What this covers

Alex Imas and Phil Trammell conduct a wide-ranging conversation with Dwarkesh Patel on what economic scarcity will remain after AGI arrives. Rather than settling on a single forecast, they argue that the long-run effects on labor, inequality, and development hinge on a cluster of empirical variables—chiefly whether demand for new goods keeps pace with automation, whether AI becomes a commoditized utility or concentrates rents, and how well ordinary people can index into AGI wealth. The discussion treats these as open questions requiring data collection and scenario planning rather than confident point estimates.

The conversation ranges across mechanisms that could sustain labor demand even amid high automation: elasticity of demand and increasing variety could redirect spending toward new services and goods; the O-ring model explains why keeping a human in part of a job can sometimes block automation of the remainder. They examine what becomes scarce by analogy—whether AI behaves like electricity (diffusing benefits widely) or social media (concentrating them)—and what this means for developing countries caught between leapfrogging and abandonment. A thread throughout is the possibility that non-satiating wealth holders, if freed from inheritance dissipation, could come to dominate the economy, though Imas pushes back on whether such preferences are truly intrinsic or whether diminishing marginal utility constrains them. The speakers disagree on whether UBI or universal basic capital is preferable, whether slow automation is safer than rapid transition, and how much safety must be traded away to avoid concentrated AI power. Throughout, they identify crucial missing data—task-level demand elasticities, detailed employment flows, modern O*NET catalogs—as the prerequisite for any reliable forecast.

Sharpest takeaway

Imas and Trammell argue that the long-run effect of AGI on labor share, inequality, and global development is genuinely indeterminate and hinges on a few key variables—demand satiation, increasing variety, elasticity, and how easily the AI economy can be indexed—so the productive move is scenario-mapping and data collection, not confident point forecasts.

  • Whether labor share collapses depends on whether new capital varieties keep arriving fast enough to outrun satiation, which is an empirical question we lack data on.
  • Demand collapse and negative growth require implausibly hard bounds on rich consumers' demand plus refusal to invest.
  • Whether developing countries and ordinary people share in AGI gains depends on whether AI behaves like commoditized electricity or rent-capturing social media, i.e. how indexable it is.

The claims · ranked46 claims · weighted by value

This asset isn't compiled yet

You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.

0.87

Non-satiating accumulators have always existed but have not taken over the economy because of 'dissipation shocks'—they die and pass wealth to heirs who squander it or fail to match economic growth, or place it in foundations that spend it—so if longevity or aligned trusts remove these shocks, a few such agents could come to dominate the economy.

causalhigh valuecontestednovelty 4/4durability 4/4· Phil Trammell

“They just haven’t really taken over the economy historically because there have been these dissipation shocks, as they’re called. They’ve given it to their kids who squandered it, or they put it in foundations which spent it.”

0.86

Whether automating most tasks of a job raises or lowers employment depends on elasticity of demand: in the O-ring model, if automating nine of ten tasks makes the human more productive and cuts prices, sufficiently elastic demand can increase total purchases and lead to more hiring; Jevons paradox only occurs for highly elastic goods, whereas inelastic goods like oil, insulin, or food do not scale demand with falling price.

causalhigh valueestablishednovelty 3/4durability 4/4· Alex Imas

“If that translates into a price effect where the product is actually cheaper, and if demand responds enough where it’s being bought more and the service is being used more, that could actually lead to more hiring.”

0.83

If some agents do not satiate in capital—e.g. they are intrinsically engaged by accumulation or by exploring the universe—then, being rational, they maintain a higher savings rate and in the long run come to hold most of the wealth, so the overall capital share converges to the capital share of that non-satiating person's spending, which is effectively one.

causalhigh valuecontestednovelty 4/4durability 4/4· Phil Trammell

“the person who doesn’t satiate in capital is going to, if they’re being rational, have a higher savings rate. So in the long run, they’re going to have most of the wealth, and the overall capital share will basically be the capital share of that person’s spending, which is going to be one.”

