5 ms·

It's always possible to make the number go up by divorcing the number from any real-world meaning. I think the "Singularity", as AI proponents call it, will ac

It's always possible to make the number go up by divorcing the number from any real-world meaning. I think the "Singularity", as AI proponents call it, will actually happen, and will likely happen pretty close to its predicted date of 2029. It's just that the form it'll take will be "I'll have your AI talk to my AI", AI botnets clickfrauding AI ad networks, AI newspapers serving up content to AI social media users, AI job boards reading AI generated resumes of people that don't actually exist, AI agents at work generating walls of text that are summarized by AI email readers and then are never read by humans anyway. Meanwhile, actual humans go fuck off to raise their kids while being paid by AI finance and HR departments that have no idea if you're actually working, because AI mouse-jigglers will provide a convincing simulacrum. Of course, the AI-inflated currency won't buy anything that you actually want, and so all actual commerce will function on personal relationships, barter, and local currency, kinda like it did in pre-industrial times. But in terms of real products that affect people's lives in positive ways, capitalism has a problem. There are going to be fewer real people in the future, which means a shrinking market for everything, which makes it harder to justify technological improvements and amortize their R&D costs over a large market. Many of the incentives in capitalism assume growth; when that assumption is violated, the game theory collapses to everybody fighting over a share of the shrinking pie before it disappears.

