other 2/3 busy
I've never worked at any company where there was any limit to the work to be done. Sales people don't give a shit what your product can do, only what they can sell, and they never sleep.
You can always find things to do but how much of that work has a positive ROI, or contributes to your business? You always have some bottlenecks, let's say implementing features was your bottleneck before AI, and you had a team of 15 product engineers with a 1y roadmap. With Claude Code a team of 5 was able to get that done in let's say 3 months. You don't magically come up with 9 months of work for 15 people right away. And your bottleneck will now be sales or something else, but your engineers won't convert to becoming sales people (or at least not all of them), and you might only need 1-2 more sales people, not 15.
> This sounds like an attempt to rationalize the fact that your business isn't that effective, otherwise adding more people would result in making more money.
Yes, or that businesses are expecting a slow down in the economy that hinders their ability to sell (i.e. their customers are going to cutback on spending)
This was the case last year (or maybe it was the year before) where technology companies saw their customers reducing spend and tightening belts.
The current economy feels hard to figure out, in that the market keeps going up but so is inflation and the struggle of the everyday American at least.
Perhaps that is leading technology companies to be more conservative in how much they produce.
But if we're assuming that AI is highly effective, shouldn't that lead to short-term growth in the economy as inflation drops? Shouldn't people then be expected to spend even more?
>I've never worked at any company where there was any limit to the work to be done. Sales people don't give a shit what your product can do, only what they can sell, and they never sleep.
The issue is how much of that work is "valuable" in the sense = makes money.
I have both been in projects and seen projects which were canceled once it turned out they didn't make money (bad sales? bad product? bad market fit? a bit of everything?). This you can only afford when you have money to spare (= with debts? high profits...?).
With the interest rates so high, how can a company justify hiring dozens/hundreds of people more? It's a risk, and what I am seeing now is that companies are shrinking left and right to focus on the business that makes money and reduce headcount on what they believe doesn't make money at all, or it's a cost too high for their "long term strategy" or whatever. Right now the only metrics that they are caring about is EBIDTA. They don't even care anymore about ARR, they are becoming irrelevant as long as they stay within a range (we want 20% increase, but we're ok with 5%).
The AI will replace everything and everyone is working out pretty well for Anthropic/OpenAI, though.
Predicting which projects will be valuable, directing resources in their direction, and making sure the right people are doing the right work with as few distractions as possible is the definition of high quality management and leadership.
Most managers are mediocre, and many are poor. Some are lucky for one or two projects but can't keep it up consistently.
Companies with a lot of wealth often scattershot random projects - some of which are directly competitive - in the hope that one will stick.
The people who have the insight and intuition to skip this and hit the mark directly are incredibly rare.
A lot of business culture is a set of cargo cult "solutions" that pretend to address this problem.
So you never worked at a consulting agency working on bids for outsourced development.
Only if all roles in a company are bottlenecked in the same way.
Example: natural monopoly in some geographically-locked domain. Just to grab an example. You have 150 people in the field, can't let them go because the company still needs hands and eyes on the ground. You have 15 people in some other, paper-pushing department. Thanks to AI advances, you only need 10 of those now.
> This sounds like an attempt to rationalize the fact that your business isn't that effective,
That's actually it. The part that can be sped up with AI don't change how slow everything else still takes. If you need 2 weeks to see the results of a change before AI, you still need that after AI.
Basically, your business is not keeping pace with development.
I’ve never worked anywhere where lack of work or projects was the bottleneck. Time, headcount, and budget were always the constraints.
Even in cost centers like IT or ops, there’s usually an endless backlog of work, technical debt, support requests, and improvements that never get prioritized because resources are limited.
Try consulting agencies, or freelancing as external contractor.
> It isn't as if there are enough projects around to keep the other 2/3 busy, so eventually when there are enough of those people on bench they have to find something else.
Eh, are they doing that though?
Anecdotally, firing people “on the bench” isn't whats happening.
Read the tweets. Listen to people still working at these big corps. They are gutting teams and pushing more work onto people because they believe they can be more productive, not because they are.
Lets that sink in.
People are being made redundant on the basis that leadership believes that in the future they will have an over capacity and theyre cutting early to avoid the bench scenario.
It is speculative.
