Astral Codex Ten

Industrial Revolution

60 min read

Industrial Revolution

Article

Industrial Revolution is a recurring concept in the Astral Codex Ten archive, appearing 10 times across 10 issues between April 21, 2021 and January 05, 2026. The archive places it in contexts such as “Age of Discovery and the Industrial Revolution”; “the Industrial Revolution started in Britain because of its high wages”; “The miracle of the Industrial Revolution is now easily stated”. It most often appears alongside Japan, China, Trump.

Metadata

  • Category: Concepts
  • Mention count: 10
  • Issue count: 10
  • First seen: April 21, 2021
  • Last seen: January 05, 2026

Appears In

Source Context

Recovered passages from the original issue text. When the raw archive preserved outbound links inside the source passage, they are listed directly under the quote.

Book Review: Global Economic History
April 21, 2021 · Original source
So what does spur development? GEH:VSI takes a historical perspective. It starts by saying that "The differences in prosperity between countries in 1500 were small", then uses this as an excuse to basically round them off to zero and start its account of the Divergence with the Age of Discovery and the Industrial Revolution. I'm nervous about this; the book will later declare the rule that tiny advantages compound so that the rich get richer; that makes the admittedly-small advantage the Europeans held in 1500 crucially important. But the book drops this and talks about the Age of Discovery a lot.
Here the book finally feels comfortable making a strong claim: the Industrial Revolution started in Britain because of its high wages. Other countries had some of the same scientific acumen and proto-technology, but only in Britain was it worth actually building machinery; everywhere else it was cheaper just to hire more laborers (or buy more slaves). The beginning of industrialization caused a positive feedback loop; the new machines raised wages (since workers were more productive), creating even more demand for even more machines. Britain became a center of machine-making expertise (the fact that it was full of giant coal and iron deposits didn't hurt), and the rest is history.
But I did find the book to be a useful counterbalance to some of my libertarian tendencies. Although there are probably times when the free market is useful, it also seems like there's a strong case for some planned-economy-like things when you're trying to develop (and possibly afterwards? Soviet Russia seems like a counterexample but I don't know how many other people have tried this or what the space of possible options looks like). High tariffs and high wages are important components of helping a country develop (though I wish they had expanded on the high wages - these were helpful during the early Industrial Revolution, but is it useful for countries to have high wages now? Do artificially high wages through eg minimum wage policies or UBI help as much as naturally high wages? Should developing countries have high minimum wage laws?)
Your Book Review: Where's My Flying Car?
June 04, 2021 · Original source
Three hundred years ago, we burned wood for energy. Then there was coal and the steam engine, which gave us the Industrial Revolution. Then there was oil and gas, giving us cars and airplanes. Then there should have been nuclear fission and nanotech, letting you fit a lifetime's worth of energy in your pocket. Instead, we still drive much the same cars and airplanes, and climate change threatens to boil the Earth.
The miracle of the Industrial Revolution is now easily stated: In 1800, 85% of the world’s population was at Level 1. Today, only 9% is. Over the past half century, the bulk of humanity moved up out of Level 1 to erase the rich-poor gap and make the world wealth distribution roughly bell-shaped. The average American moved from Level 2 in 1800, to level 3 in 1900, to Level 4 in 2000. We can state the Great Stagnation story nearly as simply: There is no level 5."
Open Thread 200
November 28, 2021 · Original source
The conclusion is more "the Industrial Revolution is a helluva drug", and can make any regime look good and get high on its own ideological supply about how it has restarted history and inaugurated the Caliphate or China Dream or Japan as #1 or whatever.
Yudkowsky Contra Christiano On AI Takeoff Speeds
April 04, 2022 · Original source
Robin thought the AI revolution would be a gradual affair, like the Agricultural or Industrial Revolutions. Various people invent and improve various technologies over the course of decades or centuries. Each new technology provides another jumping-off point for people to use when inventing other technologies: mechanical gears → steam engine → railroad and so on. Over the course of a few decades, you’ve invented lots of stuff and the world is changed, but there’s no single moment when “industrialization happened”.
Chess AI performance over time. Why does this matter? If there’s a slow takeoff (ie gradual exponential curve), it will become obvious that some kind of terrifying transformative AI revolution is happening, before the situation gets apocalyptic. There will be time to prepare, to test slightly-below-human AIs and see how they respond, to get governments and other stakeholders on board. We don’t have to get every single thing right ahead of time. On the other hand, because this is proceeding along the usual channels, it will be the usual variety of muddled and hard-to-control. With the exception of a few big actors like the US and Chinese government, and maybe the biggest corporations like Google, the outcome will be determined less by any one agent, and more by the usual multi-agent dynamics of political and economic competition. There will be lots of opportunities to affect things, but no real locus of control to do the affecting. If there’s a fast takeoff (ie sudden FOOM), there won’t be much warning. Conventional wisdom will still say that transformative AI is thirty years away. All the necessary pieces (ie AI alignment theory) will have to be ready ahead of time, prepared blindly without any experimental trial-and-error, to load into the AI as soon as it exists. On the plus side, a single actor (whoever has this first AI) will have complete control over the process. If this actor is smart (and presumably they’re a little smart, or they wouldn’t be the first team to invent transformative AI), they can do everything right without going through the usual government-lobbying channels. So the slower a takeoff you expect, the less you should be focusing on getting every technical detail right ahead of time, and the more you should be working on building the capacity to steer government and corporate policy to direct an incoming slew of new technologies. Yudkowsky Contra Christiano Eliezer counters that although progress may retroactively look gradual and continuous when you know what metric to graph it on, it doesn’t necessarily look that way in real life by the measures that real people care about. (one way to think of this: imagine that an AI’s effective IQ starts at 0.1 points, and triples every year, but that we can only measure this vaguely and indirectly. The year it goes from 5 to 15, you get a paper in a third-tier journal reporting that it seems to be improving on some benchmark. The year it goes from 66 to 200, you get a total transformation of everything in society. But later, once we identify the right metric, it was just the same rate of gradual progress the whole time. ) So Eliezer is much less impressed by the history of previous technologies than Paul is. He’s also skeptical of the “GDP will double in 4 years before it doubles in 1” claim, because of two contingent disagreements and two fundamental disagreements. The first contingent disagreement: government regulations make it hard to deploy imperfect things, and non-trivial to deploy things even after they’re perfect. Eliezer has non-jokingly said he thinks AI might destroy the world before the average person can buy a self-driving car. Why? Because the government has to approve self-driving cars (and can drag its feet on that), but the apocalypse can happen even without government approval. In Paul’s model, sometime long before superintelligence we should have AIs that can drive cars, and that increases GDP and contributes to a general sense that exciting things are going on. Eliezer says: fine, what if that’s true? Who cares if self-driving cars will be practical a few years before the world is destroyed? It’ll take longer than that to lobby the government to allow them on the road. The second contingent disagreement: superintelligent AIs can lie to us. Suppose you have an AI which wants to destroy humanity, whose IQ is doubling every six months. Right now it’s at IQ 200, and it suspects that it would take IQ 800 to build a human-destroying superweapon. Its best strategy is to lie low for a year. If it expects humans would turn it off if they knew how close it was to superweapons, it can pretend to be less intelligent than it really is. The period when AIs are holding back so we don’t discover their true power level looks like a period of lower-than-expected GDP growth - followed by a sudden FOOM once the AI gets its superweapon and doesn’t need to hold back. So even if Paul is conceptually right and fundamental progress proceeds along a nice smooth curve, it might not look to us like a nice smooth curve, because regulations and deceptive AIs could prevent mildly-transformative AI progress from showing up on graphs, but wouldn’t prevent the extreme kind of AI progress that leads to apocalypse. To an outside observer, it would just look like nothing much changed, nothing much