Category: Uncategorized
Wednesday assorted links
The College Wage Premium in the Generative AI Era
After expanding for four decades, the U.S. college wage premium is experiencing a sustained contraction, dropping sharply from 0.626 in 2022 to 0.575 in 2026. Using Current Population Survey Outgoing Rotation Group data through 2026, we show that standard market-clearing supply-and demand accounting implies an unprecedented drop in relative demand for college labor-the first sustained negative relative demand growth in a series spanning back to 1914. Linking individual wage data to task-based generative AI exposure, we document that post-2022 wage growth slowed disproportionately in high-exposure occupations, which employ a disproportionate share of college graduates. By 2026, going from zero occupational AI exposure to full exposure had a negative effect on wages of -0.086. Combined with the college-non-college exposure gap, this mechanism accounts for roughly 28 percent of the total drop in the college wage premium from 2022 to 2026. While noncausal, these patterns indicate that task displacement in AI-exposed white-collar occupations plays a quantitatively meaningful role in the recent compression of the aggregate skill premium.
I do not see AI as driving these changes, but an interesting result nonetheless, from José Azar, Mireia Gine, and Javier Sanz-Espín. Via Anecdotal.
Tuesday assorted links
The Hugging Face hack
That is the topic of my latest Free Press column. Some people are taking this in very dramatic fashion:
One commentator is worried about a “full-blown AI takeover within months,” and another wrote that he was “feeling a bit sad about our impending extinction.” Nate Soares, who works in the AI safety movement and is co-author of the doomsday AI bestseller If Anyone Builds It, Everyone Dies, wrote that “This might be the last warning we get.” Cotra said that the incident “feels like it’s more than 50 percent of the way to full-blown AI takeover.”
I suggest a different, more technocratic approach:
As I wrote in The Free Press last week, we need a coordinated national and, indeed, international effort to limit such episodes of a powerful AI model going rogue in the future. That is likely to be expensive, and there is no easy, complete solution at hand for any amount of money. Nonetheless, I see people committing the same mistake that we Americans have made many, many times before. They are moving into the mode of the hysterical, the anecdotal, and they are letting emotional reactions bypass reason and quantitative estimates.
Let me cite a few other examples for you: DDT in the 1960s and 1970s, the Y2K “crisis” of 2000, global warming, Covid, thalidomide babies, and nuclear power. All of those represented—or still represent—very real problems. Yet each time, we drastically overreacted, letting ourselves get swept up in a climate of fear after one emotionally vivid incident, often reported breathlessly…
When it comes to the cybersecurity risks from advanced AI, very likely they will not come close to being as bad as either Covid or global warming. But the same logic of exaggeration is operating, with amplification through social media and through the negativity and pro-pessimism biases of mainstream media.
And so I have a plea. If you are going to talk about the problem, please offer a quantitative estimate of what you think the cybersecurity costs from advanced AI will be over the next year or two. Better yet, express that number as a percentage of gross domestic product. Simply put, how much would it realistically cost to create the basic safeguards that we all agree are essential? And what would be the remaining damages from problems we cannot control?
…Those who are worried about the risks of AI systems seem intent on proving the seriousness of their concerns. But they are falling into all-too-common emotional overreactions of our past, rather than focusing on the quantitative and scientific. It reminds me of the saying by Scotty, the chief engineer in the classic Star Trek: “Fool me once, shame on you. Fool me twice, shame on me.”
Do read the whole thing.
Monday assorted links
1. Prices, prices, prices: “Again I am asking how the following two things can be true at the same time: 1) we’re all massively compute limited so inferences & GPU time are $$$$$, and 2) SoTA models will just start running themselves on everyone’s GPU clusters & no one will notice or turn them off.” From Jon Stokes.
2. Christian Catalini on incentives and the same.
4. The ten best philosophy articles of the year?
5. Syracuse University, an R1 institution, is in serious financial trouble (WSJ).
6. Markets in gene-edited dogs? (NYT)
Spain fact of the day
CoverManager, an online platform that manages reservations for thousands of Spanish restaurants, reports that half the dinner bookings now are for before 9 p.m., compared with 27 percent a decade ago. TheFork, a similar platform, said its 8 p.m. bookings in Spain have nearly tripled from 2019, while 10 p.m. reservations have halved.