0.82

Whether AI's gains are broadly shared depends on whether AI behaves like electricity—where downstream users capture most benefits and no provider amasses outsized power (ConEd)—or like social media, where the platform captures the rents; if every future S&P company succeeds by leveraging AI, the gains are indexable, whereas concentrated rent-capture makes indexing AGI hard.

forecasthigh valuecontestednovelty 4/4durability 3/4· Alex Imas

“Is AI going to be like electricity or social media? ... with electricity, a lot of the downstream benefits actually came to the users of the electricity rather than the actual entity producing it. On the other hand, with social media, it was the opposite case... the rents went to the platform.”

0.80

David Ricardo correctly predicted in the 1820s that the value-creating jobs of his era would be automated, but he missed that structural change would make automated goods cheap, freeing income to be spent on new services and new jobs—the lump-of-labor fallacy—which is why prime-age employment is now near record highs.

causalhigh valueestablishednovelty 2/4durability 4/4· Alex Imas

“What David Ricardo ended up missing is that you have these economics of structural change, where everything that got automated became cheap. People had more money to spend, and then they started spending it on services. This is the lump-of-labor fallacy.”

0.78

A Mongolian in 1400 forecasting what would be scarce after automation would have wrongly concluded that satiation in horse-transport, yogurt, and yurts would leave all spending flowing to human singers; instead wealth and machines expanded the range of non-singer goods, keeping the singer share negligible—suggesting increasing variety, not a collapse into human-intrinsic goods, is the central expectation.

forecasthigh valuecontestednovelty 3/4durability 4/4· Phil Trammell

“As we’ve accumulated more wealth and more advanced machines, we’ve expanded the range of things other than singers to spend our money on, and the share spent on singers has stayed negligible. Likewise, that’s my central prediction about how the future unfolds, though it could go either way.”

0.78

O-ring dynamics—where one faulty component (like the Challenger seal) destroys the whole product—explain why entire jobs are not yet automated: AI must hit extreme reliability or it degrades the finished good. But the same logic runs the other way once AI is advanced: a human kept for one-tenth of a job can pull down the quality or speed of the AI-produced remainder, so automating nine-tenths to a lower standard may not be worth it (Gans and Goldfarb).

causalhigh valuecontestednovelty 3/4durability 4/4· Dwarkesh Patel

“O-ring theory refers to the fact that the Challenger shuttle blew up because one component malfunctioned, and it destroyed the whole thing. Maybe that’s a more general model of how goods are produced in the economy.”

0.77

Although poor education makes retraining a weak strategy for poor countries, transformative AI could enable leapfrogging—as mobile banking became more prevalent in Nigeria than Germany—letting them skip intermediate development stages and achieve astronomical growth.

forecasthigh valuecontestednovelty 3/4durability 3/4· Alex Imas

“there are cases where, in developing countries, you had this leapfrogging effect with, for example, mobile banking. It’s much more prevalent in Nigeria than it is in Germany.”

0.75

Standard macro models wrongly treat output as a single substance allocable one-for-one to capital or consumption; in an AI world there is investment-specific technical change where one robot turns into many robots next year (high own-unit returns) but the price of capital goods falls fast relative to consumption, so each unit of capital gives up far less consumption than before and the real interest rate depends on which units you measure in.

definitionhigh valuecontestednovelty 4/4durability 3/4· Phil Trammell

“What’s happening is what would be called investment-specific technical change. The price of capital is falling relative to the price of consumption, instead of doing the standard macro thing of saying there’s just output, this chimera of a thing called output.”

