Wonderfully clear and well-reasoned. What did Marx say about pessimism arising from fully understanding his argument and its insights? Is there a reasonable solution to the growing crisis of AI? Can society be organized to share the wealth? Probably not. So what is the next best solution from a utilitarian perspective?
This is tangential to your point of course, but I’m deeply skeptical about the basic “We’re laying people off now thanks to AI” narrative. That could very quickly become the same genre of convenient excuse that “oh, it was the shoplifting” was for retail chains to shut down money-losing stores after they overexpanded.
I know there is real efficiency being unlocked. I also know that companies like to fire the labor they resent having to pay the second they have a bad quarter. It would be more than convenient to tap into the Zeitgeist like this and chalk it up to the miraculous productivity gains of AI.
There's a story by Polish sci-fi writer Stanislaw Lem from 1971 called "Twenty-fourth Voyage" in which a traveler (Tichy) arrives on a planet covered with vast fields of glass-like disks. He learns the planet had a consumer based economy, but when production was automated, workers were displaced, vast surpluses of unused goods accrued, and the economy began to collapse as the consumer base shrank. The leaders of the planet (which includes a senate-like body called the Plenum Moronicum) ask their Claude-like AI (the "Voluntary Universalizer of Absolute Order") to solve the problems of their economy. First it creates an army of robots who purchase the surplus goods. Then it creates an opulent "Rainbow Palace" which seems to be a gaudy, sublime tourist attraction, containing untold wonders, but it is actually a front for a machine that turns most of the non-robot residents of the planet into disks, thus solving their economic crisis.
While all of the above is worth working on, probably the one thing that is politically feasible almost immediately, and which would go a long way towards solving the problem, is the Pigouvian tax which would give corporations a reason to restrain their use of A. The best proposal I've seen is a simple tax on tokens (the units of use of computation now standard for pricing). See Peter Coy's column in the NYTimes https://www.nytimes.com/2026/06/13/business/dealbook/ai-tax-proposals.html
I see a similar situation unfolding, as in the efforts to switch from coal to greener energy sources. No leadership, either here or in England, gave real thought to the workers and their families being displaced. Of course, at this time America doesn't really have any leadership unless you are a member of the oligarchy, and they have zero concerns for the average person. This would be a great time for political leaders to step forward and talk about how to make economic adjustments to assist those most likely to be displaced by AI. Maybe an increase in the unemployment tax rate for everyone displaced by AI. This would help to fund unemployment insurance. They could create re-education programs that people can access before they are laid off and need money. The creation of a government/business partnership to address the concerns of the public about AI and the needs of those workers who will see their jobs disappear by automation. None of these suggestions are easy to implement, but the consequences of ignoring this looming problem are very severe.
A data center isn't an ideal widget, it is full of machines whose raw materials are gained with blood, violence, and destruction of environments. That alone puts myself in the "smash the machine" camp, especially as the scarcities compound and imperialists accept more and more bloody and destructive tactics to acquire scarce minerals.
If we're trying to find a path between "smash the machines" and "let it rip," the most important questions aren't about how and where we apply AI technology, but what forms of AI we're incorporating into the circuits of production. The "how" and "where" will never stay put; it's just the nature of AI to move past fixed domains of application (even if notions of "general" AI are mostly techno-optimist tripe.) The real consequences of AI come down to the tech itself. The most salient example right now would be generative models, since they're very different from what was called AI just a few years earlier. Their political effects are much more insidious and difficult to detect, let alone regulate, especially if we assume that these effects will be measurable and easy to articulate without above-average familiarity with the technology.
This is all to say that critiques of AI that take a traditional Marxist approach need to be supplemented with more tech-savvy analyses. (Not at all a criticism of this post)
Wonderfully clear and well-reasoned. What did Marx say about pessimism arising from fully understanding his argument and its insights? Is there a reasonable solution to the growing crisis of AI? Can society be organized to share the wealth? Probably not. So what is the next best solution from a utilitarian perspective?
This is tangential to your point of course, but I’m deeply skeptical about the basic “We’re laying people off now thanks to AI” narrative. That could very quickly become the same genre of convenient excuse that “oh, it was the shoplifting” was for retail chains to shut down money-losing stores after they overexpanded.
I know there is real efficiency being unlocked. I also know that companies like to fire the labor they resent having to pay the second they have a bad quarter. It would be more than convenient to tap into the Zeitgeist like this and chalk it up to the miraculous productivity gains of AI.
There's a story by Polish sci-fi writer Stanislaw Lem from 1971 called "Twenty-fourth Voyage" in which a traveler (Tichy) arrives on a planet covered with vast fields of glass-like disks. He learns the planet had a consumer based economy, but when production was automated, workers were displaced, vast surpluses of unused goods accrued, and the economy began to collapse as the consumer base shrank. The leaders of the planet (which includes a senate-like body called the Plenum Moronicum) ask their Claude-like AI (the "Voluntary Universalizer of Absolute Order") to solve the problems of their economy. First it creates an army of robots who purchase the surplus goods. Then it creates an opulent "Rainbow Palace" which seems to be a gaudy, sublime tourist attraction, containing untold wonders, but it is actually a front for a machine that turns most of the non-robot residents of the planet into disks, thus solving their economic crisis.
While all of the above is worth working on, probably the one thing that is politically feasible almost immediately, and which would go a long way towards solving the problem, is the Pigouvian tax which would give corporations a reason to restrain their use of A. The best proposal I've seen is a simple tax on tokens (the units of use of computation now standard for pricing). See Peter Coy's column in the NYTimes https://www.nytimes.com/2026/06/13/business/dealbook/ai-tax-proposals.html
I see a similar situation unfolding, as in the efforts to switch from coal to greener energy sources. No leadership, either here or in England, gave real thought to the workers and their families being displaced. Of course, at this time America doesn't really have any leadership unless you are a member of the oligarchy, and they have zero concerns for the average person. This would be a great time for political leaders to step forward and talk about how to make economic adjustments to assist those most likely to be displaced by AI. Maybe an increase in the unemployment tax rate for everyone displaced by AI. This would help to fund unemployment insurance. They could create re-education programs that people can access before they are laid off and need money. The creation of a government/business partnership to address the concerns of the public about AI and the needs of those workers who will see their jobs disappear by automation. None of these suggestions are easy to implement, but the consequences of ignoring this looming problem are very severe.
A data center isn't an ideal widget, it is full of machines whose raw materials are gained with blood, violence, and destruction of environments. That alone puts myself in the "smash the machine" camp, especially as the scarcities compound and imperialists accept more and more bloody and destructive tactics to acquire scarce minerals.
If we're trying to find a path between "smash the machines" and "let it rip," the most important questions aren't about how and where we apply AI technology, but what forms of AI we're incorporating into the circuits of production. The "how" and "where" will never stay put; it's just the nature of AI to move past fixed domains of application (even if notions of "general" AI are mostly techno-optimist tripe.) The real consequences of AI come down to the tech itself. The most salient example right now would be generative models, since they're very different from what was called AI just a few years earlier. Their political effects are much more insidious and difficult to detect, let alone regulate, especially if we assume that these effects will be measurable and easy to articulate without above-average familiarity with the technology.
This is all to say that critiques of AI that take a traditional Marxist approach need to be supplemented with more tech-savvy analyses. (Not at all a criticism of this post)