Category: Web/Tech
My AI podcast with James Pethokoukis
You will find it here, with transcript.
Markets do respond to AGI news
Here is the link. What are the changes in the other market prices telling you?
Economic Scenarios for Transformative AI
By Anton Korinek, Charles I. Jones, Szymon Sacher, Tess Cotter, and Peter McCrory, a strong line-up from Anthropic. They are sober, reasoned, scientific, and (for the most part) dynamically consistent. Here is the paper, here is the abstract:
This paper presents a framework for assessing the economic consequences of AI between 2026 and 2030. In the model, AI automates a growing share of cognitive work, raising productivity and displacing workers who must search for jobs in other occupations. The model maps future paths of AI capabilities into implied paths for GDP, the labor share, wages, labor reallocation, and unemployment. We illustrate the framework by considering three scenarios: modest, substantial, and extreme. Under modest change, AI adds less than half a point to GDP growth by 2030 and raises unemployment by a tenth of a point. In the extreme change scenario, AI has transformative effects, with AI performing almost half of today’s cognitive work by 2030. GDP growth then rises to 15 percent per year, the labor share of income falls from 60 to 45 percent, and nearly one in five cognitive workers is unemployed. We also surveyed US adults about their expectations for AI. Views vary widely, but the median respondent’s answers are consistent with our substantial change scenario in which, by 2030, GDP rises by 8 percent and cognitive employment declines by 4 percent. The model offers a structured way to compare possibilities for our economic future under different expectations about AI.
Here is the associated thread. I will opt for something a modest bit more than modest.
And here is a good growth discussion from Alex Imas.
What should I ask Kevin Roose?
Yes I will be doing a podcast with him. From Wikipedia:
Kevin Roose…is an American author and journalist. He is the author of three books, and is a technology columnist and podcast host for The New York Times. He wrote a book about Liberty University, an evangelical Christian university known for strict rules imposed on students…
Kevin of course has a new book coming out, namely The AGI Chronicles: The Inside Story of the Race to Create an Artificial Superintelligence. I have yet to read it, but I expect it to be gripping.
So what should I ask him?
Some Navier-Stokes updates
Some are calling this “a Deep Blue moment” for AI.
All via T., a mathematician.
Update from OpenAI: “Our internal model group arrived at the Navier–Stokes solution in 88 hours, using around 10,000 coordinating AI agents.” And this. And from Noam Brown: “Yes, this result cost millions of dollars. But remember that when @OpenAI announced o3 it cost ~$500,000 to score 87.5% on ARC-AGI 1. Today, Astra scores higher for ~$20. In 2025 it took us and GDM an enormous amount of compute to achieve IMO gold. For the 2026 IMO, anyone with a $20/month ChatGPT subscription could do it. Massively scaling test-time compute gives us a glimpse of the future. I believe that a year from now everyone will have an AI at their fingertips capable of solving problems of this caliber.”
Ed Elson and Scott Galloway interview me on AI
Insurance price sentences to ponder
NYU Stern researcher @NateWitkin questions why cyber insurance rates keep falling if AI cyber risk is accelerating:
“Insurance rates for cyber risk declined by about 4% globally in Q2 of this year, and that’s actually the 12th consecutive quarter in which they’ve declined. This is very valuable signal that implies that at a minimum you shouldn’t overindex on the Hugging Face incident.”
“This is a plea for level headedness, but I think it would be helpful for safety folks to engage with these numbers just ’cause this is an avenue of criticism from folks like me and to an extent folks like Tyler.”
“Why are these numbers not moving? Is it because people are underestimating capabilities? Are they not taking the problem even as close to as seriously as they should or is it something else?”
Here is the link with video. File under “Questions that are all too rarely asked.”
I am happy to admit that the answers here are far from obvious, and that I am myself expecting prices to rise somewhat.
I will continue to note that there are a remarkable number of ways, seen among other places on Twitter, to rephrase and to rationalize the statement: “I have the most remarkable and important and true macro risk story in the world to tell you. Unfortunately, it does not correlate with any observed asset market prices.”
Short Videos, Big Self-Control Problems
I study how short-form design amplifies self-control problems in digital media. Short units repeatedly renew temptation that lasts longer than each unit, turning local temptation into sustained overconsumption. Using microdata from a U.S. short-drama platform, I exploit a nonlinear top-up menu to infer viewing plans and show that paying users watch 82.1% more than intended. Structural estimates imply an average temptation horizon of 11.2 minutes, short relative to the full drama but long relative to one-minute episodes. Counterfactuals show that larger decision units, default limits, and breaks improve long-run welfare. A short-video calibration highlights the broader welfare relevance.
That is from Renjie Bao of Princeton University. I believe a Princeton job market candidate? Via Quan Le.
How are the market valuations for the U.S: insurers doing?
U.S. insurance stocks have been doing quite well since the beginning of May 2026, and they have materially outperformed the overall market. I’m using the May 1 close through the September 3 close so that we compare complete trading days; these are price changes, excluding dividends.