0.73

Economists lack the basic data—consumer demand elasticities, tracking of jobs created or destroyed, an up-to-date O*NET task database—needed to forecast AI's labor effects, so a 'Manhattan Project for data' is required before reliable predictions are possible.

factualhigh valuecontestednovelty 3/4durability 3/4· Alex Imas

“We don’t have data on consumer demand elasticities. We don’t know what they are. We’re not really tracking what jobs are getting created or destroyed. The O*NET database, with all of the tasks and different jobs, has been rarely updated and is super low quality.”

0.73

A slow 'drip' automation scenario may be among the worst outcomes because of political economy: rather than mass unemployment, displaced workers move into lower-paying sectors and become underemployed (as phone operators did over 1920-1940), and even a fast 2-3% rise in unemployment can flip the political winds and trigger a national emergency.

causalhigh valuecontestednovelty 3/4durability 3/4· Alex Imas

“What you might see is people not really being unemployed en masse, but moving into sectors that pay them less money. This is what happened with phone operators between 1920 and 1940.”

0.73

UBI risks a dangerous power-sharing arrangement because people become dependent on a government check rather than on labor-derived income, making basic needs hostage to whoever is in power; universal basic capital is preferable because it gives people an ownership share and property rights, but it faces the hard problem of targeting which assets to put in portfolios.

normativehigh valuecontestednovelty 3/4durability 3/4· Alex Imas

“When that is no longer the case and we are at the mercy of the elected official for basic needs, that feels like a power-sharing arrangement that’s really dangerous.”

0.73

For both displaced workers and developing countries, indexing into AI wealth (buying the relevant supply chains or broad market) is a cleaner and more likely-to-succeed strategy than retraining or building local data centers, but its feasibility depends on how concentrated returns are and on whether ordinary people's main capital—houses—is poorly suited to complement AI production.

normativehigh valuecontestednovelty 3/4durability 3/4· Dwarkesh Patel

“I think you guys are suggesting something closer to just buying the index of AGI. That’s probably a much cleaner strategy and much more likely to succeed.”

0.73

The messy-middle and developing-country dangers are limited because the same scenario implies a very high interest rate (equivalently, rapidly falling prices for everything except human-intrinsic goods), so even a little savings turns into a lot of consumption next year; things must go very wrong—capital productive enough to automate work but not enough to push interest rates high—for this cushion to fail.

causalhigh valuecontestednovelty 3/4durability 3/4· Phil Trammell

“the interest rate would be way higher, and/or, equivalently, the price of everything except human-intrinsic goods would be falling really rapidly... A little bit of savings would turn into a lot of consumption next year.”

0.73

Middle-income developing countries face a fork: AI technology could dissipate to places like Nigeria and level the playing field, or—lacking resources to train models or own hardware—they could be left behind, especially as automation lets developed countries produce commodities domestically and remove even the consumer-market role these countries played.

forecasthigh valuecontestednovelty 3/4durability 3/4· Alex Imas

“there’s another world where, because they don’t have enough resources, they’re not training the models, they don’t have the hardware, and they just completely get left behind. And because of automation, we can produce commodities in developed countries now. Then we don’t even have the consumer market. That world looks pretty bad.”

0.73

The right approach to AGI's economic effects is scenario planning—starting from a premise (e.g. labor share is zero, or stays the same) and writing models of what could produce it—which both clarifies the space of outcomes and reveals what data we should collect, rather than asserting confident point forecasts.

normativehigh valuecontestednovelty 3/4durability 3/4· Alex Imas

“What is really useful is to think about the potential scenarios, map them out, and say what dimension of scarcity will generate each scenario.”

0.73

Nothing has yet been fully automated when you look at network-adjusted factor shares down the whole supply chain—e.g. US computer and electronic products have a stable network-adjusted capital share around 50%, not 100%—but the coming qualitative shift is some goods reaching a network-adjusted capital share of one, where the entire supply chain is automated with no human we intrinsically value.

factualhigh valuecontestednovelty 3/4durability 3/4· Phil Trammell

“Computer and electronic products in the US have a very stable network-adjusted capital share of around 50%. It’s not 100%. I do think there’s this qualitative shift that I think we agree is coming, which is that there will be at least some goods whose network-adjusted capital share goes to one.”

