The AI Layoff Trap
arxiv.orgThe trick is bypassing the human consumer as well. Companies satisfy (human) consumer needs as a byproduct of profit maximization. But human consumers are inefficient. They have to sleep, require medical care, etc.
A purely machine economy would be far more efficient. Therefore in the limit we should eliminate reliance on human labor and consumption to build a more perfect and efficient world.
Idea! Maybe these now redundant humans can be turned into a kind of battery, so they serve as a source of energy for the machines?
Perhaps it's then smart to make the humans have a brain/computer interface, to make then dream/think they are living in a normal society so they don't revolt.
Agents everywhere!!
Do you like what I've done with the place?
A tragedy is that supposedly the original idea was to use human minds for processing power, which would have been far superior to the thermodynamically laughable idea to use them as power sources.
You jest, but isn't this the logical conclusion? A sufficiently smart AGI has no need for humanity, at all.
I know people like to make Matrix jokes about what the future of AGI would be like, but I suspect it would be more like the relationship between humans and Cheela in Dragon's Egg.
But then, what need do we have for the AI?
At this rate, I'm convinced our future AI overlords will put humanity down much more humanely than we are/will
viz. Iain M. Banks culture, where the AI keep humans as pets
The "economy on a chip" thought experiment .
The humans consume to fulfill needs, how do those needs get fulfilled in a post human economy?
Large populations of humans not getting their basic needs fulfilled is a feature. That way, those who do have lots of resources get to have even more intense feelings of superiority and that of having high status. That is, after all, the point of wealth accumulation (at least a certain point where more money does not change your life quality meaningfully). For status, its the delta between you and others that matters most. If raising your own wealth has diminishing quality of life returns, then lowering other peoples is the only effective way of increasing the delta.
Elon way overpaid when offering that lady a horse in exchange for an erotic massage
Having large amounts of people with unfulfilled needs is not exactly a novel idea.
Just train machines on the huge corpus of human needs so they can need things like no human has needed things before.
What can possibly go wrong
>The humans consume to fulfill needs
That's not how the capital class thinks of human consumption.
I use AIs for coding with moderate success, but the more I work with them, the more I am convinced that "intelligence on tap" is a pipe dream, especially in domains where logical thinking in novel (ie not-in-dataset) contexts is required.
Recently, I tasked it to study a new Czech building permit law in conjunction with some waste disposal regulations and the result was just tragic. The model (opus 4.6) just could not stop drawing conclusions from obsolete regulations in its training dataset, even when given the fulltext of the new law. The usual "you are totally right" also applied and its conclusions were most of the time obviously wrong even to a human with cursory knowledge of the subject.
I ended with studying the relevant regulations myself over the weekend.
I wonder what percentage of the job space truly depends on the current edge we have over machines.
I think it's reasonable to worry that way before machines are more reliable than the average human (let alone more reliable than a highly trained human) they can pose a significant disruption to the job market which will send shockwaves throughout society
That is why we need functioning states -- free markets won't save you in such a case. Though I found it is hard to explain especially to U.S. people, who put "regulation" on par with f words :)
"The model (opus 4.6) just could not stop drawing conclusions from obsolete regulations in its training dataset"
To be fair, humans are also often like this. If some rule/law/model was deeply ingrained into them, they often cannot stop thinking in terms of that rule, even if they are clearly in a new context (like a new country).
When the mandatory speed limit in my country was reduced from 60km/h to 50km/h in cities, 95 percent of people instantly adapted.
But that is pretty much the same rule, just the numbers slightly adjusted. What do you think would happen if they changed traffic from the right to the left lane?
Heh, that would be surely funny :) But most people at least know there is a new permit law and if they are not sure, they are to seek expert guidance. The model is even with explicit notification unable to reflect upon this fact. How it is supposed to be useful then?
Oh, most people would know in theory for sure, but if they go into driving, habit would kick in and they end up on the wrong lane pretty quickly.
At least that is what happened to me in australia and I only had a year of driving practice back then, but driving on the right side was already deeply ingrained and I had to be really aware of what I did.