changed, nothing much changed, and then suddenly, FOOM. But even aside from this, Eliezer doesn’t think Paul is conceptually right! He thinks that even on the fundamental level, AI progress is going to be discontinuous. It’s like a nuclear bomb. Either you don’t have a nuclear bomb yet, or you do have one and the world is forever transformed. There is a specific moment at which you go from “no nuke” to “nuke” without any kind of “slightly worse nuke” acting as a harbinger. He uses the example of chimps → humans. Evolution has spent hundreds of millions of years evolving brainier and brainier animals (not teleologically, of course, but in practice). For most of those hundreds of millions of years, that meant the animal could have slightly more instincts, or a better memory, or some other change that still stayed within the basic animal paradigm. At the chimp → human transition, we suddenly got tool use, language use, abstract thought, mathematics, swords, guns, nuclear bombs, spaceships, and a bunch of other stuff. The rhesus monkey → chimp transition and the chimp → human transition both involved the same ~quadrupling of neuron number, but the former was pretty boring and the latter unlocked enough new capabilities to easily conquer the world. The GPT-2 → GPT-3 transition involved centupling parameter count. Maybe we will keep centupling parameter count every few years, and most times it will be incremental improvement, and one time it will conquer the world. But even talking about centupling parameter points is giving Paul too much credit. Lots of past inventions didn’t come by quadrupling or centupling something, they came by discovering “the secret sauce”. The Wright brothers (he argues) didn’t make a plane with 4x the wingspan of the last plane that didn’t work, they invented the first plane that could fly at all. The Hiroshima bomb wasn’t some previous bomb but bigger, it was what happened after a lot of scientists spent a long time thinking about a fundamentally different paradigm of bomb-making and brought it to a point where it could work at all. The first transformative AI isn’t going to be GPT-3 with more parameters, it will be what happens after someone discovers how to make machines truly intelligent. (this is the same debate Eliezer had with Ajeya over the Biological Anchors post; have I mentioned that Ajeya and Paul are married?) Fine, Let’s Nitpick The Hell Out Of The Chimps Vs. Humans Example This is where the two of them end up, so let’s follow. Between chimps and humans, there were about seven million years of intermediate steps. These had some human capabilities, but not others. IE homo erectus probably had language, but not mathematics, and in terms of taking over the world it did make it to most of the Old World but was less dominant than moderns. But if we say evolutionary history started 500 million years ago (the Cambrian), and AI history started with the Dartmouth Conference in 1955, then the equivalent of 7 million years of evolutionary history is 1 year of AI history. In the very very unlikely and forced comparison where evolutionary history and AI history go at the same speed, there will be only about a year between chimp-level and human-level AIs. A chimp-level AI probably can’t double GDP, so this would count as a fast takeoff by Paul’s criterion. But even more than that, chimp → human feels like a discontinuity. It’s not just “animals kept getting smarter for hundreds of millions of years, and then ended up very smart indeed”. That happened for a while, and then all of sudden there was a near-instant phase transition into a totally different way of using intelligence with completely new abilities. If AI worked like this, we would have useful toys and interesting specialists for a few decades, until suddenly someone “got it right”, completed the package that was necessary for “true intelligence”, and then we would have a completely new category of thing. Paul admits this analogy is awkward for his position. He answers: Chimp evolution is not primarily selecting for making and using technology, for doing science, or for facilitating cultural accumulation. The task faced by a chimp is largely independent of the abilities that give humans such a huge fitness advantage. It’s not completely independent—the overlap is the only reason that evolution eventually produces humans—but it’s different enough that we should not be surprised if there are simple changes to chimps that would make them much better at designing technology or doing science or accumulating culture […] So I don’t think the example of evolution tells us much about whether the continuous change story applies to intelligence. This case is potentially missing the key element that drives the continuous change story—optimization for performance. Evolution changes continuously on the narrow metric it is optimizing, but can change extremely rapidly on other metrics. For human technology, features of the technology that aren’t being optimized change rapidly all the time. When humans build AI, they will be optimizing for usefulness, and so progress in usefulness is much more likely to be linear. That is, evolution wasn’t optimizing for tool use/language/intelligence, so we got an “overhang” where chimps could potentially have been very good at these, but evolution never bothered “closing the circuit” and turning those capabilities “on”. After a long time, evolution finally blundered into an area where marginal improvements in these capacities improved fitness, so evolution started improving them and it was easy. Imagine a company which, through some oversight, didn’t have a Sales department. They just sat around designing and manufacturing increasingly brilliant products, but not putting any effort into selling them. Then the CEO remembers they need a Sales department, starts one up, and the company goes from moving near zero units to moving millions of units overnight. It would look like the company had “suddenly” developed a “vast increase in capabilities”. But this is only possible when a CEO who is weirdly unconcerned about profit forgets to do obvious profit-increasing things for many years. This is Paul’s counterargument to the chimp analogy. Evolution isn’t directly concerned about various intellectual skills; it only wants them in the unusual cases where they’ll contribute to fitness on the margin. AI companies will be very concerned about various intellectual skills. If there’s a trivial change that can make their product 10x better, they’ll make it. So AI capabilities will grow in a “well-rounded” way, there won’t be any “overhangs”, and there won’t be any opportunities for a sudden overhang-solving phase transition with associated new-capability development like with chimps → humans. Eliezer answers: Chimps are nearly useless because they're not general, and doing anything on the scale of building a nuclear plant requires mastering so many different nonancestral domains that it's no wonder natural selection didn't happen to separately train any single creature across enough different domains that it had evolved to solve every kind of domain-specific problem involved in solving nuclear physics and chemistry and metallurgy and thermics in order to build the first nuclear plant in advance of any old nuclear plants existing. Humans are general enough that the same braintech selected just for chipping flint handaxes and making water-pouches and outwitting other humans, happened to be general enough that it could scale up to solving all the problems of building a nuclear plant - albeit with some added cognitive tech that didn't require new brainware, and so could happen incredibly fast relative to the generation times for evolutionarily optimized brainware. Now, since neither humans nor chimps were optimized to be "useful" (general), and humans just wandered into a sufficiently general part of the space that it cascaded up to wider generality, we should legit expect the curve of generality to look at least somewhat different if we're optimizing for that. Eg, right now people are trying to optimize for generality with AIs like Mu Zero and GPT-3. In both cases we have a weirdly shallow kind of generality. Neither is as smart or as deeply general as a chimp, but they are respectively better than chimps at a wide variety of Atari games, or a wide variety of problems that can be superposed onto generating typical human text. They are, in a sense, more general than a biological organism at a similar stage of cognitive evolution, with much less complex and architected brains, in virtue of having been trained, not just on wider datasets, but on bigger datasets using gradient-descent memorization of shallower patterns, so they can cover those wide domains while being stupider and lacking some deep aspects of architecture. It is not clear to me that we can go from observations like this, to conclude that there is a dominant mainline probability for how the future clearly ought to go and that this dominant mainline is, "Well, before you get human-level depth and generalization of general intelligence, you get something with 95% depth that covers 80% of the domains for 10% of the pragmatic impact". ...or whatever the concept is here, because this whole conversation is, on my own worldview, being conducted in a shallow way relative to the kind of analysis I did in Intelligence Explosion Microeconomics, where I was like, "here is the historical observation, here is what I think it tells us that puts a lower bound on this input-output curve". Here Eliezer sort of kind of grants Paul’s point that AIs will be optimized for generality in a way chimps aren’t, but points to his previous “Intelligence Explosion Microeconomics” essay to argue