0.73

Ordinary people's main capital—a house, often only a part of one—derives its value from land being close to other humans and relational proximity, making it uniquely ill-suited to complement AI, robots, or the goods the wealthy will bid up, and implying a Georgist land tax would not raise enough for the needed programs.

factualhigh valuecontestednovelty 3/4durability 3/4· Dwarkesh Patel

“What is the value of a house currently? It’s that the land is close to other humans and modulo relational stuff that is just not going to be the main factor of production in the future.”

0.73

The Citrini scenario of AI causing negative economic growth requires implausible conditions: capital owners must hit a hard bound on demand (not just diminishing sensitivity), refuse to deploy their wealth as investment, and have income reallocated from workers to them—because amid genuine abundance and an expanding technological frontier, unlike the Depression, it is very hard to generate negative growth.

forecasthigh valuecontestednovelty 3/4durability 3/4· Alex Imas

“So you need demand to be bounded, like a hard bound, not even a soft diminishing sensitivity. You need for them to eventually say, “I’ve had enough. I don’t want to spend any more money.” And for that money to not enter as investment. Then you can get negative growth.”

0.73

Whether a job is 'relational' is a task-level empirical question: jobs are bundles of tasks (task-based model), and a job belongs to the relational sector if consumers have measurable willingness to pay extra for a human to remain in the loop on at least one task even when all others are automated—which requires conjoint-analysis data we don't yet have.

definitionhigh valuecontestednovelty 3/4durability 3/4· Alex Imas

“If the consumer is willing to pay more for a product or service where every single task is automated except for that one part where the doctor is delivering the diagnosis and providing support, we would call that job part of the relational sector.”

0.72

Commoditizing frontier AI involves a trade-off with safety: fewer frontier companies give each a buffer to slow down for safety, but a relatively large leader-laggard gap can coexist with widely distributed public ownership, so the benefit/risk tension is smaller than imagined; Dwarkesh argues commodification's diffusion is worth it partly because concentrated labs create a clear political target, as seen in the Defense Production Act threat against Anthropic.

normativehigh valuecontestednovelty 3/4durability 2/4· Phil Trammell

“For safety purposes, you might want fewer frontier companies so that each one has a buffer in case they want to slow things down to make things safer.”

0.71

An intrinsic preference for pure accumulation is rare; historically wealth accumulation is embedded in social interaction—seeking admiration among peers and community—and is followed by diminishing marginal utility once hedonic and status needs are met, a basic feature of preferences noted by Rousseau and St. Augustine; the stylized fact is titans accumulate capital and then spend it.

factualhigh valuecontestednovelty 2/4durability 3/4· Alex Imas

“a lot of the time, at least historically, the wealth accumulation process is part of a large social interaction amongst peers, amongst the community, where you want to be admired in some way. The stylized fact of titans of industry is you accumulate the capital, and then you buy a bunch of stuff.”

0.70

A von Neumann probe is a real possible entity that is an extremely greedy optimizer—high marginal value for each new solar system because it converts into more probes—but such self-replicating accumulation may not register in GDP, which counts only final consumption and investment goods, so whether labor share looks high may be an accounting artifact of how AI beings are counted.

forecasthigh valuefringenovelty 4/4durability 3/4· Phil Trammell

“we only count final consumption goods and investment goods as GDP.”

0.70

Fully automating the supply chain for everything outside the human-intrinsic sector has an ambiguous effect on overall capital share: if we satiate quickly in those goods, their quantity goes to infinity but marginal utility falls to zero faster, so spending share can still collapse toward the human-intrinsic sector.

causalhigh valuecontestednovelty 3/4durability 3/4· Phil Trammell

“if we fully automate the supply chains for everything else, and we satiate in everything else really fast, then the quantity of everything that’s not a ballerina goes to infinity, but the marginal utility in that stuff goes to zero faster than the quantity is rising.”