But to be clear, I am not arguing models have real understanding of anything - I know they don't. My point was humans can be similar in pretending to have understood something, but if their core was modeled different, they will fall into old patterns again quickly.
neo-luddism dressed up in economic jargon. the authors suggest the only effective tool is to tax companies based on how much automation they achieve. Penalizing efficiency is a guaranteed recipe for stagnation and if we'd done this at any point in our past we would have not made it out of the dark ages
I don't disagree. However, if we change nothing, one likely alternative outcome is planet of 100 trillionaires, 10000 concubines, and an ever-shrinking ghetto of scavenging paupers. The solution is to turn the un-meritocratic nature of this particular bit of technical change against it.
As the value of labor plummets, more GDP will accrue to capital. But to whose, exactly? Let's categorize individual investing performance as a function of luck, corruption and skill. Only skill is meritocratic, and there is no good reason to reward the other two. Things have trended away from skill in recent decades.
As AI automation progresses, it provides more of the skill. Eventually all investing decisions will be AI-based, democratizing the process but effectively leaving luck and corruption in control of who wins.
At that point, there's just no good reason to reward individual investment performance. Since luck averages out, corruption will largely determine who the 100 trillionaires are.
The solution is to tax away the portion of investment returns that are not based on skill, which will trend towards 100%.
Note that it's usual that companies gets tax on surplus, not income. So we are already 'punishing' the efficient ones, we are just doing it in a relatively neutral way.
In systems with progressive income tax, the total tax income from a company with 1 employee making X$ is more than if the company had 2 employees making X/2$, so essentially 'punishing' using highly skilled labour over more less skilled ones.
There are no perfect taxes, and current tax systems have adapted from a lot of practical concerns. Some of those is that's it's easier to tax money as they are moving around vs when they are sitting still (wealth or property tax), and it's easier to tax people than abstract entities like companies, since people have a harder time moving. And for the same reason, it's easier to tax the middle class than the owner class, since the richer you are the easier it is to move yourself and those you care about to wherever taxes are low these days.
All these practical concerns have made it such one of the most common ways for the state to get a share of the productivity of its society, is from income tax. But this is not a 'economic law' that if must be like that. If more and more of the productivity and wealth creation in society is produced such that there is little employment income involved, we will have to find other ways to tax it.
Your not looming at anything after the first order effects. The idea of work os that people participate in the economy, what does a post work economy look like? How do people have the cashflows necessary to participate in things like housing and food and stuff when their way of contributing to the economy was automated away?
I don't believe there will be a "post work" economy. Some people will turn to farming, others will turn to crime and mischief (hunter gatherers).
I expected when we actually get to that point, we'll have an even worse version of the K-shaped economy.
It's not penalizing efficiency, it's taxing a commodity.
Everyone is efficient, thus, the efficiency has to support what we'd call our economy/society.
I can foresee a token tax in the future in some ways, but then that would kill private LLMs if hat came true so i'm not sure anyone has an answer, but neo-luddism seems short sighted.
I agree, and I often wonder why companies pay any tax at all, rather than just hitting the shareholders as their wealth grows. There was a post a few months ago about taxing unrealized gains that was very interesting I thought.
The company itself has an impact on our society and needs to be "governed", so it seems reasonable for them to pay for that governance. Actually, it seems unfair that you can claim no profit and get out of paying for that governance.
I'd be really interested to know if companies pay the true cost of their impact to society, or if individual income tax has to pick up the tab.
Companies do generally act as tax collectors rather than the final bearers of the cost, as taxes just increase the price for the end consumer.
Let companies manage to automate 100% of their processes by having their own country/law/constitution!
Companies produce goods which people consume. If you hand everything over to the oligarch class, how will people consume products built by companies???
The Chicago School would probably hate it
You haven't thought this through.
It’s funny, I think most people roll their eyes when Trump says things like “you’ll be tired of winning, you’ll say ‘please no more winning’”.
But recommendations to tax efficiency are unironically that (just dressed in more serious language). “Please stop giving us what we want so efficiently, we want to work more for it!”
Those winning and those asking for this to stop are two different categories of people. The former are the capital holders and the latter are those with no source of capital in their future. If you can merge the two categories, we may talk again. Until then, you need to come up with something better than trickle down in a world where there is no trickling.
> But recommendations to tax efficiency are unironically that (just dressed in more serious language). “Please stop giving us what we want so efficiently, we want to work more for it!”