What this article is arguing, is that, that is stupid.
If, in the future, you need to spin up new initiatives, youve screwed yourself by disposing of your excess capacity in the magical hope that your current capacity will magically increase itself by … spending more money on tokens.
Its just nonsense.
AI is just an excuse for poor historical decisions and unfortunate global economic conditions.
The “cuts due to AI” will be real. There will be people sitting idle as the models improve and people learn to use them better.
… but right now?
they're not. Im not. My friends arent. My former work mates arent. The people left at these companies arent. The people being cut werent (except perhaps, at meta)
Its stock price hype theatre.
In consulting firing people “on the bench” happens regularly if the sales pipeline cannot keep everyone busy, there is only so much agencies pay from their own pocket if there are no outsourcing deals landing.
In some of those well known offshoring companies, being on bench automatically means a downcut on the salary as cost measure.
> if the sales pipeline cannot keep everyone busy
That has nothing to do with AI.
It is happening due to global economic conditions.
Im not saying the bench doesn't exist, Im saying its not full of people because you have two consultants doing all 59 jobs with AI.
There are definitely places making cuts of staff who are not on the bench.
You have a team of two consultants, for what used to be a team of five, kind of.
No need to take the 69 out of a magic hat.
The other three join the on bench pool.
When economy gets worse, the lucky two are the ones staying.
If you're increasing productivity, you should be doing more and growing, yes? If you're cutting payroll costs and trying to have the same level of capacity...your business sucks and you deserve your stock tanking.
Layoffs are a strong signal that a business is not investing in growth and is just trying to wring more profit from the same thing. If investors were rational, they'd walk away.
Maybe replacing the expensive C-suite with an LLM would help make better, growth-oriented decisions.
> Layoffs are a strong signal that a business is not investing in growth and is just trying to wring more profit from the same thing. If investors were rational, they'd walk away.
Not always. A buggy whip maker in 1920 should be laying off people. No amount of investment in buggy whips will bring that market back.
A layoff is saying that the investment will not pay off. So long as the company is cutting the right things they are good. Many layoffs are not done with a proper cut of the work do be done and so are bad, but that doesn't mean they are always bad.
I think the issue is that at any given point the work required to fire a 100-200k engineer is less than making that engineer do work to make you that much.
If you're growing obviously it's stupid to fire but if you have plateau'd the easiest gain is to trim the fat, so to speak. Also once a company is acquired by a private fund, everyone becomes a title and a number in a spreadsheet, that's all.
In consulting you don't make magically customers out of nowhere, and there is a limited pool of customers to feed from.
Again, shortsighted. There's no reason a business has to have a single product. If you run out of customers for cars, you make HVAC and front end loaders.
A lot of companies simply have no direction and aren't looking to build new products. They had a success, rotated in some myopic execs, flipped into rent-seeking mode and are trying to wring more cash out of the same progressive enshittified product.
Generally you can only profitably expand into adjacent products. Making and selling cars and HVACs are completely different, meaning that your expansion will be starting from zero knowledge. Furthermore, the sales channel of HVACs, sales strategies, etc. is not likely to have much in common with that for cars, so you are essentially creating a whole new startup company. (In the case of HVACs, it would be a startup in a commodity market, which would make no sense, because commodity markets have no real profits.) Doing one thing well is not just Unix philosophy, it is also a sound business strategy. Of course, usually adjacent things that could be done well suggest themselves, but often there is a limit to these.
Perhaps you've heard of Mitsubishi, Toyota Group or Samsung? Notably, Mitsubishi is involved in making cars, HVAC and heavy equipment.
Or, for that matter, Apple. If they subscribed to that philosophy, the iPhone wouldn't exist. Honestly, they would have kept trying to make the Apple II.
That dilutes your experience and risks losing focus on the customers you already know how to serve well and in turn can destroy your company.
It works for some it fails for others.
The rational thing is to analyze the opportunity cost of the investment, which is dependent on the always fluctuating prices.