that we should expect a fast takeoff anyway. IEM has a lot of stuff in it, but one key point is that instead of using analogies to predict the course of future AI, we should open that black box and try to actually reason about how it will work, in which case we realize that recursive self-improvement common-sensically has to cause an intelligence explosion. I am sort of okay with this, but I feel like a commitment to avoiding analogies should involve not bringing up the chimp-human analogy further, which Eliezer continues to do, quite a lot. I do feel like Paul succeeded in convincing me that we shouldn’t place too much evidential weight on it. The Wimbledon Of Reference Class Tennis “Reference class tennis” is an old rationalist idiom for people throwing analogies back and forth. “AI will be slow, because it’s an economic transition like the Agricultural or Industrial Revolution, and those were slow!” “No, AI will be fast, because it’s an evolutionary step like chimps → humans, and that was fast!” “No, AI will be slow, because it’s an invention, like the computer, and computers were invented piecemeal and required decades of innovation to be useful.” “No, AI will be fast, because it’s an invention, like the nuclear bomb, and nuclear bombs went from impossible to city-killing in a single day.” “No, AI will be slow, because it will be surrounded by a shell-like metallic computer case, which makes it like a turtle, and turtles are slow.” “No, AI will be fast, because it’s dangerous and powerful, like a tiger, and tigers are fast!” And so on. Comparing things to other things is a time-tested way of speculating about them. But there are so many other things to compare to that you can get whatever result you want. This is the failure mode that the term “reference class tennis” was supposed to point to. Both participants in this debate are very smart and trying their hardest to avoid reference-class tennis, but neither entirely succeeds. Eliezer’s preferred classes are Bitcoin (“there wasn't a cryptocurrency developed a year before Bitcoin using 95% of the ideas which did 10% of the transaction volume”), nukes, humans/chimps, the Wright Brothers, AlphaGo (which really was a discontinuous improvement on previous Go engines), and AlphaFold (ditto for proteins). Paul’s preferred classes are the Agricultural and Industrial Revolutions, chess engines (which have gotten better along a gradual, well-behaved curve), all sorts of inventions like computers and ships (likewise), and world GDP. Eliezer already listed most of these in his Intelligence Explosion Microeconomics paper in 2013, and concluded that the space of possible analogies was contradictory enough that we needed to operate at a higher level. Maybe so, but when someone lobs a reference class tennis ball at you, it’s hard to resist the urge to hit it back. Recursive Self-Improvement This is where I think Eliezer most wants to take the discussion. The idea is: once AI is smarter than humans, it can do a superhuman job of developing new AI. In his Microeconomics paper, he writes about an argument he (semi-hypothetically) had with Ray Kurzweil about Moore’s Law. Kurzweil expected Moore’s Law to continue forever, even after the development of superintelligence. Eliezer objects: Suppose we were dealing with minds running a million times as fast as a human, at which rate they could do a year of internal thinking in thirty-one seconds, such that the total subjective time from the birth of Socrates to the death of Turing would pass in 20.9 hours. Do you still think the best estimate for how long it would take them to produce their next generation of computing hardware would be 1.5 orbits of the Earth around the Sun? That is: the fact that it took 1.5 years for transistor density to double isn’t a natural law. It’s pointing to a law that the amount of resources (most notably intelligence) that civilization focused on the transistor-densifying problem equalled the amount it takes to double it every 1.5 years. If some shock drastically changed available resources (by eg speeding up human minds a million times), this would change the resources involved, and the same laws would predict transistor speed doubling in some shorter amount of time (naively 0.000015 years, although realistically at that scale other inputs would dominate). So when Paul derives clean laws of economics showing that things move along slow growth curves, Eliezer asks: why do you think they would keep doing this when one of the discoveries they make along that curve might be “speeding up intelligence a million times”? (Eliezer actually thinks improvements in the quality of intelligence will dominate improvements in speed - AIs will mostly be smarter, not just faster - but speed is a useful example here and we’ll stick with it) Paul answers: Summary of my response: Before there is AI that is great at self-improvement there will be AI that is mediocre at self-improvement. Powerful AI can be used to develop better AI (amongst other things). This will lead to runaway growth. This on its own is not an argument for discontinuity: before we have AI that radically accelerates AI development, the slow takeoff argument suggests we will have AI that significantly accelerates AI development (and before that, slightly accelerates development). That is, an AI is just another, faster step in the hyperbolic growth we are currently experiencing, which corresponds to a further increase in rate but not a discontinuity (or even a discontinuity in rate). The most common argument for recursive self-improvement introducing a new discontinuity seems be: some systems “fizzle out” when they try to design a better AI, generating a few improvements before running out of steam, while others are able to autonomously generate more and more improvements. This is basically the same as the universality argument in a previous section. Eliezer: Oh, come on. That is straight-up not how simple continuous toy models of RSI work. Between a neutron multiplication factor of 0.999 and 1.001 there is a very huge gap in output behavior. Outside of toy models: Over the last 10,000 years we had humans going from mediocre at improving their mental systems to being (barely) able to throw together AI systems, but 10,000 years is the equivalent of an eyeblink in evolutionary time - outside the metaphor, this says, "A month before there is AI that is great at self-improvement, there will be AI that is mediocre at self-improvement." (Or possibly an hour before, if reality is again more extreme along the Eliezer-Hanson axis than Eliezer. But it makes little difference whether it's an hour or a month, given anything like current setups.) This is just pumping hard again on the intuition that says incremental design changes yield smooth output changes, which (the meta-level of the essay informs us wordlessly) is such a strong default that we are entitled to believe it if we can do a good job of weakening the evidence and arguments against it. And the argument is: Before there are systems great at self-improvement, there will be systems mediocre at self-improvement; implicitly: "before" implies "5 years before" not "5 days before"; implicitly: this will correspond to smooth changes in output between the two regimes even though that is not how continuous feedback loops work. I got a bit confused trying to understand the criticality metaphor here. There’s no equivalent of neutron decay, so any AI that can consistently improve its intelligence is “critical” in some sense. Imagine Elon Musk replaces his brain with a Neuralink computer which - aside from having read-write access - exactly matches his current brain in capabilities. Also he becomes immortal. He secludes himself from the world, studying AI and tinkering with his brain’s algorithms. Does he become a superintelligence? I think under the assumptions Paul and Eliezer are using, eventually maybe. After some amount of time he’ll come across a breakthrough he can use to increase his intelligence. Then, armed with that extra intelligence, he’ll be able to pursue more such breakthroughs. However intelligent the AI you’re scared of is, Musk will get there eventually. How long will it take? A good guess might be “years” - Musk starts out as an ordinary human, and ordinary humans are known to take years to make breakthroughs. Suppose it takes Musk one year to come up with a first breakthrough that raises his IQ 1 point. How long will his second breakthrough take? It might take longer, because he has picked the lowest-hanging fruit, and all the other possible breakthroughs are much harder. Or it might take shorter, because he’s slightly smarter than he was before, and maybe some extra intelligence goes a really long way in AI research. The concept of an intelligence explosion seems to assume the second effect dominates the first. This would match the observation that human researchers, who aren’t getting any smarter over time, continue making new discoveries. That suggests the range of possible discoveries at a given intelligence level is pretty vast. Some research finds that the usual pattern in science is constant rate of discovery from exponentially increasing number of researchers, suggesting strong low-hanging fruit effects, but these seem to be overwhelmed by other considerations in AI right now. I think Eliezer’s position on this subject is shaped by assumptions like: If you have an AI as intelligent as Elon Musk today, then tomorrow you can run it on more hardware with a bit of normal human algorithmic progress, and get one twice as intelligent. So even if it would take Elon years to make a breakthrough, long before those years are up you’ll have an AI that can make breakthroughs much faster.