You're trying to make it sound ridiculous, but most people aren't pure consumers. They're laborers and consumers. Policies that hurt while wearing the consumer hat may be more than justified by the benefits while wearing the labor hat.
If robotics progress starts to pick up, I'll take this more seriously. Right now, there's practically infinite demand for labor in construction, manufacturing, agriculture and many other industries. All kinds us good projects that could be happening, if you dig into why, labor intensive work is a factor. Why didn't the hydroponics project take off? Why is that still an empty lot instead of a new home? Why isn't there live theatre in this small city? Why is there a pot hole in the bike lane?
Isn’t this more a function of how the American construction market is just really messed up somehow (corruption?)? In China, actual things get built fairly cheaply and quickly. You just don’t see workers hanging around watching one guy dig a hole like you do in the states. I would guess that automation is the only way out of the mess we are in, since just throwing more money and people at the problem just seems to make it worse.
What’s different about the market in China that enables this?
The usual answer is slave labor, like in the middle east. But some combination of an extremely poor job market with laborers that can't leave, and can't do anything else.
Could you expand on the corruption claim?
Trades seem to have high barriers to entry and have stringent unproductive working rules. I’m not really sure, but does to make sense that construction prices have risen so much so fast without even considering the cost of materials? The public sector is much worse, of course, where a short jaunt tram at LAX costs more than a 75 mile HSR run in China. We obviously aren’t competitive in building things anymore.
Somewhat related, but I know construction is the one of the only industries that has gotten less productive in the past 100 years. You'd think more machinery and such would make it more productive, but no.
I think part of that is safety and labor cost, but I think another part is that construction is contracted almost always.
n = 1 …
Go to Miami, Florida, and see how virtually all public projects magically go to Cuban-American-owned companies — even huge multinationals with far greater skill, capacity, and efficiencies can't seem to land the good work.
Infinite demand, maybe, but not at wages that most people are willing to accept. Of course, if there's literally no other work, then previously-middle-class people will take what's available and become homeless because the wage doesn't pay the bills (which are, in places, extremely inflated due to decades of jaw-droppingly bad housing and transport policies). Sounds like a highly desirable future.
Yes, but thanks to Baumol's cost disease productivity increases in other sectors can have spillover effects in terms of wages.
That assumes the current high wages are here to stay. This seems unlikely if AI consumes most white collar jobs.
Um, I expressly said that high wages wouldn't stay? If the choices are either being jobless and homeless, or doing some menial cotton-harvesting job while still being homeless, we got a slight social problem. The GP said that there's a lot of demand for menial labor. That demand only exists if you don't have to actually pay for said labor. In other words, it's not demand at all.
>> If robotics progress starts to pick up, I'll take this more seriously. Right now, there's practically infinite demand for labor in construction, manufacturing, agriculture and many other industries.
Don't be so sure: https://www.nytimes.com/2026/04/08/business/economy/blue-col...
Lucky not to live around small towns that were killed by the introduction of robotics?
Yes there is some demand for labour in fields like agriculture , and many rather not pick the work and survive elsewhere, because feudal lords rather pay peanuts for the hard work.
So you work in one of these fields, right? Hydroponics, homebuilding, theatre construction, pothole repair?
I currently work as a software engineer, but I've worked in the past in restaurants (dishwasher/prep cook), doordashing, as a musician, as moving help. If AI automates software I'll just do something else.
Spoken like a 2007 bond trader
All of those are basically low wage garbage jobs. We have little to no respect for them as a society, too, to add insult to injury.
If it were up to me, I'd be flipping burgers right now. But I literally can't do that, it doesn't pay enough for that to be a job I can take. Id have to, like, find roommates.
Good luck with physical labor when you reach your 50s.
Fair point
Yeah I am always disappointed in how little there is automated in construction and how slow humans are in this activity. It feels like an exclave of the Dark Ages in the Information Age.
Construction is interesting because productivity has actually fallen in recent decades.
Safety and quality have increased.
AI layoffs are very shortsighted IMO and should be viewed by investors as a sign of weakness in management or the business itself.
If everyone is going to increase productivity by some factor k per employee, then kx is the new norm of overall productivity of x employees.
If you lay off some percentage Y of your work force, then your expected gains will only be k(x(100-y)/100). In other words, you will not recognize the same productivity gains as your competitors that chose not to lay off.