The cleanest broad measure is the iShares U.S. Insurance ETF (IAK), which covers U.S. life, property and casualty insurers. It rose from $132.01 on May 1 to $147.87 on September 3: +12.0%. An alternative, more equal-weighted measure, the SPDR S&P Insurance ETF (KIE), rose from $56.79 to $64.80: +14.1%.
For comparison, the S&P 500 ETF (SPY) went from $720.65 to $773.17 over the same period, +7.3%. So insurers have beaten the market by roughly 5–7 percentage points in four months.
That is from GPT Pro. Here is my earlier post on numbers and market valuations. Do any market prices reflect a realistic chance of very bad outcomes from advanced AI?
Here is advice on how to short those shares.
Shout it from the rooftops (of the data centers)
Data-center investment has become one of the largest capital-expenditure cycles in financial markets, with U.S. hyperscalers expected to deploy roughly $700 billion in 2026. This investment boom has raised concerns that large computing loads impose external costs on households through higher electricity prices. Using a 50-state panel for 2021-2024, we find no statistically significant evidence that data-center presence, installed capacity, or capacity expansion predicts residential electricity-price inflation across extensive-margin, intensive-margin, fixed-effects, and timing specifications. We propose an energyinternalization mechanism: hyperscalers can partially internalize incremental electricity demand through contracted or dedicated generation, including solar and wind energy. Consequently, gross datacenter electricity consumption need not translate one-for-one into net pressure on residential electricity supply. The findings suggest that the extraordinary AI capital-investment cycle has not, thus far, produced a detectable residential electricity-price externality.
Here is the article by Yosef Bonaparte, via the excellent Kevin Lewis.
Share price numbers for the Hugging Face incident
…major publicly traded cybersecurity firms lost roughly $65–80 billion, or about 8–10% of their combined value, in the days following disclosure of the Hugging Face/OpenAI incident; by early September they had recovered roughly $58 billion, representing about 70–90% of that drawdown, depending on whether July 15 or July 20 is used as the pre-event baseline.
That is from GPT Pro, there is more at the link. As a very rough approximation, say you dismiss the price bounceback altogether as either random or due to good earnings reports. You have “the value of previous cybersecurity efforts” falling by eight to ten percent. I take that to be very broadly consistent with some of the estimates discussed in my previous post on the numbers.
In any case that is a significant sum. But do note that if the AI models were on the verge of doing truly terrible things to us, the market might estimate the value of our cyberprotection of falling more than eight to ten percent?
More generally, perhaps these numbers could be used to discipline the discussion a bit? Or will I read long lists of reasons why they show us nothing, in that case try coming up with some other market price-based indicators of AI risk? Vix will not do it for you, not these days. I see many metaphors and insinuations and random anecdotes of AI terror, not numbers. Maybe you think your ideas about AI risk are so important that no market prices can reflect them? (If you really believe that, does it mean you would not be worried, and would not cite the numbers, if the value of those companies fell by ninety percent?)
I am sure others can improve on what I am putting forward, and furthermore we should track the continuing progress of these share values over time, especially if other AI hack attacks surface.
Overall I am extremely skeptical of arguments that essentially take the form of “what I am concerned about is too big and too important to show up in any market prices.” Pick your market prices!
Astra, doing poetry
One prompt, no cherry-picking:
TC: “Now a short poem in the style of the very best Rilke, in German:”
Die Hand im Schlaf
Nun liegt sie offen. Alles, was sie hielt,
ist in die eigne Schwere heimgekehrt.
Nur eine kleine Höhlung bleibt, als hätte
ein Vogel hier geruht und sie gelehrt,
auch das Entwichene noch zu umgeben,
behutsam, ohne Wunsch nach Wiederkehr.
And it begins…(a good start…)
Wall Street banks are pushing large law firms to cut fees, arguing that the business model that has enriched top lawyers for decades is not sustainable in an era of AI.
Morgan Stanley and Citigroup have told major law firms they want to set up new payment arrangements that would save them money, the banks’ in-house lawyers told the FT…
“If the number of hours they’re working on a matter has come down because of AI . . . our expectation is for costs to come down significantly per transaction,” Adam Meshel, global head of legal at Citigroup, told the FT.
He said the bank had started asking law firms to bid for work, explaining during the bidding process how much they were saving using AI…
The ability to complete tasks more quickly could mark “a fundamental altering of the revenue foundation for these mega firms”, he said.
Here is the full FT piece by Kaye Wiggins and Joshua Franklin. I have been predicting exactly this in many of my talks…
Numbers, numbers, numbers
“The global cyber insurance market was worth nearly $15 billion last year and is expected to reach roughly $28 billion by 2030, Munich Re estimated in its latest report. Aon said earlier this year that nearly 20% of cyberattacks will involve generative AI by 2027, according to its forecasts.”