Some businesses can grow, some cannot grow, some grow at different rates. The risk adjusted (subjective) prices determine whether or not an investor should walk away.
we're not taking into consideration the new bottleneck: product growth. imo it's the growth/expansion of the product that creates new work and opportunities for more work. at the moment, product growth is human-driven--because it's a risk to take a product in any direction, and whoever makes it has to bear the full responsibility. ideas that cause product growth are still generated by humans (and rightly so), perhaps also augmented by ai.
there's also what the market is ready for. we've said that some products failed because they were ahead of their time. it's even more true now where the power of ai, ai productivity, etc could take the product far beyond the markets expectation. what does that lead to? the deliberate slow rate of growth means power/potentials have to be controlled somewhat. so even without hiring, if ai is adopted as a first-class tool within the organization, there will be surplus resources that need to be shed somehow.
On the scale of a company, augmenting is replacing. If a worker plus AI can do the work of two workers without AI (but cheaper), you go for that; and it doesn't matter how good or bad AI is without the human.
The point is if a worker plus AI can do the work of two workers without AI, then why not keep both workers and have them both use AI to have the equivalent of four non-assisted workers?
Because you don’t have enough work that really needs doing, at least in that particular area. You cut engineers because the bottleneck to increased revenue isn’t software features or bugs, it’s marketing/sales; human beings’ limited attention for which there is now more competition than ever; and customers’ available funds.
ETA: this is sometimes (though not always) very different for a mature company than an early stage startup.
That is a very convenient message for marketing and sales people. The fact that their whole job is crafting messages shouldn’t raise any eyebrows.
Ha! I’d never thought about it like that but…yeah.
I suspect another big part of it is that marketing and sales are relatively easy to measure and to scale.
You can hire one, two, or three new salespeople and expect that revenue will change more or less proportionately. Fixing (or ignoring) a handful issues doesn’t scale so smoothly—-there are jumps where the product suddenly seems much better/worse.
Because the entire structure of the business is designed for approximately the amount of work it currently does and likely has no particular immediate use for twice as much work in most departments.
In 20 years I have never been on a team that didn't have twice as much work as we had people to do it.
Businesses are not magically efficient
Your experience of how the world works is usually because of what work you have done. You can't grow twice as many crops, sell twice as many groceries, drive twice as many busses, because of AI - fundamentally there's a consumption problem as well.
Many businesses are not bottlenecked by processes that are computer based.
But the firms in the headlines doing layoffs after layoffs aren't growing crops, selling groceries, or driving busses... They're knowledge work roles in companies selling intangible products and services. It's large corporations doing this much more than SMBs.
They’re also the ones constantly hiring and recruiting because internally nearly everyone benefits to having more people “under them”, and there’s a massive HR/Talent team that doesn’t go into hibernation after a 20% workforce reduction. Organizations want to grow, not because they need to but because it’s in the best interest of nearly all individuals still on the inside.
And I’m sure those companies also have “backlogs" due to limited labor/labor costs. There are always shelves to face, vehicles with deferred maintenance, and so on.
Obviously, there are limits: I’m not sure what my local grocery store or bus line would do with 100 new workers, but I have no doubt they could put a few people to work right away.
I think you're viewing this from the perspective of someone who has a functioning brain and plentiful concepts and ideas that aren't being built because you're labour-constrained. Companies like Meta simply don't have productive uses for all of that human + AI labour. Meta spends tens of billions a year paying people to throw shit at the wall and see what sticks. If the idea well you're going to is running dry, AI with a smaller number of humans can slop out the stuff you do want to build more efficiently is their implicit argument, especially when you don't care about quality (as is the case with Meta). Layoffs are also being used to tell a story around efficiency to investors while companies wait for the billions they're plowing in AI actually show profit.
If one woman can produce a baby on 9 months, why can't you get 9 women pregnant and produce a baby in one month?
There's only so much to do and coordination costs (already burdensome) become overwhelming.
Except that comforting C-suite narrative does not reflect reality. 2026 agents both increase productivity by knocking clearly specified but error-prone and tedious tasks out of the park whilst simultaneously vexing and annoying their users with hallucinations and downright lies on tasks with intrinsic ambiguity. This is made worse by the token providers with their constant tweaks to their deployments to cut costs w/o losing accuracy which flat doesn't work out well for the end user.
The demand doesn't necessarily double.
Diminishing returns on additional labor.
I think there are risks:
- AI pricing is variable, probably the cheapest it will ever be right now
- AI produces a lot more shit for humans to review, and you will always need humans. If you don’t focus on keeping things simple you will probably play yourself unless you’re good at separating out blast radiuses.