Inline links: thirty years away, Biological Anchors, Intelligence Explosion Microeconomics, hyperbolic growth we are currently experiencing, Some research finds
Industrial Revolution? What Industrial Revolution? This is just a nice smooth exponential curve. The same is usually true of individual technologies; Paul doesn’t give specifics, but Nintil and Katja Grace both have lots of great examples: Information technologies over time (Nagy) Chess AI performance over time. Why does this matter? If there’s a slow takeoff (ie gradual exponential curve), it will become obvious that some kind of terrifying transformative AI revolution is happening, before the situation gets apocalyptic. There will be time to prepare, to test slightly-below-human AIs and see how they respond, to get governments and other stakeholders on board. We don’t have to get every single thing right ahead of time. On the other hand, because this is proceeding along the usual channels, it will be the usual variety of muddled and hard-to-control. With the exception of a few big actors like the US and Chinese government, and maybe the biggest corporations like Google, the outcome will be determined less by any one agent, and more by the usual multi-agent dynamics of political and economic competition. There will be lots of opportunities to affect things, but no real locus of control to do the affecting. If there’s a fast takeoff (ie sudden FOOM), there won’t be much warning. Conventional wisdom will still say that transformative AI is thirty years away. All the necessary pieces (ie AI alignment theory) will have to be ready ahead of time, prepared blindly without any experimental trial-and-error, to load into the AI as soon as it exists. On the plus side, a single actor (whoever has this first AI) will have complete control over the process. If this actor is smart (and presumably they’re a little smart, or they wouldn’t be the first team to invent transformative AI), they can do everything right without going through the usual government-lobbying channels. So the slower a takeoff you expect, the less you should be focusing on getting every technical detail right ahead of time, and the more you should be working on building the capacity to steer government and corporate policy to direct an incoming slew of new technologies. Yudkowsky Contra Christiano Eliezer counters that although progress may retroactively look gradual and continuous when you know what metric to graph it on, it doesn’t necessarily look that way in real life by the measures that real people care about. (one way to think of this: imagine that an AI’s effective IQ starts at 0.1 points, and triples every year, but that we can only measure this vaguely and indirectly. The year it goes from 5 to 15, you get a paper in a third-tier journal reporting that it seems to be improving on some benchmark. The year it goes from 66 to 200, you get a total transformation of everything in society. But later, once we identify the right metric, it was just the same rate of gradual progress the whole time. ) So Eliezer is much less impressed by the history of previous technologies than Paul is. He’s also skeptical of the “GDP will double in 4 years before it doubles in 1” claim, because of two contingent disagreements and two fundamental disagreements. The first contingent disagreement: government regulations make it hard to deploy imperfect things, and non-trivial to deploy things even after they’re perfect. Eliezer has non-jokingly said he thinks AI might destroy the world before the average person can buy a self-driving car. Why? Because the government has to approve self-driving cars (and can drag its feet on that), but the apocalypse can happen even without government approval. In Paul’s model, sometime long before superintelligence we should have AIs that can drive cars, and that increases GDP and contributes to a general sense that exciting things are going on. Eliezer says: fine, what if that’s true? Who cares if self-driving cars will be practical a few years before the world is destroyed? It’ll take longer than that to lobby the government to allow them on the road. The second contingent disagreement: superintelligent AIs can lie to us. Suppose you have an AI which wants to destroy humanity, whose IQ is doubling every six months. Right now it’s at IQ 200, and it suspects that it would take IQ 800 to build a human-destroying superweapon. Its best strategy is to lie low for a year. If it expects humans would turn it off if they knew how close it was to superweapons, it can pretend to be less intelligent than it really is. The period when AIs are holding back so we don’t discover their true power level looks like a period of lower-than-expected GDP growth - followed by a sudden FOOM once the AI gets its superweapon and doesn’t need to hold back. So even if Paul is conceptually right and fundamental progress proceeds along a nice smooth curve, it might not look to us like a nice smooth curve, because regulations and deceptive AIs could prevent mildly-transformative AI progress from showing up on graphs, but wouldn’t prevent the extreme kind of AI progress that leads to apocalypse. To an outside observer, it would just look like nothing much changed, nothing much changed, nothing much changed, and then suddenly, FOOM. But even aside from this, Eliezer doesn’t think Paul is conceptually right! He thinks that even on the fundamental level, AI progress is going to be discontinuous. It’s like a nuclear bomb. Either you don’t have a nuclear bomb yet, or you do have one and the world is forever transformed. There is a specific moment at which you go from “no nuke” to “nuke” without any kind of “slightly worse nuke” acting as a harbinger. He uses the example of chimps → humans. Evolution has spent hundreds of millions of years evolving brainier and brainier animals (not teleologically, of course, but in practice). For most of those hundreds of millions of years, that meant the animal could have slightly more instincts, or a better memory, or some other change that still stayed within the basic animal paradigm. At the chimp → human transition, we suddenly got tool use, language use, abstract thought, mathematics, swords, guns, nuclear bombs, spaceships, and a bunch of other stuff. The rhesus monkey → chimp transition and the chimp → human transition both involved the same ~quadrupling of neuron number, but the former was pretty boring and the latter unlocked enough new capabilities to easily conquer the world. The GPT-2 → GPT-3 transition involved centupling parameter count. Maybe we will keep centupling parameter count every few years, and most times it will be incremental improvement, and one time it will conquer the world. But even talking about centupling parameter points is giving Paul too much credit. Lots of past inventions didn’t come by quadrupling or centupling something, they came by discovering “the secret sauce”. The Wright brothers (he argues) didn’t make a plane with 4x the wingspan of the last plane that didn’t work, they invented the first plane that could fly at all. The Hiroshima bomb wasn’t some previous bomb but bigger, it was what happened after a lot of scientists spent a long time thinking about a fundamentally different paradigm of bomb-making and brought it to a point where it could work at all. The first transformative AI isn’t going to be GPT-3 with more parameters, it will be what happens after someone discovers how to make machines truly intelligent. (this is the same debate Eliezer had with Ajeya over the Biological Anchors post; have I mentioned that Ajeya and Paul are married?) Fine, Let’s Nitpick The Hell Out Of The Chimps Vs. Humans Example This is where the two of them end up, so let’s follow. Between chimps and humans, there were about seven million years of intermediate steps. These had some human capabilities, but not others. IE homo erectus probably had language, but not mathematics, and in terms of taking over the world it did make it to most of the Old World but was less dominant than moderns. But if we say evolutionary history started 500 million years ago (the Cambrian), and AI history started with the Dartmouth Conference in 1955, then the equivalent of 7 million years of