senordevnyc 11d ago
But in terms of real products that affect people's lives in positive ways, capitalism has a problem. There are going to be fewer real people in the future, which means a shrinking market for everything, which makes it harder to justify technological improvements and amortize their R&D costs over a large market. I thought the US population was predicted to peak in like 2080. Also, hasn’t per-capita consumption in real dollars gone up way more than 10x over the last century? So even if the population in 2126 is slightly lower than in 2026, people’s actual consumption could be much higher. What am I missing?
- nostrademons 11d ago
  The peak has been moved up by decades in the last year. The 2080 figure is from Nov 2023 [1], before Trump administration cuts in immigration. The Jan 2026 CBO report predicted a peak of 2056 [2]; a similar Institute for Family Studies report from July 2026 [3] predicts a peak in 2054. The delta is entirely because of immigration. If you count only natural fertility (births - deaths), population growth nearly went negative during COVID (distorted somewhat by the high death rate), briefly recovered, and is predicted to cross again for good in 2030. The Trump administration's signature campaign promise is to clamp down on immigration, and even get an estimated 12-15M illegal immigrants to self-deport. If this happens, the population will immediately shrink. Immigration and immigrant families are what is propping up the population; if you assume net-zero migration (which IIRC was one of the scenarios modeled in the Census projection), the population already peaked in 2023. [1] https://www.congress.gov/crs_external_products/IN/HTML/IN12327.web.html https://www.congress.gov/crs_external_products/IN/HTML/IN123... [2] https://www.cbo.gov/publication/61879 https://www.cbo.gov/publication/61879 [3] https://ifstudies.org/report-brief/the-demographic-dead-end-2026-state-of-fertility-report https://ifstudies.org/report-brief/the-demographic-dead-end-...
  - senordevnyc 10d ago
    Fair point, although I think it’s pretty hard to predict what immigration will look like over the next few decades. But yes, the peak could come in a couple decades, not a half century, so stipulated. Does that meaningfully change whether capitalism today is facing a shrinking market? Any thoughts on the second part of my comment?
    - nostrademons 10d ago
      The second part is a more complex issue, and also a completely separate one from the question of when/whether/at what number the U.S. population will peak. It is theoretically possible to grow per-capita GDP without growing population, and this would normally be seen as better from a consumer perspective. There are big questions about what that would actually mean, though, and the game theory is all screwed up. Per-capita GDP is, literally speaking, the value of all the economic transactions needed to support a person. In general higher per-capita GDP is seen as better, because if those transactions weren't beneficial, why would you make them? But this is not always the case when you look at society in aggregate and then compare to individual welfare. To see why, consider a small society [1] where everybody owns their own plot of land, farms it, generates enough food to feed their family, and only trades occasionally because most of what they need day-to-day can be produced with their own labor. Such a society would have an extremely low GDP, because most of the work done consists of internal transactions that aren't counted. We pay rent to our landlord, who pays 10% of that to their property manager, who periodically lays out a few thousand to contractors to fix things, all of which is counted in GDP; but somebody who just owns their house outright and fixes it themselves pays zero, and zero gets counted as GDP. We pay a few hundred bucks at the grocery store, who then pays the distributor, who pays the farmers, who distributes wages to their laborers and rents their tractor from John Deere and pays Monsanto for genetically-modified seeds that can't reproduce because you gotta have them coming back for more; but the person who grows or gathers their own food pays nothing, and nothing is counted in GDP. Apparently it didn't take a whole lot of labor either; contemporary hunter-gatherer tribes spend about 12-19 hours/week laboring to obtain food and shelter [2], while I'd bet that most of today's urban poor spend significantly more than that. The normal counterpoint [3] to this is that today's urban poor (let alone wealthy) aren't working just for food and shelter, there's a whole lot of technological development thrown in that just wouldn't have happened without capitalism. So take the family on Medicaid that's working 2 jobs just to survive in a tiny beaten-down rented apartment. They also have clean water, and access to vaccines, and antibiotics when their kid gets sick, and a car that lets them go to the beach on weekends, and they can store food for the winter in a refrigerator, and they sit on a toilet to poop instead of burying it in a hole, and many other things we take for granted. They might even have a TV and a cell-phone and access to millions of hours of entertainment, since those have gotten so cheap relative to rent. It's just not directly comparable to a hunter-gatherer lifestyle. But the counter to the counterpoint is that most of those technological innovations have depended upon an increasing population together with capitalism. The actual mechanism here is that capitalism commoditizes goods where there is an oversupply, making it unprofitable to continue producing them. That means that young people entering the workforce for the first time are strongly disincentivized from being say a farmer or a paperboy or a domestic servant, and instead are incentivized to move up-market to innovation industries like software engineering or biotech. Capitalism is the ratchet that keeps people inventing new things, but the inventing (and even the maintenance of these new technologies) is done by actual people. If there are not enough people, than folks who would otherwise go into technological industries will have to work instead at the basic tasks of keeping society