Yes I realize it is more complex than that, because of reduced opex, but there are diminishing returns very quickly.
There are productivity gains to be had by reducing amount of communication and internal layers IMO.
I believe there are also diminishing returns on new features/products (in general case). So you won't really need that many people.
Let’s take AGI to its inevitable raw conclusion. Not by the definition (ab)used by clueless VCs screaming about abundance, but by what is already happening using the worst case:
The abundance of mass layoffs and job displacement due to funding and building of AI systems is the true definition of AGI.
We might as well get there faster instead of delaying it. You have already seen Oracle and Block attributing their layoffs to AI so it is happening right now.
So why delay any further and just get it over with.
Get where faster? Get what over with?
Aren’t you talking about destroying livelihoods? Pushing people into poverty and/or homelessness? What is the benefit exactly?
I guess the argument would go that a new economic model will be required at that stage.
There isn't much point in having people do jobs they don't like which are trivial to automate just for money, but at the point where there isn't enough economically useful things for everyone to do, the current system falls down.
> What is the benefit exactly?
Well one benefit would be international competitiveness. The country that does it slowest will be the country doing more work for less output.
> but at the point where there isn't enough economically useful things for everyone to do
This assumes that for example a person who has been an artist for 20 years, can easily enough switch professions to a machinist, and the only reason for them not to do it is because the economy has no need for another machinist. An insane way to think. This is not how humans work.
Let me see any HN dweller go from their cushy home office to butchering animals for meat on 12-hour shifts for example... Oh and btw, no safety net to give you food, housing and healthcare while you learn the new craft!
> Let me see any HN dweller go from their cushy home office to butchering animals for meat on 12-hour shifts for example
I think that's the reality of lots of people when they face any redundancy situation - People take up jobs that they wouldn't traditionally want to do in order to survive or look after their family. I don't necessarily see why people on HN would be different.
> I guess the argument would go that a new economic model will be required at that stage.
> ...but at the point where there isn't enough economically useful things for everyone to do, the current system falls down.
Not necessarily. To quote the Bobs from Office Space: "He won't be receiving a paycheck anymore, so it will just work itself out naturally." No need to change, just let the plebs die out.
The paper is suggesting such a new economic model. Do you have a another proposal?
Well, the paper suggests taxing businesses who automate which is more like trying to continue the current economic model by reducing the use of AI.
I suspect if the 'AI-Maximalists' are true, it will need to be a more fundamental rewrite (e.g. UBI, Socialism etc)
Exactly!
As of now, there is no benefit to regular working people. Perhaps in the future, great abundance will occur, but as of now, there will only be job loss, fear, neo-luddism, and blame.
Believe me when I say that I know people, some close to me, that are experiencing fear due to automated systems being installed and tested where they work. They are essentially witnessing start of their automated replacement robot workforce.
Whatever is planned in terms of AI being used to help people needs to happen, sooner rather than later, because all I am seeing is chaos in the horizon.
(⧘⟃⨅⟄⧘)≋≋≋⦻
I am thoughoughly unconvinced that the “AI-based layoffs” are actually caused by AI displacing workers and aren’t just regular layoffs caused by other factors with a smokescreen of “Actually we’re laying people off because we’re doing really well, please don’t dump your stock”.
The article is saying that the solution here isn’t to just throw up our hands and commit suicide as a nation, it’s to simply tax the AI, punishing the negative externality.
Seems like the obvious answer to the prisoner’s dilemma problem where everyone wants to lay off their workforce, but expects that they’ll be the only ones to get this bright idea.
What’s a bit hard for me to rationalize here is why are market shifts considered a negative externality here? We didn’t tax moulding machines because they reduced the demand for sculptors.
Don’t get me wrong, I think the end goal of “Tax those who can pay for it to build a social safety net” is reasonable, I just don’t buy the “negative externalities” argument.
Well because if you don't do something then everyone loses their jobs, and then there's no more consumers, and then the economy implodes, and then everyone dies, including the people utilitizing AI.
It's kind of similar to how nobles fucked themselves in the middle ages. One would think having a lot of serfs is good, but no actually. Having a functioning economy would be better, and in the long run feudal economies stagnated.
And that's why I, today, am effectively much more rich than a feudal lord.