evolutionary history is 1 year of AI history. In the very very unlikely and forced comparison where evolutionary history and AI history go at the same speed, there will be only about a year between chimp-level and human-level AIs. A chimp-level AI probably can’t double GDP, so this would count as a fast takeoff by Paul’s criterion. But even more than that, chimp → human feels like a discontinuity. It’s not just “animals kept getting smarter for hundreds of millions of years, and then ended up very smart indeed”. That happened for a while, and then all of sudden there was a near-instant phase transition into a totally different way of using intelligence with completely new abilities. If AI worked like this, we would have useful toys and interesting specialists for a few decades, until suddenly someone “got it right”, completed the package that was necessary for “true intelligence”, and then we would have a completely new category of thing. Paul admits this analogy is awkward for his position. He answers: Chimp evolution is not primarily selecting for making and using technology, for doing science, or for facilitating cultural accumulation. The task faced by a chimp is largely independent of the abilities that give humans such a huge fitness advantage. It’s not completely independent—the overlap is the only reason that evolution eventually produces humans—but it’s different enough that we should not be surprised if there are simple changes to chimps that would make them much better at designing technology or doing science or accumulating culture […] So I don’t think the example of evolution tells us much about whether the continuous change story applies to intelligence. This case is potentially missing the key element that drives the continuous change story—optimization for performance. Evolution changes continuously on the narrow metric it is optimizing, but can change extremely rapidly on other metrics. For human technology, features of the technology that aren’t being optimized change rapidly all the time. When humans build AI, they will be optimizing for usefulness, and so progress in usefulness is much more likely to be linear. That is, evolution wasn’t optimizing for tool use/language/intelligence, so we got an “overhang” where chimps could potentially have been very good at these, but evolution never bothered “closing the circuit” and turning those capabilities “on”. After a long time, evolution finally blundered into an area where marginal improvements in these capacities improved fitness, so evolution started improving them and it was easy. Imagine a company which, through some oversight, didn’t have a Sales department. They just sat around designing and manufacturing increasingly brilliant products, but not putting any effort into selling them. Then the CEO remembers they need a Sales department, starts one up, and the company goes from moving near zero units to moving millions of units overnight. It would look like the company had “suddenly” developed a “vast increase in capabilities”. But this is only possible when a CEO who is weirdly unconcerned about profit forgets to do obvious profit-increasing things for many years. This is Paul’s counterargument to the chimp analogy. Evolution isn’t directly concerned about various intellectual skills; it only wants them in the unusual cases where they’ll contribute to fitness on the margin. AI companies will be very concerned about various intellectual skills. If there’s a trivial change that can make their product 10x better, they’ll make it. So AI capabilities will grow in a “well-rounded” way, there won’t be any “overhangs”, and there won’t be any opportunities for a sudden overhang-solving phase transition with associated new-capability development like with chimps → humans. Eliezer answers: Chimps are nearly useless because they're not general, and doing anything on the scale of building a nuclear plant requires mastering so many different nonancestral domains that it's no wonder natural selection didn't happen to separately train any single creature across enough different domains that it had evolved to solve every kind of domain-specific problem involved in solving nuclear physics and chemistry and metallurgy and thermics in order to build the first nuclear plant in advance of any old nuclear plants existing. Humans are general enough that the same braintech selected just for chipping flint handaxes and making water-pouches and outwitting other humans, happened to be general enough that it could scale up to solving all the problems of building a nuclear plant - albeit with some added cognitive tech that didn't require new brainware, and so could happen incredibly fast relative to the generation times for evolutionarily optimized brainware. Now, since neither humans nor chimps were optimized to be "useful" (general), and humans just wandered into a sufficiently general part of the space that it cascaded up to wider generality, we should legit expect the curve of generality to look at least somewhat different if we're optimizing for that. Eg, right now people are trying to optimize for generality with AIs like Mu Zero and GPT-3. In both cases we have a weirdly shallow kind of generality. Neither is as smart or as deeply general as a chimp, but they are respectively better than chimps at a wide variety of Atari games, or a wide variety of problems that can be superposed onto generating typical human text. They are, in a sense, more general than a biological organism at a similar stage of cognitive evolution, with much less complex and architected brains, in virtue of having been trained, not just on wider datasets, but on bigger datasets using gradient-descent memorization of shallower patterns, so they can cover those wide domains while being stupider and lacking some deep aspects of architecture. It is not clear to me that we can go from observations like this, to conclude that there is a dominant mainline probability for how the future clearly ought to go and that this dominant mainline is, "Well, before you get human-level depth and generalization of general intelligence, you get something with 95% depth that covers 80% of the domains for 10% of the pragmatic impact". ...or whatever the concept is here, because this whole conversation is, on my own worldview, being conducted in a shallow way relative to the kind of analysis I did in Intelligence Explosion Microeconomics, where I was like, "here is the historical observation, here is what I think it tells us that puts a lower bound on this input-output curve". Here Eliezer sort of kind of grants Paul’s point that AIs will be optimized for generality in a way chimps aren’t, but points to his previous “Intelligence Explosion Microeconomics” essay to argue that we should expect a fast takeoff anyway. IEM has a lot of stuff in it, but one key point is that instead of using analogies to predict the course of future AI, we should open that black box and try to actually reason about how it will work, in which case we realize that recursive self-improvement common-sensically has to cause an intelligence explosion. I am sort of okay with this, but I feel like a commitment to avoiding analogies should involve not bringing up the chimp-human analogy further, which Eliezer continues to do, quite a lot. I do feel like Paul succeeded in convincing me that we shouldn’t place too much evidential weight on it. The Wimbledon Of Reference Class Tennis “Reference class tennis” is an old rationalist idiom for people throwing analogies back and forth. “AI will be slow, because it’s an economic transition like the Agricultural or Industrial Revolution, and those were slow!” “No, AI will be fast, because it’s an evolutionary step like chimps → humans, and that was fast!” “No, AI will be slow, because it’s an invention, like the computer, and computers were invented piecemeal and required decades of innovation to be useful.” “No, AI will be fast, because it’s an invention, like the nuclear bomb, and nuclear bombs went from impossible to city-killing in a single day.” “No, AI will be slow, because it will be surrounded by a shell-like metallic computer case, which makes it like a turtle, and turtles are slow.” “No, AI will be fast, because it’s