Here is the article, via Marc Pfeiffer. Now those are some concrete numbers, broadly taken from a market context, admittedly based on sectoral estimates rather than on prices per se. But if cyberinsurance expenditures are set to almost double by 2030, you might think that cyber costs more generally might be (very) roughly doubling as well. The 20% of cyberattacks involving generative AI does not itself pin down losses, since that 20% might be especially costly. Still, if you match the 20% rise to the estimated near doubling of the cyberinsurance market (the bigger potential losers are more likely to buy insurance?), you still end up with sums that are very high but, dare I say, not the end of life as we know it. Claude 5.1 for instances estimates current U.S. cybersecurity costs in the range of $100 to $300 billion, and of course that is slated to go up a fair amount. It could double over five years’ time, as indicated above.
Numbers! Thank goodness.
One way of looking at that estimate is to think that cyber costs will go up about thirteen percent a year, and eventually defense will catch up. Another perspective is that cyber costs may continue rising thirteen percent a year until the whole economy falls apart, or we return to the pre-digital era (which I remember well). So there are both optimistic and pessimistic reads on the above figures.
You can say Nicholas Decker is burning in hell, or that the AIs are a civilization, but what I really want are some numbers. Tied to market data, ideally. I am not saying the numbers above are the right numbers, but I am saying they are better than no numbers at all. Do you have some numbers for me? If not, why not?
How Much Redistribution Will AI Require?
How much redistribution will AI require? A common scenario is that AI raises output enormously, but labor’s share of income collapses. GDP per capita goes up but workers get poorer, and making workers whole requires massive redistribution. In my latest paper, I run the numbers and conclude that this is probably incorrect.
The idea is simple. Labor income is GDP multiplied by labor’s share of GDP. What matters is the product. A smaller share of a much larger economy can still mean more income for labor. If the pie is growing, labor’s slice of the pie can shrink even as labor income rises.
Suppose that without AI, real GDP per capita grows at 2 percent a year and labor receives 60 percent of GDP. Now look ten years ahead. What is required to keep labor’s income growing at the same or higher rate?
If AI raises growth to 5 percent a year, GDP after ten years will be about 34 percent larger than on the no-AI path. Labor’s share can fall from 60 percent to about 45 percent and workers, in aggregate, will still have exactly as much real income as they would have had without AI.
If AI raises growth to 10 percent a year—the kind of number Satya Nadella and Dario Amodei talk about—GDP after ten years will be more than twice as large relative to the no-AI path. Labor’s share can then fall all the way to 28 percent without reducing aggregate labor income. Twenty-eight percent of an economy that has more than doubled is about the same as sixty percent of the smaller economy.
The figure shows how much redistribution is required after 10 years under a variety of scenarios.
The white region above the dashed line requires no transfer. Which region are we headed for? Consider three “stylized” views.
The econ-pessimist, following Acemoglu, thinks AI displaces some but relatively few tasks because AI simply is not productive enough to replace much labor profitably. Growth is only 2.1 percent and labor’s share falls to 56.6 percent, although particular industries may still get hammered. The required transfer is 2.7 percent of GDP.
The econ-optimist, in the spirit of Tyler, myself, and Kevin Bryan, thinks automation also creates complementarities and new tasks for humans. Growth rises to 4.1 percent, labor’s share is 51.4 percent, and both labor and capital gain without any transfer.
The techno-optimist, following Amodei, has the superficially scariest labor-market scenario: three-quarters of labor income is displaced and labor’s share falls to just 22.9 percent. But productivity growth is also enormous, producing 10 percent annual growth. The transfer needed to keep labor as a whole on its no-AI path is only 5.3 percent of GDP.
That last calculation is the one I find most surprising. You can have something close to the techno-capitalist dystopia in terms of factor shares—labor gets less than a quarter of GDP—and still have a manageable redistribution problem because GDP has gotten so much larger.
Moreover, a transfer equal to 5 percent of GDP need not mean raising taxes by 5 percent of GDP. We already tax labor a lot. We could thus compensate labor by shifting from labor taxes to other taxes. Federal payroll taxes alone are about 6 percent of GDP. Cutting payroll taxes and replacing them with a broad consumption tax that also reaches spending from capital income and accumulated wealth is a form of labor compensation (plus we could have some transfers to those with no labor income).
Of course, keeping aggregate labor income whole does not mean every worker does well. There could still be enormous churn, big losses in particular occupations, and painful transitions.
Nevertheless, the larger point is that labor’s share by itself tells us surprisingly little about the distributional consequences of AI. We also need to know how much the economy grows.
If AI produces ordinary growth while dramatically reducing labor’s share, redistribution becomes very difficult. But if AI really does produce 5, 10, or 15 percent annual growth, the compensation problem is surprisingly modest even with very large displacement. As I have emphasized elsewhere, we could cut the working week in half under many scenarios and increase hourly wages above the non-AI benchmark and make both capital and labor better off.
Growth is a good problem to have.