running, like how Matthew McConaughey's character in Interstellar [4] was a trained NASA pilot who is growing food on a farm because that is what society needs. This has actually happened in several occasions in history, eg. during the fall of the Roman Empire we saw increasing de-specialization as the urban elite class fled Rome and ended up working the fields during the Dark Ages, while during Crisis of the Late Middle Ages, you had the younger sons of the nobility returning to productive work as guild artisans (and ushering in the Rennaissance) as the peasantry was largely killed off by Black Death and the Hundred Years War and the Wars of the Roses. The second part of my response is about game theory, and specifically about the actual decision to invest that financiers and entrepreneurs make when they hear a new idea. When the market is rapidly expanding, this is usually a no-brainer: capture it or somebody else will. But when the market is contracting, you usually do not see further investment. Going back to the meaning of per-capita GDP expanding without population growing, it implies that some new and dramatically better way of doing something will come out, compete with the existing alternatives, and convince consumers to spend significantly more. This is a hard battle! Typically tech startups win when there is no alternative at all for the desired activity, or at least when they are orders of magnitude better than the alternative. A few percentage points better doesn't cut it; its not noticeable enough for consumers to switch. And investors know this, which is why it's essentially impossible to get a startup funded by pitching "I'm going to be like Comcast but cheaper." Everybody knows that ISPs in America are overpriced, but a price war means a contracting market instead of an expanding one, and battles over contracting markets are vicious. Much better to fund "I'm going to make an AI that runs businesses automatically", which is probably bullshit but it's bullshit that has very little competition and a huge total addressable market. The incentive in a contracting market isn't to invest heavily to prove yourself better than the competition. It's to milk whatever assets you have for as much cash as possible before they wink out of existence, and then take the money and run. A lot of the prosocial behavior that capitalism has resulted in was heavily conditioned on expanding markets, expanding population, and expanding payoffs for innovation. When population collapses, those incentives reverse. [1] like the Parable of the Mexican Fisherman: https://aliabdaal.com/newsletter/the-parable-of-the-mexican-fisherman/ https://aliabdaal.com/newsletter/the-parable-of-the-mexican-... [2] https://web.cs.ucdavis.edu/~rogaway/classes/188/materials/diamond https://web.cs.ucdavis.edu/~rogaway/classes/188/materials/di... [3] https://francismead.com/2015/06/06/agriculture-humanitys-worst-mistake-or-not/ https://francismead.com/2015/06/06/agriculture-humanitys-wor... [4] https://interstellarfilm.fandom.com/wiki/Joseph_Cooper https://interstellarfilm.fandom.com/wiki/Joseph_Cooper
overfeed 11d ago
> Meanwhile, actual humans go fuck off to raise their kids while being paid by AI finance and HR departments I was able to suspend my disbelief until this line. Unless Capitalism has been upended by 2029, AI companies will continue to have capitalist goals like maximizing value for their shareholders. AI HR won't pay humans that don't positively move the needle on the P&L
- nostrademons 11d ago
  It would be interesting if AI actually did optimize the P&L at most companies. Right now, the vast majority of office jobs are bullshit jobs, because the director who holds the budget is incentivized to maximize headcount rather than P&L. The job is not strictly necessary, but if the money is not spent, it disappears, and so the director has an incentive to ensure everything is spent. The reason these jobs can continue to exist is because of poor information flow within the organization. Directors distort both the importance and required resources for the projects they're working on, making them seem harder and more lucrative than they actually are. It's notoriously difficult to credit an actual transaction back to a particular product feature that made someone buy, so there is no actual verification of these. Then these project proposals are looked at by an FP&A analyst who has maybe 5 minutes to evaluate each, so of course 100 headcount to maintain a 3-screen CRUD mobile app seems reasonable. A rational, omniscient AI would look at this and say "This is ridiculous. I can code this app in 50ms." And so if we actually did put AI in charge of the P&L, they would likely fire 100% of the employees, as well as doing other shady stuff like charging all stored credit cards multiple times, gaslighting consumers as to the existence of the product, and DDoSing the court system. But again, this strengthens my thesis. So the AI has now fired 100% of the humans in the workforce and is running amuck with the financial system. What do the remaining 100% of unemployed people do? Start trading their skills for old-fashioned analog currency, like gold coins or even paper dollars. They just disconnect the AI, turn off the computers, and go about their daily business.
  - overfeed 11d ago
    > Right now, the vast majority of office jobs are bullshit jobs, because the director who holds the budget is incentivized to maximize headcount This may have been briefly true during the Zero-Interest regime. The economy isn't doing so hot right now, outside of AI. If squeezed companies doing multiple rounds of layoffs keep those "bullshit jobs", maybe they are not bullshit. Perhaps coordinating people, and keeping up with bureaucratic demands like paying taxes, not breaking the law and growing the business requires warm bodies. > What do the remaining 100% of unemployed people do? Start trading their skills for old-fashioned analog currency. History books suggest a different outcome. Firing 100% of employees in a growth-based, consumer economy only works when you're the only one using that strategy. If a critical mass of companies do the same, then you're staring at a managed decline at best.