dangerous and powerful, like a tiger, and tigers are fast!” And so on. Comparing things to other things is a time-tested way of speculating about them. But there are so many other things to compare to that you can get whatever result you want. This is the failure mode that the term “reference class tennis” was supposed to point to. Both participants in this debate are very smart and trying their hardest to avoid reference-class tennis, but neither entirely succeeds. Eliezer’s preferred classes are Bitcoin (“there wasn't a cryptocurrency developed a year before Bitcoin using 95% of the ideas which did 10% of the transaction volume”), nukes, humans/chimps, the Wright Brothers, AlphaGo (which really was a discontinuous improvement on previous Go engines), and AlphaFold (ditto for proteins). Paul’s preferred classes are the Agricultural and Industrial Revolutions, chess engines (which have gotten better along a gradual, well-behaved curve), all sorts of inventions like computers and ships (likewise), and world GDP. Eliezer already listed most of these in his Intelligence Explosion Microeconomics paper in 2013, and concluded that the space of possible analogies was contradictory enough that we needed to operate at a higher level. Maybe so, but when someone lobs a reference class tennis ball at you, it’s hard to resist the urge to hit it back. Recursive Self-Improvement This is where I think Eliezer most wants to take the discussion. The idea is: once AI is smarter than humans, it can do a superhuman job of developing new AI. In his Microeconomics paper, he writes about an argument he (semi-hypothetically) had with Ray Kurzweil about Moore’s Law. Kurzweil expected Moore’s Law to continue forever, even after the development of superintelligence. Eliezer objects: Suppose we were dealing with minds running a million times as fast as a human, at which rate they could do a year of internal thinking in thirty-one seconds, such that the total subjective time from the birth of Socrates to the death of Turing would pass in 20.9 hours. Do you still think the best estimate for how long it would take them to produce their next generation of computing hardware would be 1.5 orbits of the Earth around the Sun? That is: the fact that it took 1.5 years for transistor density to double isn’t a natural law. It’s pointing to a law that the amount of resources (most notably intelligence) that civilization focused on the transistor-densifying problem equalled the amount it takes to double it every 1.5 years. If some shock drastically changed available resources (by eg speeding up human minds a million times), this would change the resources involved, and the same laws would predict transistor speed doubling in some shorter amount of time (naively 0.000015 years, although realistically at that scale other inputs would dominate). So when Paul derives clean laws of economics showing that things move along slow growth curves, Eliezer asks: why do you think they would keep doing this when one of the discoveries they make along that curve might be “speeding up intelligence a million times”? (Eliezer actually thinks improvements in the quality of intelligence will dominate improvements in speed - AIs will mostly be smarter, not just faster - but speed is a useful example here and we’ll stick with it) Paul answers: Summary of my response: Before there is AI that is great at self-improvement there will be AI that is mediocre at self-improvement. Powerful AI can be used to develop better AI (amongst other things). This will lead to runaway growth. This on its own is not an argument for discontinuity: before we have AI that radically accelerates AI development, the slow takeoff argument suggests we will have AI that significantly accelerates AI development (and before that, slightly accelerates development). That is, an AI is just another, faster step in the hyperbolic growth we are currently experiencing, which corresponds to a further increase in rate but not a discontinuity (or even a discontinuity in rate). The most common argument for recursive self-improvement introducing a new discontinuity seems be: some systems “fizzle out” when they try to design a better AI, generating a few improvements before running out of steam, while others are able to autonomously generate more and more improvements. This is basically the same as the universality argument in a previous section. Eliezer: Oh, come on. That is straight-up not how simple continuous toy models of RSI work. Between a neutron multiplication factor of 0.999 and 1.001 there is a very huge gap in output behavior. Outside of toy models: Over the last 10,000 years we had humans going from mediocre at improving their mental systems to being (barely) able to throw together AI systems, but 10,000 years is the equivalent of an eyeblink in evolutionary time - outside the metaphor, this says, "A month before there is AI that is great at self-improvement, there will be AI that is mediocre at self-improvement." (Or possibly an hour before, if reality is again more extreme along the Eliezer-Hanson axis than Eliezer. But it makes little difference whether it's an hour or a month, given anything like current setups.) This is just pumping hard again on the intuition that says incremental design changes yield smooth output changes, which (the meta-level of the essay informs us wordlessly) is such a strong default that we are entitled to believe it if we can do a good job of weakening the evidence and arguments against it. And the argument is: Before there are systems great at self-improvement, there will be systems mediocre at self-improvement; implicitly: "before" implies "5 years before" not "5 days before"; implicitly: this will correspond to smooth changes in output between the two regimes even though that is not how continuous feedback loops work. I got a bit confused trying to understand the criticality metaphor here. There’s no equivalent of neutron decay, so any AI that can consistently improve its intelligence is “critical” in some sense. Imagine Elon Musk replaces his brain with a Neuralink computer which - aside from having read-write access - exactly matches his current brain in capabilities. Also he becomes immortal. He secludes himself from the world, studying AI and tinkering with his brain’s algorithms. Does he become a superintelligence? I think under the assumptions Paul and Eliezer are using, eventually maybe. After some amount of time he’ll come across a breakthrough he can use to increase his intelligence. Then, armed with that extra intelligence, he’ll be able to pursue more such breakthroughs. However intelligent the AI you’re scared of is, Musk will get there eventually. How long will it take? A good guess might be “years” - Musk starts out as an ordinary human, and ordinary humans are known to take years to make breakthroughs. Suppose it takes Musk one year to come up with a first breakthrough that raises his IQ 1 point. How long will his second breakthrough take? It might take longer, because he has picked the lowest-hanging fruit, and all the other possible breakthroughs are much harder. Or it might take shorter, because he’s slightly smarter than he was before, and maybe some extra intelligence goes a really long way in AI research. The concept of an intelligence explosion seems to assume the second effect dominates the first. This would match the observation that human researchers, who aren’t getting any smarter over time, continue making new discoveries. That suggests the range of possible discoveries at a given intelligence level is pretty vast. Some research finds that the usual pattern in science is constant rate of discovery from exponentially increasing number of researchers, suggesting strong low-hanging fruit effects, but these seem to be overwhelmed by other considerations in AI right now. I think Eliezer’s position on this subject is shaped by assumptions like: If you have an AI as intelligent as Elon Musk today, then tomorrow you can run it on more hardware with a bit of normal human algorithmic progress, and get one twice as intelligent. So even if it would take Elon years to make a breakthrough, long before those years are up you’ll have an AI that can make breakthroughs much faster.