0.69

If open models stay only six to nine months behind the frontier, then within months of reaching AGI everyone gains access to the resource, which would make AI more like commoditized electricity and let even a country like Nigeria simply buy the index and gain AGI access via open models.

forecasthigh valuecontestednovelty 3/4durability 2/4· Alex Imas

“If we’re indeed in a world where the open models are six months behind the frontier—or nine months—then we’ll hit AGI, we’ll hit whatever, and in six months, everybody has access to this resource.”

0.69

Indexing the economy is already not that hard—well under 20% of the total market cap of non-tiny US companies is private—and even leading labs like OpenAI and Anthropic will likely go public before long, with AI itself plausibly alleviating disclosure frictions, so the long trend toward easier indexing should continue despite a recent uptick in privatized returns.

forecasthigh valuecontestednovelty 3/4durability 2/4· Phil Trammell

“well under 20% of the total market cap of non-tiny companies in the US is private... even they look like they’re going public before too long, probably.”

0.69

The 'messy middle'—where AI automates enough jobs to cause a political crisis but doesn't generate enough new wealth to compensate the displaced and produce a Pareto improvement—is possible but occupies a narrow window, because technology powerful enough to automate many jobs would normally be growing the pie fast; it requires AI to be only a hair more productive (and cheaper) than the labor it replaces, which conflicts with how broad intelligence generalizes across white-collar tasks.

forecasthigh valuecontestednovelty 3/4durability 2/4· Phil Trammell

“My guess is that if we have the technology to automate so many jobs that it becomes a new kind of political problem, then the pie will also be growing really fast.”

0.69

Experiments show people pay more for human-made art than AI-made art and that the human premium collapses when the human work is mass-produced (500 copies) while AI shows no such effect—suggesting humans provide intrinsic relational value through connection, unlike a horse that is merely a replaceable input valued only for its output.

factualhigh valuecontestednovelty 3/4durability 2/4· Alex Imas

“With one, you get the effect that the person-produced art print is valued much higher than the AI version. Then, in a set of other conditions, we say there’s 500 of these being produced. For the human-made one, the price goes down a lot because it’s no longer seen as making a connection with this one artist. With AI there’s no difference.”

0.68

Even without an intrinsic taste for accumulation, agents may have near-unsatiating demand for future wealth for instrumental reasons—arms races over political, philosophical, or religious influence—or for total utilitarian philanthropy aimed at creating new happy beings, echoing Bostrom's astronomical-waste point about Dyson spheres turning stellar energy into happy simulations.

causalhigh valuefringenovelty 3/4durability 3/4· Phil Trammell

“there’s just total utilitarian philanthropy. When I think about why it might be good to have a lot of wealth in the future as a good classical utilitarian... the value... is to create new happy beings.”

0.68

Individual economists' point forecasts about the AI labor market are not very useful because economists disagree in every direction and have historically been terrible at forecasting; aggregate mechanisms like prediction markets and wisdom-of-the-crowd are more reliable.

normativehigh valuecontestednovelty 2/4durability 3/4· Alex Imas

“rather than thinking about individual forecasts, we should be generating prediction markets where you get aggregate forecasts and wisdom-of-the-crowd effects. The reason I think this is because we have been famously terrible at forecasting.”

0.68

Labor share has stayed roughly constant at over 60% of the economy through the Industrial Revolution and all subsequent automation—a Kaldor fact so stable that some suspect it is an accounting artifact; Atkinson shows that holding accounting constant, labor share may not have fallen at all.

factualhigh valuecontestednovelty 2/4durability 3/4· Alex Imas

“This is a Kaldor fact. We should stress this. It’s incredibly surprising that it’s over 60% after the Industrial Revolution and all of the automation we’ve ever seen. Some people are worried it’s an accounting error that it’s been so constant.”