senordevnyc 11d ago

But in terms of real products that affect people's lives in positive ways, capitalism has a problem. There are going to be fewer real people in the future, which means a shrinking market for everything, which makes it harder to justify technological improvements and amortize their R&D costs over a large market. I thought the US population was predicted to peak in like 2080. Also, hasn’t per-capita consumption in real dollars gone up way more than 10x over the last century? So even if the population in 2126 is slightly lower than in 2026, people’s actual consumption could be much higher. What am I missing?

nostrademons 11d ago
The peak has been moved up by decades in the last year. The 2080 figure is from Nov 2023 [1], before Trump administration cuts in immigration. The Jan 2026 CBO report predicted a peak of 2056 [2]; a similar Institute for Family Studies report from July 2026 [3] predicts a peak in 2054. The delta is entirely because of immigration. If you count only natural fertility (births - deaths), population growth nearly went negative during COVID (distorted somewhat by the high death rate), briefly recovered, and is predicted to cross again for good in 2030. The Trump administration's signature campaign promise is to clamp down on immigration, and even get an estimated 12-15M illegal immigrants to self-deport. If this happens, the population will immediately shrink. Immigration and immigrant families are what is propping up the population; if you assume net-zero migration (which IIRC was one of the scenarios modeled in the Census projection), the population already peaked in 2023. [1] https://www.congress.gov/crs_external_products/IN/HTML/IN12327.web.html https://www.congress.gov/crs_external_products/IN/HTML/IN123... [2] https://www.cbo.gov/publication/61879 https://www.cbo.gov/publication/61879 [3] https://ifstudies.org/report-brief/the-demographic-dead-end-2026-state-of-fertility-report https://ifstudies.org/report-brief/the-demographic-dead-end-...
- senordevnyc 10d ago
  Fair point, although I think it’s pretty hard to predict what immigration will look like over the next few decades. But yes, the peak could come in a couple decades, not a half century, so stipulated. Does that meaningfully change whether capitalism today is facing a shrinking market? Any thoughts on the second part of my comment?
  - nostrademons 10d ago
    The second part is a more complex issue, and also a completely separate one from the question of when/whether/at what number the U.S. population will peak. It is theoretically possible to grow per-capita GDP without growing population, and this would normally be seen as better from a consumer perspective. There are big questions about what that would actually mean, though, and the game theory is all screwed up. Per-capita GDP is, literally speaking, the value of all the economic transactions needed to support a person. In general higher per-capita GDP is seen as better, because if those transactions weren't beneficial, why would you make them? But this is not always the case when you look at society in aggregate and then compare to individual welfare. To see why, consider a small society [1] where everybody owns their own plot of land, farms it, generates enough food to feed their family, and only trades occasionally because most of what they need day-to-day can be produced with their own labor. Such a society would have an extremely low GDP, because most of the work done consists of internal transactions that aren't counted. We pay rent to our landlord, who pays 10% of that to their property manager, who periodically lays out a few thousand to contractors to fix things, all of which is counted in GDP; but somebody who just owns their house outright and fixes it themselves pays zero, and zero gets counted as GDP. We pay a few hundred bucks at the grocery store, who then pays the distributor, who pays the farmers, who distributes wages to their laborers and rents their tractor from John Deere and pays Monsanto for genetically-modified seeds that can't reproduce because you gotta have them coming back for more; but the person who grows or gathers their own food pays nothing, and nothing is counted in GDP. Apparently it didn't take a whole lot of labor either; contemporary hunter-gatherer tribes spend about 12-19 hours/week laboring to obtain food and shelter [2], while I'd bet that most of today's urban poor spend significantly more than that. The normal counterpoint [3] to this is that today's urban poor (let alone wealthy) aren't working just for food and shelter, there's a whole lot of technological development thrown in that just wouldn't have happened without capitalism. So take the family on Medicaid that's working 2 jobs just to survive in a tiny beaten-down rented apartment. They also have clean water, and access to vaccines, and antibiotics when their kid gets sick, and a car that lets them go to the beach on weekends, and they can store food for the winter in a refrigerator, and they sit on a toilet to poop instead of burying it in a hole, and many other things we take for granted. They might even have a TV and a cell-phone and access to millions of hours of entertainment, since those have gotten so cheap relative to rent. It's just not directly comparable to a hunter-gatherer lifestyle. But the counter to the counterpoint is that most of those technological innovations have depended upon an increasing population together with capitalism. The actual mechanism here is that capitalism commoditizes goods where there is an oversupply, making it unprofitable to continue producing them. That means that young people entering the workforce for the first time are strongly disincentivized from being say a farmer or a paperboy or a domestic servant, and instead are incentivized to move up-market to innovation industries like software engineering or biotech. Capitalism is the ratchet that keeps people inventing new things, but the inventing (and even the maintenance of these new technologies) is done by actual people. If there are not enough people, than folks who would otherwise go into technological industries will have to work instead at the basic tasks of keeping society running, like how Matthew McConaughey's character in Interstellar [4] was a trained NASA pilot who is growing food on a farm because that is what society needs. This has actually happened in several occasions in history, eg. during the fall of the Roman Empire we saw increasing de-specialization as the urban elite class fled Rome and ended up working the fields during the Dark Ages, while during Crisis of the Late Middle Ages, you had the younger sons of the nobility returning to productive work as guild artisans (and ushering in the Rennaissance) as the peasantry was largely killed off by Black Death and the Hundred Years War and the Wars of the Roses. The second part of my response is about game theory, and specifically about the actual decision to invest that financiers and entrepreneurs make when they hear a new idea. When the market is rapidly expanding, this is usually a no-brainer: capture it or somebody else will. But when the market is contracting, you usually do not see further investment. Going back to the meaning of per-capita GDP expanding without population growing, it implies that some new and dramatically better way of doing something will come out, compete with the existing alternatives, and convince consumers to spend significantly more. This is a hard battle! Typically tech startups win when there is no alternative at all for the desired activity, or at least when they are orders of magnitude better than the alternative. A few percentage points better doesn't cut it; its not noticeable enough for consumers to switch. And investors know this, which is why it's essentially impossible to get a startup funded by pitching "I'm going to be like Comcast but cheaper." Everybody knows that ISPs in America are overpriced, but a price war means a contracting market instead of an expanding one, and battles over contracting markets are vicious. Much better to fund "I'm going to make an AI that runs businesses automatically", which is probably bullshit but it's bullshit that has very little competition and a huge total addressable market. The incentive in a contracting market isn't to invest heavily to prove yourself better than the competition. It's to milk whatever assets you have for as much cash as possible before they wink out of existence, and then take the money and run. A lot of the prosocial behavior that capitalism has resulted in was heavily conditioned on expanding markets, expanding population, and expanding payoffs for innovation. When population collapses, those incentives reverse. [1] like the Parable of the Mexican Fisherman: https://aliabdaal.com/newsletter/the-parable-of-the-mexican-fisherman/ https://aliabdaal.com/newsletter/the-parable-of-the-mexican-... [2] https://web.cs.ucdavis.edu/~rogaway/classes/188/materials/diamond https://web.cs.ucdavis.edu/~rogaway/classes/188/materials/di... [3] https://francismead.com/2015/06/06/agriculture-humanitys-worst-mistake-or-not/ https://francismead.com/2015/06/06/agriculture-humanitys-wor... [4] https://interstellarfilm.fandom.com/wiki/Joseph_Cooper https://interstellarfilm.fandom.com/wiki/Joseph_Cooper