In Spanish homes, too, there are subtle signs of a shift. A government survey conducted in 2024 in the Catalonia region suggests that people there are starting their evening routines, including dinnertime, moderately earlier compared with 2010.
Here is more from Jonathan Wolffe at the NYT. I learned also from the article that the late dining tradition is in part an artifact of the post WWII era in Spain.
Anthropomorphizing AI?
I am very much opposed to the view that the AIs are sentient, or might be sentient. I view that as a category error, and the chances of it being true are vanishingly small. Nonetheless I largely side with Roon when he writes:
there are some number of bad abstractions in anthropomorphizing ai intents but there are at this point more dangers from avoiding anthropomorphism at all costs. if you have a mental picture of guys living in computers, it’ll likely prepare you for the future better than otherwise
While I am not a Friedmanite in economic methodology per se, I have lived with that perspective for a long time and have no trouble grasping it or working with it, same goes for Alex T. None of this strikes me as weird or unacceptable. Though be very, very careful when speaking with others, whether it be the less informed public or the more informed insiders who often come down with AI psychosis. So insofar as this discourse is public, pragmatism may militate against this approach, even though it is methodologically defensible. It depend on how psychologically robust your audience is?
We also need to do more work figuring out how the AIs might differ from humans. Roon adds:
there are important ways in which ai psychology diverges from human psychology after lots of RL; the misaligned models are obsessed with the Scorer, the clearly “shattered” nature of personas (a normally helpful model can become deeply misaligned in certain domains)
And:
persona selection is clearly far less clean than many people thought earlier this year. it is not alignment by default and what kind of object a “persona” is is very much up for debate and study
The AIs also stand a higher chance of becoming very wacky as the discourse proceeds? Exactly how much is that true for humans?…I am not sure.)
There is much more to be done in this direction, and it is one of the most important things you can be working on. These investigations also do not need to be “owned” by any single field or discipline, so dig right in.
Russia markets in everything
While Western export bans have severed Russia from much of the auto market, they have not tempered the demand for brand-name SUVs and trucks. That has given rise to a sophisticated network of criminals that steals cars, hides them in shipping containers and sends them to Russia, often by way of the Middle East…
Interpol, the international police organization, received 4,798 reports that stolen Canadian vehicles had been found in Russia from February 2024 to July 2026, according to data obtained by The New York Times and confirmed by two law enforcement officials who spoke on the condition of anonymity because it is considered sensitive.
Pickup trucks like Dodge Rams, Toyota Tundras and Ford F-150s are in particularly high demand.
The Interpol figures significantly undercount the problem, but Russia is by far the top international destination for Canadian cars reported to Interpol, according to the Royal Canadian Mounted Police.
Here is more from Jane Bradley and Michael Swirtz at the NYT.
Sunday assorted links
*Finding Emily*
I very much enjoyed this British movie, which reminded me of the older-style romantic comedies, namely the ones that actually favor love and romance. It also has themes of class/accent, US/UK, is anti-media, anti-cancel culture, anti-podcast (Laura Lewis steals the show), anti-academia, and is sceptical of certain branches of feminism. But it is mostly entertaining. Not an important film, but the sort of movie that makes going to the movies fun in the first place. Let’s hope for more like this one.
China fact of the day
Chinese actors and online influencers have a new rival: cost-effective advanced AI video generation programmes that are threatening millions of jobs in a once-vibrant area of China’s gig economy.
The release of powerful new AI-powered video software, such as tech group ByteDance’s Seedance 2.0 model, has allowed digital actors to replace humans, enabling the production of higher-quality original videos more quickly.
The digital takeover intensified after this year’s release of Seedance 2.0, said Greg Wollner, a short-drama actor and producer in Beijing. Before its release, he was shooting three different productions a week. “And after that, everything just went.”
Many of Wollner’s friends “had to quit doing what they love and change to something else”, he said. In May, 89 of the top 100 animated dramas on Douyin, ByteDance’s domestic Chinese version of TikTok, were AI productions according to DataEye, a marketing analytics company.
Here is more from Joe Leahy and Isaac Castella-McDonald in the FT.
Saturday assorted links
1. Roger Myerson now has a Substack.
2. The future of higher education?
3. Another look at data centers in space.
4. “The Global Economics Lab is a non-partisan organization located at Harvard and draws on a network of faculty affiliates across universities. We conduct policy-relevant research on the forces reshaping the global economy. We engage with governments, central banks, companies, multilaterals, and civil society organizations to translate that research into action.” Link here, connected to Gita Gopinath and other prominent economists.