- I see a lot of super low quality work that doesn’t solve the problem but it’s like look that guy solved the problem in one day! Promote him! Everyone is happy except for the end users who for whatever reason are being totally ignored (whose problem it fails to appropriately solve) and I saw this in accounting software so…hello eventual lawsuits?
AI is definitely not the cheapest it will ever be right now. The frontier is getting more expensive, but the same capability will get cheaper over time.
Why wouldn’t inference just keep getting better and cheaper as hardware and algorithms improve?
The typical playbook for a VC funded startup is to race to a monopoly where the company can have higher margins. Prices continuing to go down for the consumer over time would require competition to stay high in the long term, and even then it’s not clear if even current prices are profitable.
The current level of AI has plenty of inherent competition from local models. In the long term, most of the profit will probably be from very smart models that run at something closer to datacenter scale over long inference loops - where local inference can't do much and even third-party inference/small neoclouds will be severely challenged. That is a very natural "moat" and has natural cross-efficiencies with AI model training, which requires a similar scale.
When the AI models hallucinate up a catastrophe, managers will reevaluate that calculus.
Humans are accountable and act accordingly, models are not.
Yeah, it's probably be something like:
2025: agents are the future!
2026: we don't need any new employees
_stuff gets real_
2027: wow, employees are actually pretty good value.*
That's not always the case. Augmenting certainly can mean that, but it can also mean doing something that people couldn't do before.
For example, looking through meta data in a SQL environment that you didn't know existed to troubleshoot an issue. And a million other things. The odds of any employee not knowing everything are very good, even when humanity as a whole had already discovered that thing.
The AI story is a cover for the companies becoming less profitable in a bad economy, less able to invest because of the end of ZIRP, not capable of finding new markets, and developers becoming more expensive due to changes in accounting laws on the US.
People keep trying to make sense from the official story told by professional liars.
Because they aren't getting cut cause of AI. They are getting cut cause of the uncertainty re: future revenue streams AI is bringing. These layoffs are driven by fear and leadership lacking any vision for how the post AI world looks for their company. Their solution is to tighten the belts, which of course opens the window further for the AI companies to actually write software and make them obsolete.
I disagree. Lots of people are so laser focussed at only the close to the code aspects of programming they are unable to leverage the enormous increase in scope it can offer them.
There is absolutely room for head count reduction while companies restructure around this.
If you freeze the over-hiring, you'll still get people complaining about how there's no entry-level jobs anymore.
> Cutting people because of AI makes no sense, you know these people are good without AI, you'd want to keep them!
You can't put "undocumented knowledge" into a spreadsheet.
In reality there are two kinds of layoff going on.
Idiots who truly believe that AI will actually replace a significant amount of the workforce (Long term, it will not but some jobs displaced) and those who have internalised that ZIRP is over, they need to be more lean as VCs have closed their wallets and saying "We've revolutionised our workflow with AI" than "We haven't turned huge investment debt into profit in 5 years and we need to reduce headcount for even the slightest chance of that happening, or at least raise again by saying 'AI' over and over" has far better optics.
But we see layoff from companies that are making record profits, how does that fit here? Are you saying all of them are AI-idiots?
I would say that just the old "stupid" corporate practice of overhiring - for stock manipulation, reduce the talent pool for competitors etc. - Why Do Companies Overhire Just To Lay Off? - https://www.youtube.com/watch?v=KCF78a2wTHM
I think my read of it would be something like extreme short term thinking, and essentially like a push your luck game. If we can make it to the next model, maybe we can get enough productivity to reverse course.
As I understand it there are some underlying problems in the industry. Companies aren’t exiting which means VC’s aren’t getting returns and investment is slowing. In order for NVIDIA to maintain its current valuation, it has to keep growing at the rate it’s been growing which is the time during which companies have been piling on debt to perform extreme investments in AI. On top of this the growth rate is also based on extreme circular dealing.
So the whole thing isn’t sustainable, but the crash doesn’t happen until some big player blinks and their forecast / valuation goes down. And the only path to the kind of growth / gains people saw during the peak of the internet era is through AI.
Probably none of these big companies will fail, but it will be deeply uncomfortable for all of them when the music stops. And I wonder if it bankrupts a swath of the current generation of VCs.
That's the less common third kind. late 10s tech spending was absurd, it's the same excuse as the first but again, an easier way to reduce Capex for staff they don't need.
Preaching to the HN choir - great for upvotes but does anyone read anything new in this? Feels like I’ve seen the same article about this every month and I don’t think business people care.
more articles there will be about this - bigger the probability of this idea it will be in the dataset for chatgpt 6.9 thinking so the c-level of the companies would stop firing people
Even if you hire more people into jobs where AI acts a multiplier, there are still going to have to be layoffs where AI basically makes certain jobs _redundant_. I think both things are happening and will continue to happen, and happened anytime new technology came out.
The AI slop articles about AI will continue until morale improves.
Flag and move on. Too bad the HN audience eats this shit up by the spoonful.
You must be joking, the vast majority of HN loves AI and thinks they don't need to write any code anymore.
Some people love AI, some don't.
But the point about there being nothing new in the article still stands.
I don't see people complaining about the repetition of the dozens of daily topics about how LLMs will replace software engineers.
My pet theory is that AI enables programmers to be relatively more productive than other roles. So, if I want to grow my company, shouldn't I hire MORE programmers? Anyone know a good counterargument?
Along another vein, I guess I wonder with my limited knowledge of economics if the demand for programmers is elastic or inelastic.
Most businesses don’t grow just by churning out more units of software. At some point, it doesn’t matter how quickly you can churn out features if you’re not solving customer problems and convincing customers that they should pay for those solutions.
Once software becomes cheap, the bottleneck to growth shift to product design, infrastructure/manufacturing, sales and support.
> So, if I want to grow my company, shouldn't I hire MORE programmers?
You take the same pot of money and allocate it differently:
2010s: Hire 10 programmers
2020s: Hire 9 programmers and pay for the best AI money can buy
The 9 programmers with AI will be more productive than the 10 without.
Late 2020s, hire 5 developers and pay for the best AI money can buy.
2029, hire a developer and pay for the best AI money can buy.
2030, pay for the best money AI can buy.
At some point anything any nuanced business context a team needs and that requires humans today will be digested in AI onboarding.
if constructing houses or buildings are quick, the bottleneck will be moved to other things. Like city planner, regulations / legals, material procurements, furnishing / equiments, etc.
for software it will be requirement gathering, product planning, looking for buyer / customer, even brainstorming on finding what to make.
Kinda playing devil’s advocate here but: if AI is a multiplier it makes (now more) sense to get rid of net negatives
The companies cutting headcount because of AI often just aren’t creative enough to expand their portfolio. There, I said it.
Today, even someone with minimal context can ship end-to-end prototypes. That means existing employees, the people with actual domain knowledge should be able to innovate faster than ever. Anthropic has shipped one of the best coding agents on the market and still has hundreds of engineering roles open.
High-agency people with company context and ideas within each company should be getting amplified right now, not laid off. Instead they're stuck figuring out how to "tokenmaxx" because leadership mandated it and they are forced to ship software no one will use in order to keep their jobs.
This right here. Big companies laying off engineers are basically saying our backlogs are empty and we're out of ideas. This should represent opportunity for small/mid companies to finally compete with the large competitors as some of the advantages of scale are normalized.
I think we're forgetting the companies somewhere in the middle: letting employees use AI, and refine their internal process, and not wasting time on "tokenmaxxing" and just building better products, the companies you rarely hear about, because if you're doing things right people don't notice as much as if you mess things up.
I agree with the article in its entirety, but its just so easy to play devil's advocate here. There might actually be a large value add to companies who replace poor performing employees with AI. So, who are the low versus high performing employees?
The largest problem with software employment is defining software developer performance. There is no industry baseline for defining or measuring this. Its so easy for a person to be that 10x (or much greater) developer in a compatible team or be a complete failure on the wrong team.
The problem for high-performing developers (even the elitest) is that over a long enough time scale, they won't be remotely competitive with high performing AI. At the moment we need human operators to guide and supervise these AIs, but unless the tech's progress grinds to a halt for some unforeseeable reason, it won't be this way for long.
I know this is a common belief, but its an assumption lacking evidence.
I have been in multiple meetings where corporate folks parrot AI and productivity but if I ask for a business case and/or productivity savings, they go silent.
The funniest meeting has to be where someone said - "I have been using LLMs and it is terrible for our use case. Can we have an AI demo for Microsoft Copilot?" Imagine their surprise when they were told Copilot is LLM.
AI is just code for the actual economic reasons that companies are struggling with but it sounds better if you blame innovation instead ( Think AI = Actually Inflation, Actually Interest Rates, Actually Iran as the causes). The premise of this article is true because the companies who didn't shed jobs probably have a healthier balance sheet.
One datacenter rack of NVIDIA Vera Rubin GPUs costs $8 million. That's about the same cost as a team of software engineers. These data centers are not going to pay for themselves.
> One datacenter rack of NVIDIA Vera Rubin GPUs costs $8 million. That's about the same cost as a team of software engineers.
How many engineers do you think that can pay for? At 100k/year that's a pretty huge team.
Where are you finding engineers at 100K/yr all-in? Also how many of them will be as good at coding as the models you deploy on the rack?
Also you need both now. It’s not enough to just have humans, and that OpEx needs to come from somewhere.
I’m not happy about it - just the way it is.
Europe. 100k USD is about 75k GBP, which deducting National Insurance (our social security program) and pension brings you to a salary of £63k roughly. You can easily hire good devs for this amount (especially outside of london).
A large team at Meta (16 people) would cost the company more than $8m/year. That is all-in cost, including payroll taxes, benefits, etc.
“Next quarter is next quarter’s problem. Let’s talk about bonuses.”
— Every every single current tech c-suite
To take advantage of AI orgs needs to optimize for giving humans broader agency, not AI usage or skills.
These large companies aren’t well positioned for this. There’s too many human bottlenecks to every decision. You get in trouble for going outside your role. Thats why AI is drudgery at many of these places, often even negative productive.
Smaller companies though started assuming AI will let their employees stretch to broader roles, wear many hats, and given larger agency
The article makes a good point, but it's based on a polarized view that people make good judgement and AI is incapable of making good judgement.
I've worked with people who demonstrated below-average judgement, and I've seen cases of good judgement from AI. I think if a company can identify the poor performers and part wit them, there is a decent chance that the remaining people with AI in hand can more than make up the difference.
This is some hopium/copium that the corporate world will trend to being naturally fair and ethical.
Amazon was always an asshole-driven company which did topgrading layoffs annually at the start of every year. It has never suffered because of that.
What's great about this article is it's a falsifiable hypothesis. Is anyone keeping a list of companies in different sectors not firing due to AI? We can revisit this in 2 years and see how we did.
> AI Does Not Replace Judgement. It Multiplies It.
And that is a problem. If you employ people whomlove to comlkain are basically minus 0.1 of regular employe (tolerable annoyance), they get power multiplier and instead of small rants at water cooler, will be able to file federal law suits under employers ass.
And they will also get way more ways to harrass productive employes.
> It is the knowledge they carry. How the business actually operates. Where the edge cases live. Why certain decisions get made the way they do. What customers really mean when they complain about a specific issue. The context that never makes it into a process document because it does not need to — because the right person already knows.
I get this is about the immediate firing of people and losing knowledge quickly but there’s no reason this this knowledge can’t be captured actively and passively over the next year or so now there’s a massive incentive to do so. In fact using AI to capture this knowledge is a huge opportunity tons of startups are working on.
Relatively speaking, a company could obtain AGI first merely by just by keeping their critical thinkers while the rest of the industry offloads their thinking to LLMs. It wouldn’t be AGI, just GI, like we’ve always had before before GPT arrived
There are clearly defined tasks these things already do that are tedious and error-prone for a human. Using them correctly amplifies the abilities of humans. The problem is with the C-suite mentality towards cutting costs by replacing humans. It isn't going to work, but the only way they learn lessons like that is, as always, the hard way.
I don't think any company is cutting headcount because of AI. I think that's just what they say. The reality is that they're facing a cash crunch due to a bad economy. And if you state a cash crunch as the reason, your stock will fall or VCs will run away. But surprisingly, if you state AI as the reason for layoffs, your stock will rise or VCs will get excited. Same thing, just different clothes, and it works.
I still think AI is just a cover story for these job cuts. Tech companies are still "rightsizing" after unsustainable growth during the pandemic. At the same time we're clearly headed into a recession.
Investors like growth, not shrinkage. Claiming AI is replacing those jobs helps avoid the appearence of shrinkage, while also feeding the AI hype machine that many of these companies have invested heavily into.
Nobody ever knew what the right size was. There was a valve that was opened to allow hiring. And another "bung" that could be unplugged to facilitate firing. The algorithm was to open the valve when times are good/money cheap/greed is good, and opened the bung in other times. Totally decoupled from how many people (and largely what sort of people) are required. AI has changed that mechanism.
A friend of mine works at big corporate. He is in meetings all the time and barely codes. For him a 20$ AI subscription is more than enough.
I'm personally at a fintech startup and have a 100$ subscription. Our org is much leaner and its much faster to ship stuff. What used to be entire departments or teams can now be done by two or three people. Three people to ensure we share context and have a decent tram factor.
Its a little odd that you think these two examples are comparable. How many customers does his product have and how many does yours?
Of course a leaner team can ship faster. But your stakes are much lower. A production bug has less visibility and impact. Limited revenue means less stakeholders, and less opinions about how and what to do in a startup. Less customers means less customer support and feedback. The differences go on and on. I'm not sure you can attribute all of that to AI assistance.
Good luck with your fintech startup. I'm in that space too and its a bear.
The problem that software engineers and product teams often face is that the time from roadmap to feature is quite long, and AI has offered a clear, meaningful speedup to some parts, like writing tests and boilerplate. In many cases, the same ticket flow can be managed much faster with AI. So, what? We're not quite at the point where people have transitioned this into demand for more product deliveries faster. As soon as that occurs, it doesn't matter how great the AI is, because the current pace will be slow. Why stop at what the current roadmap is? Why not ask for double the features? When we gave everyone power tools that didn't dramatically make construction easier, it just enabled more complex buildings.
The misunderstanding is that people equate what these companies are saying and what they are doing. In reality this is the perfect time to fire people under cover. This happens every X years where as leadership you can say this tech is displacing you, when in truth its just that you arent needed anymore.
This is a capitalism functioning correctly as labor as in then moved to places it is needed more. Its sort of shocking to see devs complain about getting laid off. Thats the point of a high variance career. Great and terrible outcomes go hand in hand, if you want one you must expect the other.
I don't think software was ever sold as a high variance career. Start a startup! was sold that way, particularly by PG's early essays, but not not W2 employment. You do W2 because you want low variance, and you either love software or it pays well. Either way, I don't think most people were okay with terrible outcomes. (I don't think getting laid off is terrible, I think it should be an expected possibility. I've found that contracting/freelance has helped reduce my fear of losing my job, since you expect the job to end when you start it, and since you're regularly looking for a new job, you get a sense of how long it takes.)
Simply reality is very few companies are cutting because of AI. They’re just trimming bloat that’s been there the whole time. The AI narrative is broadly just PR spin and CEOs trying to look cool.
These two ideas, that (1) AI is making people more productive and (2) we are laying off employees because AI replaces them. Those two ideas are mutually incompatible. When employees become more productive, you always always always want to hire more of them, because it’s a multiplier on your outputs and therefore an efficient use of resources.
That is why one of these narratives is total bullshit. Either “layoffs from AI” are layoffs for some other structural reason in the tech sector, or “AI makes employees way more productive” is total bullshit.
I don’t know which is the scarier lie tbh
This assumes that demand is elastic and unbounded. It might be true that there's a lot of untapped demand for software, but especially considering individual firms, it's plausible that demand for their offerings might be capped. In this case it does make sense for firms to shrink.
Analogously, automation and productivity improvements in farming drastically decreased employment in the farming sector, from basically everyone to basically no one.
This is a simple economic fallacy. Your idea holds true if your business has room for growth, but you always need a market for those multiplied outputs. In an economy where every business has multiplied outputs, but demand doesn't scale with those outputs, what do you think happens?
Which companies are cutting headcount for AI though?
You mean companies blaming AI or using it as an excuse?
The real answer is stop reading the headlines or the fake news.
> Not because they have fewer costs. Because they have more capability.
This kind of sentences are typically AI-generated. What’s the point of an article about AI generated by AI?
If the article is about the craft of writing, then no point. But if the article is about communicating ideas, AI is just the means the person with the ideas chose.
However, in this case I think both were fairly vacuous. Also, sentence fragments are a hallmark of bad writing that high school English teachers try to train you out of, apparently unsuccessfully if the training data makes them so likely.
LOL who doesn't?
I don't know anyone who has been hiring in any kind of significant numbers. Everyone fires either because of alleged AI efficiencies or to pay for AI token shit or to just bla bla to investors how magically AI AI AI they are.
End result all the same. Massive job losses due to AI.
Greed is the underlining issue.
That's absolutely correct. I use LLMs heavily in my work and open source projects. These tools are effective, they can solve hard problems and they allow me to work on a wider range of tasks than could before.
But, they're also jagged in terms of functionality. When you work with a human, you can learn what their core competencies are, and then if you give them a task that falls within that domain, you can be reasonably sure they'll finish it correctly. That's not the case with LLMs. It might do one task brilliantly, and a next similar task, it just shits the bed on.
And since it has no understanding of the task in a human sense, it can't self correct, learn or improve. All its doing is stringing tokens together based on probability. So, you need a human in the loop to review what it's doing, to correct it when it makes mistakes, and to define the actual goals. My experience is that doing all that properly ends up taking up a lot of time, so your actual productivity gain per person aren't all that significant.
Companies that try to replace humans with LLMs will soon find that they end up with a whole bunch of code that doesn't actually work, and they have no hope of fixing. The double edge of LLMs is that they're really good at generating a lot of wrong code really fast.
This is a pro-AI piece. They realized they overplayed their hands when college graduates started booing them and their new line is "AI won't replace workers, it will make them more valuable!"
Another bury-your-head-in-the-sand article. It sucks but AI replacing coders is very real. I haven’t coded since last year but I’ve pushed more features at a rate faster than I ever could in my entire career. I pushed new code while chatting on WhatsApp with my friends yesterday using 2 prompts.
“Remove this old feature.” “Are you sure you didn’t break anything?”
That was it. Then I manually tested it to make sure nothing was broken. Then I did a brief review before posting it for code review and then pushed it. What would have taken me probably about 1 day to go through and figure out code changes and then actually change the code took me about 30 seconds.
To think that we need to maintain the post-Covid hiring bloat is nonsense. I’m not so arrogant to think that someone with an llm can’t replace me, if I survive a few more years in this industry I’ll be amazed and grateful.
Has this ever happened? Companies are just keep going along. There's these consequences for things that hurt people
Capitalists usually recoup gains rather than re-investing in people or working in new verticals. This gels with companies who would rather do anything else than grow.
We need stronger goverment and socialist revokution! Strong socialist goverment invest it to people! Socialist goverment would not starve people, torture, or just send them somwehere else (those examples do not count)
Wasn't really trying to take the socialist view. Merely stating what usually happens. Grow the business usually means make more money in the easiest way possible (cost cutting usually) rather than invest in future gains.
I once was the final step in an interview process. My flag to the team was that this person had been laid off something like four times in a row. I said, layoffs don’t cut the great people, so it says something about the applicant.
The team insisted this person had a big heart, a lot of passion, and they’d make a difference. We extended an offer and they accepted.
They ended up being unremarkable in every way. It was a drag on the company. We let them go a year later and set ourselves back on a topic our customers cared about.
Anyway, some people are great. Some are not. The idea that there’s this hidden special knowledge trapped inside teams, if only the executives would unleash it, is nonsense. I personally feel for anyone who loses their job, but companies will always pick the most productive and efficient path to a result.
Also so does the labor market. How many people, when given the option for more pay and benefits, decline out of loyalty to their employer? At least in tech, this is almost never the case.
I think we’re a bit lost. On one hand, certain people don’t trust these big employers and think they’re evil, and on the other hand really really want their job at said evil megacorp.
Companies are cutting headcount to fund AI projects, not necessarily because of the gains they are getting from it. And frankly in many companies, cutting layers and reducing team handoffs will be a net gain - even if AI didn't exist.