> We didn’t tax moulding machines because they reduced the demand for sculptors.
AI isn't promising to be one machine, it's promising to be a general intelligence that could potentially send unemployment over 50%. We're talking about every truck driver, every warehouse worker, basically everyone who sits in front of a screen for a living being unemployed. This really changes things from 'social safety net' to 'prevent civil unrest that threatens continuity of government.'
> it’s to simply tax the AI, punishing the negative externality.
That "simply" is working overtime here.
Great paper.
> If AI displaces human workers faster than the economy can reabsorb them
Big if.
Consequential enough that it's reasonable to plan for it.
While I agree with the general sentiment that this requires monitoring and study, the abstract is _very_ tendentious, lays multiple hypothesis as facts and doesn’t provide any measurement or alternatives to their preferred solution.
This isn’t a scientific study, it’s a militant manifesto
DApplying a Pigouvian tax to Ai companies might be difficult. None of them actually generate profit, nor have a model to ever do so. I think the huype machien needs to stop spinning before a solution like that would become practical.
Start by shifting taxation from worker incomes to corporate incomes?
What do you mean?
The way I see OP: It's the Luddite Fallacy.
Textile workers worried about machinery saying it's going to result in collapse of the economy when they lose their job.
But every time there's a new technology that threatens some jobs because of an increase in efficiency but investment thus can be placed in other locations creating new jobs. different kinds of jobs.
It's not different this time. Each time the luddite movement is wrong. They are solely concerned with their own selfish concerns and the demands to stop the technological improvement will not be heeded.
Past performance is not a reliable indicator of future results.
Your logic equates to "there will always be jobs for humans to do", which is naive. Remember that we're counting down in the number of things we do better than inorganic stuff. At some point our bodies (admittedly impressive when compared to other animals) will be surpassed in enough aspects that there isn't anything where we can provide enough value to live off.
Thanks for the downvote.
>will be surpassed in enough aspects that there isn't anything where we can provide enough value to live off.
Regardless of that. AI is here to stay. Nobody is going to revoke the constitutions and criminalize AI to prevent this from happening.
What's going to happen will happen.
> Regardless of that. AI is here to stay. Nobody is going to revoke the constitutions and criminalize AI to prevent this from happening.
> What's going to happen will happen.
Username checks out.
I agree, but that was not the point of contention.
I did not downvote you. I argued why your model of what the future will look like is wrong. That point still stands.
So... the solution is basically "pay tax on the demand that you're destroying".
We can all hate on the premise (ai is good enough to do this) and/or the solution presented (centrally enforced taxation), but you gotta admit:
the messaging from SV's AI leaders about how "ai will take all your jobs" is confused as fuck, because if so, who will be on the consuming end of things?
If such tax would introduce an asymmetry that will favour human employment, then there should be enough buying power to create some demand
They don't care because they are set.
> the messaging from SV's AI leaders about how "ai will take all your jobs" is confused as fuck, because if so, who will be on the consuming end of things?
Maybe SV's AI leaders and other assorted trillionaires. A capitalist economy that drops any pretense of serving the needs of anyone except a tiny elite.
Still looking for the AI in the room. Where is it exactly? Surely nobody is claiming LLMs are AI?
AI doesn't need to be AGI to be useful. Surely, you've tried Claude Code before and found it more helpful than Clippy?
I'm not talking about AGI. Microsoft Word is also more helpful than Clippy, is it also AI?
While technically inaccurate, there is so much talk around it that when someone says AI everyone assumes LLM at this point.
Doing the translations and asset creations for CMS content, that used to be done previously by human teams, for example.
Of course it's AI. It's not AGI yet.
If AI displaces human workers faster than the economy can reabsorb them, it risks eroding the very consumer demand firms depend on.
That is a huge "if" though. I am not sure either that the latter falls from this. When the US transitioned away from assembly lines or agriculture dominated, it's not as if consumer spending consequently collapsed.
When did the US transition away from agriculture?
When did the US transition away from assembly lines?
I don’t think you have thought through either one of these and I don’t think they are comparable to what we expect to see for AI’s changes to the job market.
here is the data. Farming as a percentage of the labor force fell considerably https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEi...
Especially also since AI providers are struggling with scaling up.