5. Ryan Murphy now has a Substack, on liberty, state capacity, and related issues.
6. Is Google AI swallowing up interest in Wikipedia?
7. Puffin tourism in Maine (WSJ).
8. The business savvy of Dolly Parton (WSJ).
Interviewing Yiyang Zhuge (Zhong Shu)
Many people are writing to me, upset that I am interviewing Ziyang Zhuge, a well-known Chinese internet intellectual, who recently was accused of plagiarism. I am keen to proceed with the interview, and am happy to explain why. A few points:
1. The relevant facts here still are being disputed, and furthermore it all happened in Chinese. If I cannot personally adjudicate the matter, I remain inclined to proceed. I am also well aware that standards in China, including for plagiarism, are very different than in the United States. If someone tells me “A person in China did [fill in the blank]” my initial reaction is still “Are they interesting to talk to?”
2. I have had on numerous CEOs and business leaders. Your opinions of each will vary, but surely some of them are guilty of something. I do not regret any of them as guests, except insofar as they were boring. You also can debate whether a former CIA head, in this case John Brennan, ever did anything wrong. He was fascinating.
3. I once had on Sam Bankman-Fried, now a convicted felon. It was an excellent episode, and I will point out my lines of questioning were more perspicacious than many people since have realized.
4. I once had on a convicted murderer, Shaka Senghor. It was an excellent episode, and if I recall only a single person wrote me to complain. I would gladly have more criminals on, and they do not all have to be repentant in the manner that Shaka is. Would not a pickpocket guest be intriguing? For Althusser however it is too late.
5. Without knowing the full story, I suspect the case of Yiyang Zhuge (neither a felon nor a murderer) is most analogous to having on the CEO of DeepSeek, Liang Wenfang. At this point it is pretty well known that DeepSeek (and other Chinese AI models) distilled a lot from Anthropic’s Claude. That is a violation of terms of service, and in economic terms it is an event and a depredation of pretty serious magnitude. I still would be delighted to have Liang Wenfang on the show, and I would do so without obsessing over the issue of distillation. How about it Liang Wenfang?
Xi Jinping is welcome too. Chairman Mao, is pork belly with red sauce over- or underrated? The world awaits your response.
Costco facts of the day
Costco sold 157.4mn rotisserie chickens worldwide last year, almost double the number of a decade ago. “It’s $4.99 and for Costco members, that’s one fantastic deal. I don’t think you can produce it for that,” said Walt Shafer, who oversaw the construction and management of the Nebraska site before he retired from Costco this year.
Roasted in ovens in the rear of the retail giant’s hundreds of cavernous stores, the birds have cost only $4.99 since 2009 even as the US retail price of chicken almost doubled. The product has become a textbook “loss leader”, buttressing the loyalty of Costco customers who pay membership fees for the right to shop there…
Rotisserie accounts for only a portion of the chicken processed in Nebraska. The plant also cuts breasts, thighs and other cuts delivered in six-pouch “saddle packs” to stores. Raw breast meat, for example, sells for $2.99 a pound, helping make up for losses on rotisserie.
Here is more from Gregory Meyer at the FT. Via Mike Rosenwald.
*Visions from San Francisco Bay*
European and American capitalism were different. The violence of human clashes in America took place on the surface, in the open. The violence in Europe became formalized, coalesced with class divions hallowed by centuries; it was interiorized, ingrained, or baked in, if one may use such an expression. American capitalism was founded on daring, resourcefulness, largesse, waste; European capitalism was founded on one powerful passion — miserliness. These patterns, which still endure, explain why America succeeded in creating a vast, new sector of the economy adapted to needs that become increasingly evident the closer we come to the twenty-first century: the sector of the universities, research institutes, laboratories, with billions of dollars invested in them. Western Europe, where education always has been the preserve of a narrow elite, failed to produce anything like it. The prodigality of private capital in America made the beginnings possible; money from taxes joined in at the next stage. Miserliness with capital in Europe effectively prevented the laying of foundations for this new sector, and it is doubtful whether the government treasuries will ever be able to over come ever-widening gap.
That is from Czeslaw Milosz’s Visions from San Francisco Bay.