0.68

The negative narrative around AI dominates partly because it is psychologically much easier to imagine losing something that exists than to imagine a good thing that does not yet exist, so it is easier to warn that liked jobs will disappear than to vividly describe a future utopia.

causalhigh valuecontestednovelty 2/4durability 3/4· Alex Imas

“It’s more difficult to imagine a good thing that doesn’t exist than losing something that exists. It’s much easier for somebody to go on a podcast and say, “These jobs that you like, they’re going away,” than for somebody to spin up a utopia which doesn’t exist yet.”

0.68

Many professions will keep humans in the loop for regulatory and accountability reasons—licensing, the need for a fireable or liable legal entity, and political roles like judges and jurors—that have nothing to do with the human's actual ability to perform the service, though Trammell argues these frictions are transitional and will be out-competed once AI-run institutions are more efficient.

causalhigh valuecontestednovelty 2/4durability 3/4· Alex Imas

“There’s a lot of regulatory layers that are also going to be keeping—even if there’s no relational element—humans in the loop that have nothing to do with the ability of the human to actually perform the service.”

0.66

The relational sector—services and goods where a human being in the loop is itself part of the product's value—will remain scarce after broad automation, because humans are naturally scarce, so value will continue to accrue there even as other things stop being scarce.

forecasthigh valuecontestednovelty 3/4durability 3/4· Alex Imas

“Something like the relational sector, which is defined as services and goods where the fact that a human was in the loop is part of the value of that product. Because humans are naturally scarce, if we have automation where a lot of other things stop being scarce, we will still have scarcity in the things that humans are involved in and in the loop for.”

0.65

Moore's law can be read pessimistically as the value of computation halving every 18 months: we run out of new uses for computation so fast that falling value, not just falling cost, is what sustains the trend—so price reflects both cheaper transistors and a declining marginal value of each transistor.

definitionhigh valuecontestednovelty 3/4durability 2/4· Phil Trammell

“I like the pessimistic framing of Moore’s law: every 18 months, the value of computation halves. We’re running out of uses for computation so fast that it’s sustaining Moore’s law.”

0.64

A wealth or capital tax may lack a politically sustainable low equilibrium—as happened with income tax, which started low for wartime and escalated to marginal rates near 40-50%—and a capital tax risks distorting investment if investors fear the government will take and dilute ever-larger shares of companies.

forecasthigh valuecontestednovelty 2/4durability 3/4· Dwarkesh Patel

“One concern with the wealth tax is that there’s no politically sustainable equilibrium at a 0.5% wealth tax. This happened with the income tax, of course. It starts low, it’s for war or something, and then it slowly escalates.”

0.61

Redistribution design should separate three questions—how revenue is raised, what is taxed, and how it is distributed—so that, for example, a broad-based consumption or VAT-style tax could fund the government buying shares of AI companies and distributing them to everyone, which differs from directly redistributing the stocks and avoids populist expropriation of particular companies.

normativehigh valuecontestednovelty 2/4durability 3/4· Phil Trammell

“It’s worth separating how the revenue is raised, what’s taxed, and how it’s distributed. It could be that the government hands out shares of Anthropic to everyone by a broad-based tax and then buying Anthropic.”

0.60

Software appears to be a particular kind of good for which demand stays highly elastic—we keep wanting more as it gets cheaper—unlike goods such as food where satiation caps demand even in the long run; this is not an inherent property of all markets.

factualhigh valuecontestednovelty 2/4durability 2/4· Dwarkesh Patel

“The claim with software is that it is not some inherent property of markets that as it gets cheaper, you’ll just keep wanting more of it. The thing about software is this is a particular kind of good where as it gets cheaper, we’ll want more and more of it.”