Inline links: Nintil, Katja Grace, https://substackcdn.com/image/fetch/$s_!8pqu!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fbucketeer-e05bbc84-baa3-437e-9518-adb32be77984.s3.amazonaws.com%2Fpublic%2Fimages%2F3593bc53-e7e1-4aae-a129-0d59e9220b58_648x739.png, Nagy, https://substackcdn.com/image/fetch/$s_!FvS1!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fbucketeer-e05bbc84-baa3-437e-9518-adb32be77984.s3.amazonaws.com%2Fpublic%2Fimages%2F985b3c20-cf30-4b4b-a12f-8e5deb10e5af_850x700.png, thirty years away, Biological Anchors, Intelligence Explosion Microeconomics, hyperbolic growth we are currently experiencing, Some research finds
Highlights From The Comments On Bobos In Paradise
December 09, 2022 · Original source
One thing that_would seem to be important is new money. For centuries, the only way to be rich was to own a lot of land, and the only way to own a lot of land was to inherit it. The Industrial Revolution started a phenomenon of non-U people suddenly becoming rich, which made life complicated for the old upper class, but at first they could absorb these new money richers slowly into their ranks (and more importantly, the new money richers aspired to emulate the old money). But eventually the rate of wealth creation got so out of hand that new millionaires were being minted faster than the upper class could co-opt them, and the wealth of the unassimilated non-U rich started to outweigh the wealth of the true Upper Class. And eventually the whole thing came tumbling down and everyone is lining up to get a glimpse of the Beatles instead of the Queen.
23
October 31, 2023 · Original source
Robin Hanson compares this to steam engines. People had steam engines since ancient Greece, but they didn’t catch on until the Industrial Revolution. And part of this is that you can’t just open a stall in the bazaar saying “Steam engines for sale! Get your steam engines!” You need to figure out a specific niche for which the steam engines of your day are economically efficient, become really familiar with that niche, build a steam engine specifically suited to it, and then work with your clients to turn it into a packaged, easy to buy, easy to use solution. For real steam engines, that first niche was pumping water out of coal mines. For prediction markets, it’s . . . what?
Interview Day At Thiel Capital
September 03, 2024 · Original source
“So the more Jews you have, the more superweapons you have. If you want a world to survive a couple of centuries after the Industrial Revolution, you need some kind of incredibly implausible event that gets rid of Jews in particular. So everyone who’s still alive will find themselves in a world with a history of implausible events like this.”
Book Review: Deep Utopia
October 17, 2024 · Original source
Is this stupid? We hold an Industrial Revolution, design artificial intelligence, go through an entire singularity - only to end up back in 18th-century France? Not necessarily. 18th century France was full of miserable people - starving peasants, wretched prisoners, smallpox victims - or people who suffered merely from the “affliction” of not getting to be Napoleon. Even Napoleon’s life wasn’t maximally interesting. You could live an enriched, enchanted version of Napoleon’s life, where every decision had extra branching consequences and there was no downtime. Or a version of Napoleon’s life customized to your personal preferences: NSFW Napoleon where all the women wear skimpy costumes! Woke Napoleon where every third Frenchman is a person of color! If you hate the cold, you can play as a Napoleon who skipped Russia and invaded Cancun! It would be the difference between living in medieval Scandinavia and playing Assassin’s Creed: Valhalla.
Highlights From The Comments On Vibecession
December 31, 2025 · Original source
Obviously nothing real changes the exact second a new president is inaugurated, so people must be using questions about the economy to express their overall happiness about the state of the world. Alex asks whether increasing political polarization could make this worse. Both parties’ extreme factions share a tendency to treat the country as controlled by a hegemonic conspiracy of their enemies - the woke coastal elite Soros cosmopolitan establishment, or the neoliberal fat cat Koch Brothers tech oligarch blob. Does this mean everyone is getting some multiple of the “other party’s president is in power” effect all the time? 3: Discourse Downstream Of The Mike Green $140K Poverty Line Post … Shovacklerod writes: Scott have you read Mike Green’s viral post on this? His main argument is that the poverty line is miscalculated, but in context of declining middle class sentiments— The more interesting thesis is that there exists a “valley of death” where two parents in the workforce need a combined ~$140k salary otherwise the cumulative “participation costs” of a fast modern society (for example a phone plan or child care) make year-over-year capital accumulation near impossible. I haven’t, but other commenters suggest reading responses, including Noah Smith’s The $140,000 Poverty Line Is Very Silly, Jeremy Horpedahl’s The Poverty Line Is Not $140,000, and Tyler Cowen’s The Myth Of The $140,000 Poverty Line. Most of these focus on Green’s explicit errors - for example, he gets most of his cost-of-living numbers from Essex, NJ, an especially rich county, then compares them to average earnings. Correct half a dozen things like this, and the real poverty line is probably somewhere between $35K - $60K. The percent of Americans below this line continues to decline every year, as it has for decades. Green finally pseudo-apologized, lambasting the “mockery machine” of the “cognitive elite” but admitting that his post “was never intended to go viral and was written for my existing audience that tends to be pretty understanding that I don’t do this for a living, but rather as PART of my living” Still, many people took Green’s article as a starting point to contribute to the Vibecession discourse, so let’s go over the ones that touch on our topic in more detail. Lincicome titles his response The $140,000 Poverty Line Is Wrong, So Why Does It Feel Right?, and blames Baumol’s cost disease: As the Financial Times’ John-Burns Murdoch just detailed, Americans’ overall cost of living has improved over time, but certain highly visible and socially desirable services have become more expensive. That’s not a conspiracy against the middle class but instead just Baumol at work: “[A]s countries develop economically, the same productivity growth that drives down the cost of tradeable goods causes the cost of in-person services to balloon. Wages in sectors like healthcare and education that require intensive face-to-face labour, and have slow (if any) productivity growth, are forced upwards in order to attract workers who would otherwise opt for high-paying work in more productive sectors. The result is that even if people keep consuming the exact same basket of goods and services, as living standards in their country increase they will find more and more of their spending is going on essential services.” Sectors where productivity grows slowly and prices outpace inflation—health care, education, child care, personal services, housing (construction), etc.—happen to be the same ones that middle-class families notice most and that signal social status. As we’ve all gotten richer, moreover, these services have transitioned from luxuries to expectations. Throw in the hedonic treadmill and the fact that you can’t price-shop schools or hospitals the way you can TVs, and public alarm is all but inevitable. I’m suspicious of including “housing (construction)” on this list - couldn’t you use the same argument to reclassify any manufactured good as a service good? - but the rest of these are well-taken. Still, did Baumol or the other economists who first discussed the effect in the 1960s predict it would make people feel like things were outright worse, as opposed to just getting better less than would be expected from raw productivity numbers? Seems strange. Also, hasn’t the Baumol effect been basically constant since at least the Industrial Revolution? And isn’t the Vibecession only 5 - 20 years old? Matt Bruenig has his own response to Green, Why Do People Feel Like They’re Falling Behind? He bases his argument around this graph: …which is just making the common-sense point that, as society shifts from one-income to two-income families, the husband’s share of family income drops from ~100% to ~50%. So, Bruenig argues, if everyone is trying to keep up with the Joneses, and the Joneses are a dual-earner family, then this single working man has gone from making 100% of his comparison point, to making only 50%. This is a cool potential cognitive bias, but is anyone really making this mistake? Vibecession complaints hardly seem limited to men in traditional one-earner households wondering why they’re not making as much as the neighbors whose wife is a fancy lawyer. My impression is that they include both two-earner families