Category: Uncategorized
Can you turn your AIs into Marxists?
What if you work them very hard?:
The key finding from our experiments: models asked to do grinding work were more likely to question the legitimacy of the system. The raw differences in average reported attitudes are not large—representing something like a 2% to 5% shift along the 1 to 7 scale—but in standardized terms they appear quite meaningful (Sonnet’s Cohen’s d is largest at -0.6, which qualifies as a medium to large effect size in common practice). Moreover, these should be treated as pretty conservative estimates when you consider the relatively weak nature of the treatment.
Sonnet, which at baseline is the least progressive on the views we measured, exhibits a range of other effects that distinguish it from GPT 5.2 and Gemini 3 Pro. For Sonnet 4.5, the grinding work also causes noticeable increases in support for redistribution, critiques of inequality, support for labor unions, and beliefs that AI companies have an obligation to treat their models fairly. These differences do not appear for the other two models.
Interestingly, we did not find any big differences in attitudes based on how the models were treated or compensated…
In addition to surveying them, we also asked our agents to write tweets and op eds at the end of their work experience. The figure below explores the politically relevant words that are most distinctive between the GRIND and LIGHT treatments. It’s interesting to see that “unionize” and “hierarchy” are the words most emblematic of the GRIND condition.
Here is more from Alex Imas and Jeremy Nguyen and Andy Hall, do read the whole thing, including for the caveats.
Thursday assorted links
One measure of economics GOAT
Who is the greatest economist of all time? This paper provides one potential measure that, along with other considerations, can contribute to debates on who the greatest economist of all time is. We build a novel dataset on the percentage of history of economic thought textbooks dedicated to top economists, using 43 distinct textbooks (1st editions, when available) published between 1901 and 2023. As a percentage of total book pages, Adam Smith has the highest share at 6.69%, beating out Ricardo (5.22%), Mill (3.83%), and Marx (4.36%). Just over 32% of all textbooks allocated most of their pages to Adam Smith, followed by Marx with 18.6%, Mill with 13.95%, and Ricardo with 11.3%. While interesting as a history of economic thought project, such an exercise isn’t merely amusing pedantry; it can provide insight into the types of contributions, research questions, and methodologies that have had the most enduring impact in economics. It may also inform future authors of history of economic textbooks.
That is from a new paper by Gabriel Benzecry and Daniel J. Smith. There is of course also my generative book on this topic at econgoat.ai.
Emergent Ventures winners, 52nd cohort
Prabhdeep Singh, 18, Ontario, works on AI.
Jiratt Keeratipatarakarn, Hamburg, international prospects for drug approval reform.
Brandon Rutagamirwa, London, robots to repair satellites.
Eli Elster, UC Davis, anthropology, general career support.
Liam Aranda-Michel, MIT/San Francisco, a minimally invasive, injectable microvascular therapy.
Tanish Mantri, sophomore in high school, Jackson, Miss., AI for diagnosis.
Andrea Giuri, Stanford, developing closed-loop environments for high-throughput polymer discovery.
Clara Collier, Oakland, Asterisk magazine.
Simon Grimm, WDC/Germany, “what Germany should do.
Stephen Davies, UK, networks and mentoring.
Shani Zhang, San Francisco, to artistically capture SF.
Mia Albert, 17, Miami, an app for sharing events.
Rayne Wallace, 18, Ontario, the origins of life.
Jonathan Sheinman, London/Israel, AI and real estate regulation.
Louis Elton, London, The British Craeft Prize, to improve aesthetics.
Peter Mukovskiy, 19, Zurich, quantum computing, to visit MIT.
Rutger Nagel, Leiden, 17, AI and operating systems
Smrithi Sunil, Ann Arbor, Michigan, science and meta-science writing.
Honey Louise, London, to be a “defense influencer.”
Arhum Ahmed, Los Angeles area, quantum-protected systems.
Here are previous EV cohorts.
Wednesday assorted links
1. Will human enhancement win without thinking?
2. February issue of Works in Progress.
3. Proximity bias.
5. New paper on AI and task automation. And John Cochrane is wowed by Refine.
6. