If mainstream corporate America thought it would turn increasingly cheap intelligence into higher revenue growth and higher profit margins, then mainstream corporate America needs to go back to school.
If you want to increase your profit margins, you'll have to innovate! Adopting AI tools that somebody else created, and are equally available to your competitors, won't increase your revenue or margins. The ones who will see increased revenue and profit margins are the ones who actually innovated, by creating the AI.
This is exactly what Apollo reports: the Magnificent Seven increased their profit margins, while mainstream corporate America didn't.
> The only way more efficient operations doesn't make margins go up is if prices go down.
But in general, it is not true that more efficient operations + unchanged margins = observable lower prices. Especially when the mere existence of a technology can make non-adopters' processes overall less efficient. For example, if there is a RAM shortage, and storage becomes more expensive, then SaaS subscription prices will increase, even if all vendors adopt AI to make their operations more efficient.
Not really. Like, yes, relative to your competitors. But, if the cost of growing potatoes is ten cents per potato, and then I make something that can produce potatoes for one cent then I'd expect everyone to just have more potatoes. This would drive down the cost of potatoes, because the supply would be saturated eventually, but AI is supposed to just make more of everything. So, if AI works as advertised, then you'd probably expect a brief period of highly increased profit margins (by the consumers of AI) followed by rapid _deflation_ as the value of everyone's dollar increases exponentially. Of course, neither of those things is going to happen, because the AI sales pitch is actually a pyramid scheme, so you get the people at the top raking in cash, while everyone else tries to sucker the next tier of buyers so that they can make their buck.
I think the problem is moreso that making useful things for enterprise with AI right now is actually pretty hard because context and attention management is hard, the token cost for complex tasks scales fast, and model's lack of intuition for fuzzy decision-making.
There are hordes of grifters promising things they can't deliver and reaping in millions in capital from investors who are equally confused about the challenges of designing useful AI.
That said, AI can do useful things when properly designed. The market finding the people who can make this happen won't happen as fast as people expect.
My biggest doubt isn't whether AI will find applications for speeding up people outside of tech, but that it will atrophy learning and cognitive development to the extent that in 20 years human ability without AI will be on average much lower than it is today. Overall with AI augmentation we may be more effective, but strip away that tool and the human left behind will be less
I have a suspicion that, outside of tech, AI WILL speed up the mundane aspects of people’s jobs, that part which they spend 80% of their time getting through, so they can then do the 20% where they actually add value.
BUT the new time they unlock will be spent doing more of the mundane work that previously just didn’t get done because it was overtaken by events or it got done to the absolute barest minimum level.
As a result the 20% where they add value never really grows or changes and in fact might be less valuable because of all the extra noise created by doing more tasks in the 80%.
In 5-10 years those tasks may be done by humanoid robots. Replacing cleaners with robots is already here. In many supermarkets and warehouses robots wet and dry clean the floors. A task that would take humans hours to do.
We have heard about humanoid robots for nearly 20 years from people like Boston Dynamics, and they still cant do the most basic tasks and cost millions.
The only robots available today can barely bring a glass of water and they cost crazy amounts of thousands of dollars.
> Replacing cleaners with robots is already here. In many supermarkets and warehouses robots wet and dry clean the floors.
What do you think would happen to a hospital that only cleaned the floors ?
There are far more important places that need cleaning besides floors.
True - but it doesn't need to be a humanoid robot - there was a great video of a person that built a 'picker' - essentially a crane in their bedroom that could tidy for them. Why make it look like a human, when better and cheaper to focus on solving the problem?
To act as the Devil's Advocate here, I suppose the very high valuations for the AI companies are because some people think AI will solve those through allowing robots to do that work. Although, Healthcare is a lot broader, and seems to me to be much more complicated than the others. Surely less solvable than the others in your list - if any are in the near future.
Healthcare is probably the one that we need some sort of help with the most! I think the need for people (fine - or robots) working in Healthcare will grow faster compared to the others in the future because of population aging.
Edit: Clarified my meaning by adding "some people think" to the sentence "...because some people think AI will solve..."
Once the OS and app hooks are in place, I think its biggest win is going to be plain language automation. Even if we don't have it write all of our emails and report, simply being able to say "show me the emails Bob sent about the project" or "did Bob ever say he was going to call Phil?" and get an accurate answer will be huge. We're approaching that in programming, when I can use general terms to automate operations within a codebase. The dream of "programming" with plain language may finally be realized.
If AI helps with the mundane side of real work, great, but outside of mundane-only work, there generally isn't a conveyor belt of random mundane work backfilling the real work's now-fixed-by-AI mundane bits. Any mundane work that's gone unaddressed for long is unaddressed because management doesn't think it'll benefit the business, and if it didn't benefit the business yesterday, clearing out some unrelated mundane work doesn't automatically increase the priority of that unhelpful-to-the-business backlog of mundane work.
The mundane tasks are only about 20% of the work for white collar works. AI just eliminates the low-value mundane tasks.
Unless AI companies are willing to do the impossible and accept liability for incorrect responses/outcomes, they're never going to be trusted with any task that could incur liability (law, accounting, medicine, architecture, engineering).
Could humans review AI output? Sure. You still need humans. Funnily enough wages won't decline in the long-run - they'll rise as firms compete for those who remained employed in said roles as others fell out of favour and have gaps in work experience.
Many, many variables at play. I never take anyone seriously who never considers this (many people on here).
Wages only rise if you don't include the displaced in your average, which feels a bit dodgy in this context, even if it is technically the accepted practice.
Wages will only rise if labor power rises relative to corporate power.
If that doesn't happen, then companies will just continue to make excuses for why they need to lay off 20,000 workers after having their best quarter ever.
> that part which they spend 80% of their time getting through, so they can then do the 20% where they actually add value.
Are we still telling ourselves this lie? Nobody appreciates anyone else's job, not until it's too late. There must be a name for this phenomenon, where everyone thinks that everyone else's job is 80% pointless.
> BUT the new time they unlock will be spent doing more of the mundane work
...which will also quickly be done by "AI", uncovering more such tasks until most of the "mundane work" is done by "AI" leaving humans free to do "non-mundane work". Of course those less mundane tasks require more insight, intelligence, knowledge and initiative from those humans which means some people will just not be able to keep up with their competitors, the humans who can do the less mundane tasks and the "AI" which can do the mundane stuff faster, more reliably and cheaper.
The big question is what those humans will do once such a scenario becomes true.
The cost of training/developing a frontier model will continue to rise. The cost of inference will continue to fall. Companies don't buy models as such, they buy inference. As inference becomes a larger and larger business expense companies will optimize that spending, routing work to the "capable enough" least cost tokens.
There is a lot of human work that can be offset with tokens. On the other hand, the amount of work for the highest intelligence models will remain limited - corporate strategy, major decisions, etc. are just not as frequent.
The highest “intelligence” models can’t even handle basic tasks right now. Saying that only “corporate strategy” and “major decisions” will need those models is laughable.
When software engineers get put out of work, they'll move to other industries. They'll take their engineering mentality/skills and their affinity for Claude and they'll look at their menial, boring work they're handed and they'll say... I know... I can automate that!
The economy has many different resources with different supplies available. Making the supply of one resource infinite may sound like a huge win, but nothing is dependent on a single resource, everything requires a combination – just like in board game economies, you can't win by only producing wood.
AI represents some weird intelligence-like resource. It clearly does something of value, but that must be combined with other resources. In tech there's a lot of that, so making it (roughly) infinitely accessible has had a nice impact and we've run into our next limiting factor: human time. Companies that aren't "pure tech" though will run into much earlier limiting factors – hardware, compliance, regulation, artistic value, human time, and so on.
I suspect AI has delivered financial value but AI psychosis has cancelled it out.
It's also possible that AI has created value for consumers, but the profits flowed to the AI producers (due to scarcity) not the middle layer of the value chain, due to completion.
If mainstream corporate America thought it would turn increasingly cheap intelligence into higher revenue growth and higher profit margins, then mainstream corporate America needs to go back to school.
If you want to increase your profit margins, you'll have to innovate! Adopting AI tools that somebody else created, and are equally available to your competitors, won't increase your revenue or margins. The ones who will see increased revenue and profit margins are the ones who actually innovated, by creating the AI.
This is exactly what Apollo reports: the Magnificent Seven increased their profit margins, while mainstream corporate America didn't.
> Adopting AI tools that somebody else created, and are equally available to your competitors, won't increase your revenue or margins
The only way more efficient operations doesn't make margins go up is if prices go down. Is that what we're seeing?
Answering your factual question: Yes, of course, software is cheaper in real terms now, even despite high inflation! See e.g.
https://www.accelerance.com/blog/2026-outsourcing-rate-trend...
https://www.reuters.com/world/india/ai-reshapes-indias-it-se...
> The only way more efficient operations doesn't make margins go up is if prices go down.
But in general, it is not true that more efficient operations + unchanged margins = observable lower prices. Especially when the mere existence of a technology can make non-adopters' processes overall less efficient. For example, if there is a RAM shortage, and storage becomes more expensive, then SaaS subscription prices will increase, even if all vendors adopt AI to make their operations more efficient.
Not really. Like, yes, relative to your competitors. But, if the cost of growing potatoes is ten cents per potato, and then I make something that can produce potatoes for one cent then I'd expect everyone to just have more potatoes. This would drive down the cost of potatoes, because the supply would be saturated eventually, but AI is supposed to just make more of everything. So, if AI works as advertised, then you'd probably expect a brief period of highly increased profit margins (by the consumers of AI) followed by rapid _deflation_ as the value of everyone's dollar increases exponentially. Of course, neither of those things is going to happen, because the AI sales pitch is actually a pyramid scheme, so you get the people at the top raking in cash, while everyone else tries to sucker the next tier of buyers so that they can make their buck.
I think the problem is moreso that making useful things for enterprise with AI right now is actually pretty hard because context and attention management is hard, the token cost for complex tasks scales fast, and model's lack of intuition for fuzzy decision-making.
There are hordes of grifters promising things they can't deliver and reaping in millions in capital from investors who are equally confused about the challenges of designing useful AI.
That said, AI can do useful things when properly designed. The market finding the people who can make this happen won't happen as fast as people expect.
My biggest doubt isn't whether AI will find applications for speeding up people outside of tech, but that it will atrophy learning and cognitive development to the extent that in 20 years human ability without AI will be on average much lower than it is today. Overall with AI augmentation we may be more effective, but strip away that tool and the human left behind will be less
Literally the argument math teachers make about calculators.
Notice how even with such advanced cognitive atrophy, the patient is still able to repeat that worn-out comparison.
And math teachers have been correct. So, if anything, that suggests that the parent post is also correct.
I have a suspicion that, outside of tech, AI WILL speed up the mundane aspects of people’s jobs, that part which they spend 80% of their time getting through, so they can then do the 20% where they actually add value.
BUT the new time they unlock will be spent doing more of the mundane work that previously just didn’t get done because it was overtaken by events or it got done to the absolute barest minimum level.
As a result the 20% where they add value never really grows or changes and in fact might be less valuable because of all the extra noise created by doing more tasks in the 80%.
This is working theory and likely full of holes.
How will AI do mundane work like:
- Cleaning - Picking up rubbish - Stocking shelves - Healthcare - Fruit picking
Thats only the start, while a lot of people do work on a computer, far more dont.
In 5-10 years those tasks may be done by humanoid robots. Replacing cleaners with robots is already here. In many supermarkets and warehouses robots wet and dry clean the floors. A task that would take humans hours to do.
We have heard about humanoid robots for nearly 20 years from people like Boston Dynamics, and they still cant do the most basic tasks and cost millions.
The only robots available today can barely bring a glass of water and they cost crazy amounts of thousands of dollars.
> Replacing cleaners with robots is already here. In many supermarkets and warehouses robots wet and dry clean the floors.
What do you think would happen to a hospital that only cleaned the floors ?
There are far more important places that need cleaning besides floors.
True - but it doesn't need to be a humanoid robot - there was a great video of a person that built a 'picker' - essentially a crane in their bedroom that could tidy for them. Why make it look like a human, when better and cheaper to focus on solving the problem?
https://hackaday.com/2026/06/15/building-a-ceiling-based-cra...
To act as the Devil's Advocate here, I suppose the very high valuations for the AI companies are because some people think AI will solve those through allowing robots to do that work. Although, Healthcare is a lot broader, and seems to me to be much more complicated than the others. Surely less solvable than the others in your list - if any are in the near future.
Healthcare is probably the one that we need some sort of help with the most! I think the need for people (fine - or robots) working in Healthcare will grow faster compared to the others in the future because of population aging.
Edit: Clarified my meaning by adding "some people think" to the sentence "...because some people think AI will solve..."
Once the OS and app hooks are in place, I think its biggest win is going to be plain language automation. Even if we don't have it write all of our emails and report, simply being able to say "show me the emails Bob sent about the project" or "did Bob ever say he was going to call Phil?" and get an accurate answer will be huge. We're approaching that in programming, when I can use general terms to automate operations within a codebase. The dream of "programming" with plain language may finally be realized.
Its not even a theory tbh. Its just a statement - utterance.
If AI helps with the mundane side of real work, great, but outside of mundane-only work, there generally isn't a conveyor belt of random mundane work backfilling the real work's now-fixed-by-AI mundane bits. Any mundane work that's gone unaddressed for long is unaddressed because management doesn't think it'll benefit the business, and if it didn't benefit the business yesterday, clearing out some unrelated mundane work doesn't automatically increase the priority of that unhelpful-to-the-business backlog of mundane work.
The mundane tasks are only about 20% of the work for white collar works. AI just eliminates the low-value mundane tasks.
Unless AI companies are willing to do the impossible and accept liability for incorrect responses/outcomes, they're never going to be trusted with any task that could incur liability (law, accounting, medicine, architecture, engineering).
Correct.
Could humans review AI output? Sure. You still need humans. Funnily enough wages won't decline in the long-run - they'll rise as firms compete for those who remained employed in said roles as others fell out of favour and have gaps in work experience.
Many, many variables at play. I never take anyone seriously who never considers this (many people on here).
Wages only rise if you don't include the displaced in your average, which feels a bit dodgy in this context, even if it is technically the accepted practice.
[flagged]
Not interested in conversing with a person who can't even keep their previous comment in their context window.
Good day, sir.
Wages will only rise if labor power rises relative to corporate power.
If that doesn't happen, then companies will just continue to make excuses for why they need to lay off 20,000 workers after having their best quarter ever.
> that part which they spend 80% of their time getting through, so they can then do the 20% where they actually add value.
Are we still telling ourselves this lie? Nobody appreciates anyone else's job, not until it's too late. There must be a name for this phenomenon, where everyone thinks that everyone else's job is 80% pointless.
> BUT the new time they unlock will be spent doing more of the mundane work
...which will also quickly be done by "AI", uncovering more such tasks until most of the "mundane work" is done by "AI" leaving humans free to do "non-mundane work". Of course those less mundane tasks require more insight, intelligence, knowledge and initiative from those humans which means some people will just not be able to keep up with their competitors, the humans who can do the less mundane tasks and the "AI" which can do the mundane stuff faster, more reliably and cheaper.
The big question is what those humans will do once such a scenario becomes true.
The cost of training/developing a frontier model will continue to rise. The cost of inference will continue to fall. Companies don't buy models as such, they buy inference. As inference becomes a larger and larger business expense companies will optimize that spending, routing work to the "capable enough" least cost tokens.
There is a lot of human work that can be offset with tokens. On the other hand, the amount of work for the highest intelligence models will remain limited - corporate strategy, major decisions, etc. are just not as frequent.
The highest “intelligence” models can’t even handle basic tasks right now. Saying that only “corporate strategy” and “major decisions” will need those models is laughable.
Previously: https://news.ycombinator.com/item?id=50224463
609 points by the-mitr 1 day ago | hide | past | favorite | 645 comments
When software engineers get put out of work, they'll move to other industries. They'll take their engineering mentality/skills and their affinity for Claude and they'll look at their menial, boring work they're handed and they'll say... I know... I can automate that!
The economy has many different resources with different supplies available. Making the supply of one resource infinite may sound like a huge win, but nothing is dependent on a single resource, everything requires a combination – just like in board game economies, you can't win by only producing wood.
AI represents some weird intelligence-like resource. It clearly does something of value, but that must be combined with other resources. In tech there's a lot of that, so making it (roughly) infinitely accessible has had a nice impact and we've run into our next limiting factor: human time. Companies that aren't "pure tech" though will run into much earlier limiting factors – hardware, compliance, regulation, artistic value, human time, and so on.
It's because other industries are buying expensive tokens, and there's no real ROI yet.
I thought tokens were already sold below what they cost the AI companies, no?
AI is the anti midas golden finger. Everyone who touches it is down, and those at the center are down hundreds of billions.
AI is the mierdas touch
How does this compare to financial payoff from Bitcoin and NFT adoption?
I suspect AI has delivered financial value but AI psychosis has cancelled it out.
It's also possible that AI has created value for consumers, but the profits flowed to the AI producers (due to scarcity) not the middle layer of the value chain, due to completion.
[dupe] Discussion on source: https://news.ycombinator.com/item?id=50224463
[flagged]