0.59

Despite intense attention, there is currently no clear evidence of mass AI-driven automation or unemployment; even in the most exposed sectors like software engineering you have to squint to see anything, with at most junior developer hiring running slightly below trend and senior demand if anything rising.

factualhigh valuecontestednovelty 2/4durability 1/4· Alex Imas

“They just recently released a report, and you really have to squint to see anything happening... even looking at software engineering, the most exposed sectors, there’s just not really anything going on.”

0.55

For perhaps the first time the falling-value pattern may have reversed: an H100 costs more to rent now than three years ago despite superior technology and more total compute, because as models get smarter the opportunity cost of compute rises—an instance of increasing variety creating new high-value demand for capital.

factualhigh valuecontestednovelty 2/4durability 1/4· Dwarkesh Patel

“The famous fact here is that an H100 costs more to rent now than it did three years ago, even though we have much superior technology and much more compute in the world. Because as models get smarter, the opportunity cost of compute gets higher.”

0.53

The wealthiest people already exhibit non-satiating capital accumulation: figures like Zuckerberg and Musk could convert their stock to dividend consumption but instead let wealth compound into more data centers, fabs, or Mars infrastructure, showing an almost Nick Landian preference for accelerating capital that—without any change to human nature—can determine what the future economy produces.

factualhigh valuespeaker onlynovelty 3/4durability 2/4· Dwarkesh Patel

“most of his wealth is just stock in Meta. As a controlling shareholder, he could say, “Meta, turn all this wealth into dividend income...” Instead, he would rather have his wealth compound and have Meta build more data centers.”

0.52

A fully autonomous AI with its own welfare and decision-making has no inherent reason to prefer dealing with humans, so the relational sector's value depends entirely on whether human preferences for human connection persist among humans, not on AI demand.

forecasthigh valuespeaker onlynovelty 2/4durability 3/4· Alex Imas

“If there’s an AI that has its own welfare, is fully autonomous, and is making its own decisions that are welfare-relevant, to be honest, I have absolutely no prior that it would prefer to deal with humans. There’s no reason.”

0.48

Natural selection may strengthen rather than erode human preference for other humans: a person who feels a moral aversion (in Haidt's framework) to offloading social interaction onto AI is more likely to find a mate and reproduce than one indifferent between humans and AI, so given ongoing selection, human-relational preferences could intensify.

forecasthigh valuespeaker onlynovelty 3/4durability 2/4· Alex Imas

“The other one has almost a moral emotion—using Jonathan Haidt’s framework—against offloading those sorts of social interactions to an AI. Which of those two people are going to reproduce, find a mate, all of these sorts of things? I think the answer is clear.”

0.48

Layoff narratives can become public coordination devices in which firms lay off workers mainly to signal AI adoption and avoid looking behind the times, producing a keeping-up-with-the-Joneses cascade that can leave firms worse off than before the layoffs.

causalhigh valuespeaker onlynovelty 3/4durability 2/4· Alex Imas

“Then you’re going to just get a cascade effect of firms needing to keep up with the Joneses in terms of starting to lay people off. That’s super worrying, where the firm might actually be worse off after the layoffs than before.”

0.34

The economic questions about AGI are deeply interconnected: whether there are runaway concentrated gains connects to technical questions like recursive self-improvement, continual learning, and online learning, which in turn determine whether countries like Uganda will have any purchase on the returns of AGI.

factualcontestednovelty 1/4durability 2/4· Dwarkesh Patel

“every question is connected to every other. That question about whether there’s runaway gains connects to questions about recursive self-improvement... These are just forecasting technical questions which then impact whether Uganda will have any purchase on the returns of AGI.”

0.30

The proposal to privatize Social Security amounted to giving everybody a basket of stocks, similar in spirit to universal basic capital funded by a broad-based tax buying equities, and it has worked so far though its longevity is uncertain.

factualestablishednovelty 1/4durability 2/4· Alex Imas

“That was the proposal for Social Security, by the way. That was privatizing Social Security... Privatizing Social Security was basically giving everybody a basket of stocks.”