overfeed 11d ago

> Meanwhile, actual humans go fuck off to raise their kids while being paid by AI finance and HR departments I was able to suspend my disbelief until this line. Unless Capitalism has been upended by 2029, AI companies will continue to have capitalist goals like maximizing value for their shareholders. AI HR won't pay humans that don't positively move the needle on the P&L

nostrademons 11d ago
It would be interesting if AI actually did optimize the P&L at most companies. Right now, the vast majority of office jobs are bullshit jobs, because the director who holds the budget is incentivized to maximize headcount rather than P&L. The job is not strictly necessary, but if the money is not spent, it disappears, and so the director has an incentive to ensure everything is spent. The reason these jobs can continue to exist is because of poor information flow within the organization. Directors distort both the importance and required resources for the projects they're working on, making them seem harder and more lucrative than they actually are. It's notoriously difficult to credit an actual transaction back to a particular product feature that made someone buy, so there is no actual verification of these. Then these project proposals are looked at by an FP&A analyst who has maybe 5 minutes to evaluate each, so of course 100 headcount to maintain a 3-screen CRUD mobile app seems reasonable. A rational, omniscient AI would look at this and say "This is ridiculous. I can code this app in 50ms." And so if we actually did put AI in charge of the P&L, they would likely fire 100% of the employees, as well as doing other shady stuff like charging all stored credit cards multiple times, gaslighting consumers as to the existence of the product, and DDoSing the court system. But again, this strengthens my thesis. So the AI has now fired 100% of the humans in the workforce and is running amuck with the financial system. What do the remaining 100% of unemployed people do? Start trading their skills for old-fashioned analog currency, like gold coins or even paper dollars. They just disconnect the AI, turn off the computers, and go about their daily business.
- overfeed 11d ago
  > Right now, the vast majority of office jobs are bullshit jobs, because the director who holds the budget is incentivized to maximize headcount This may have been briefly true during the Zero-Interest regime. The economy isn't doing so hot right now, outside of AI. If squeezed companies doing multiple rounds of layoffs keep those "bullshit jobs", maybe they are not bullshit. Perhaps coordinating people, and keeping up with bureaucratic demands like paying taxes, not breaking the law and growing the business requires warm bodies. > What do the remaining 100% of unemployed people do? Start trading their skills for old-fashioned analog currency. History books suggest a different outcome. Firing 100% of employees in a growth-based, consumer economy only works when you're the only one using that strategy. If a critical mass of companies do the same, then you're staring at a managed decline at best.