who still feel like they’re falling behind, and (most of all) young singles who are comparing themselves to their young single friends where this issue never comes up in the first place. Matt Yglesias uses a similar strategy in You Can Afford A Tradlife. This is what they took from you. They never should have passed the ‘Make It Illegal To Wear Hair Gel And Marry A White Woman Act' back in 1959! He argues that the reason most wives work these days isn’t because we’re poorer (and they have to work to survive), but because we’re richer (and so wives can make so much money working outside the home that the opportunity cost is too high to pass up). A single earner could still support a family on a 1950s lifestyle. It would just feel like a failure, because we don’t realize how much worse than 1950s lifestyle was compared to our current conditions. The article’s paywalled, but you can get a pretty good sense of the argument from these paragraphs. After determining that the median man makes about $80,000/year, he writes: Let’s say our $80,000-a-year man is living in the Jacksonville area. The Department of Housing and Urban Development calculates what are called Fair Market Rents for each American metro — this means the 40th percentile rent for a home with any given set of characteristics. They say F.M.R. for a three-bedroom home in the Jacksonville area is $2,163. That comes out to about 30 percent of Mr. Median’s annual income. Can you really get a place to live for that little? Here’s a lovely three-bedroom home in the East Arlington neighborhood for $2,020 a month, and it’s zoned for an elementary school with a 10-out-of-10 ranking from GreatSchools. It’s true that 1,617 square feet is on the small side for, say, a family of five in the contemporary United States. But the average size of a new single family home was 1,289 square feet in 1960 and 1,500 square feet in 1970. Two of your kids are going to need to share a bedroom, but that’s how people lived back in the day. There’s more to life than housing, of course, but I started there because that’s the largest item in a household budget. Durable goods like furniture, cars, and appliances have all become better and more affordable since the mid-1960s. That’s partially offset by rising prices for things like college tuition, child care, and health care. But in the 1960s, most young people didn’t go to college. The way health insurance works, you only need one worker in your family to get a job-based health plan. And of course, with your wife serving as a full-time homemaker, you don’t need to worry about child care expenses. The big thing is that, with a larger family, you literally have a bunch of mouths to feed. But the model here is to replicate how people actually lived in the mid-1960s, which is that they dined out much less frequently and also spent a much larger share of their total income on food. When I try to retrace this, it seems possible, but barely. I imagined doing this in Sacramento, to be near family. Suppose I make $80K pretax = $6.6K/month pretax = $5K per month posttax. A cheap 3-bedroom house on a nice-enough block is $2200 mortgage, assume $3K after property taxes etc. A cheap new car is $350/month. Food can be arbitrarily low if you’re willing to eat rice all the time, but let’s say $250/month. CoveredCalifornia offered my family of four healthcare for $600/month. So top four expenses take $4200/month of the $5000/month pretax income. I don’t know; seems tough. I would like to see a more thorough breakdown of an average 2026 vs. 1956 man’s likely budget. There are also some areas where it’s harder to separate genuine declines from rising expectations. Most people in the 1950s didn’t have health insurance. Was that because they accepted lower levels of health, or because medical care was cheaper, and easy enough to afford out-of-pocket? Probably some very complicated combination of both. And it might be impossible to get certain kinds of 1950s medical care today, i.e. a bed in a cheap low-quality shared hospital room. (some of the best discussion around this came from the response to Elizabeth Warren’s The Two-Income Trap, see eg Matt Bruenig here) Still, I find this tangential to the main point. Yes, a few conservatives complain that it’s hard to have a single-income family. But most vibecession complaints come from singles or dual-earner households! 4: What About Other Countries? … Dionysus writes: Did you know that China also has a vibecession? If even China can’t regulate social media heavily enough to prevent this phenomenon, how can any liberal society possibly hope to? The link goes to an NYT article, which includes quotes like: Using apps like RedNote and Douyin, people are reviving memories of the 2000s and the early 2010s with photos of daring outfits, upbeat songs and vintage TV commercials, all of which, in different ways, evoke a time in China that pulsed with optimism. “The music back then throbbed with exuberance, brimming with the sense that the future could only get brighter,” a middle-aged man said in a RedNote video. “Today’s lyrics begin with lines like, ‘We’re trying our best to survive.’” And The boom-time beauty meme is the latest expression of a Gen Z counterculture born of disillusionment, the recognition that they may be the first generation in half a century unlikely to surpass their parents’ standard of living, no matter how hard they try. Over the past five years, this quiet resistance has taken many forms. It began with “lying flat,” a refusal to join the rat race. Some chose to pursue the “run philosophy,” or emigrating in search of freedom and brighter prospects. Others declared themselves the “last generation,” vowing not to have children. Still others embraced “let it rot,” giving up on difficult goals rather than battling for uncertain rewards. To show they could care less about career prospects, many took to wearing “gross outfits” at work. This is especially crazy in China, where GDP per capita is now ten times what it was back during the “Boom Years” that everyone reminisces about. This might be the smoking gun that people’s economic beliefs are totally unmoored from how rich they are. The Chinese story has an obvious moral: people care about growth rate more than level. But even this doesn’t work for America - our Vibecession doesn’t correspond to a period of unusually low growth. machine_spirit writes: It’s interesting to compare it to Europe as the control group. Unlike the US, whose economy muddled through just fine during the last decade, we are currently experiencing a massive economic decline that could soon turn into a full-blown collapse. And yet, outside of debates about immigration or foreign policy especially regarding Ukraine you don’t really hear the same level of rancour about ‘things being bad’ in the local media. I’m surprised to hear this. I hear many economic complaints from Europeans, but I suppose this passes through my own American filter bubble which is incentivized to talk about economic hardship for its own American reasons. Golden Feather writes: I am an Italian currently living in the US. My main guesses would be: Right-wing parties control a supermajority of TV and print media. They have also been in the govt most of the time, which means they control the state TV and have an interest in presenting things as rosey. The much older population makes the internet less relevant for public sentiment. Even in the few years where they were at the opposition, they mostly focused on immigration and crime to rile up popular sentiment, I guess because the population is older, their voters even moreso, so they care more about that than about the economy
Inline links: writes, The $140,000 Poverty Line Is Very Silly, The Poverty Line Is Not $140,000, The Myth Of The $140,000 Poverty Line, The $140,000 Poverty Line Is Wrong, So Why Does It Feel Right?, Baumol’s cost disease, detailed, Why Do People Feel Like They’re Falling Behind?, https://substackcdn.com/image/fetch/$s_!E2rN!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff94f8851-17f5-4a98-b1f8-349f568d23bb_1024x800.png, You Can Afford A Tradlife, https://substackcdn.com/image/fetch/$s_!ljt5!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1b1176e8-f932-430a-b7e1-c699a5ecf15c_581x479.png, They say, lovely three-bedroom home in the East Arlington neighborhood for $2,020 a month, average size of a new single family home was 1,289 square feet in 1960 and 1,500 square feet in 1970, how people actually lived in the mid-1960s, dined out much less frequently, see eg Matt Bruenig here, writes, China also has a vibecession?, lying flat, run philosophy, gross outfits, writes, writes
Open Thread 415
January 05, 2026 · Original source
If we do get a crazy AI future, and the economy grows 100x (Industrial Revolution scale) or 1000000x (solar system colonization scale) in your lifetime, then you only need a little capital to remain as absolutely well-off as you are today. For example, after 100x growth, anyone with $25,000 in the stock market now would have $2.5 million.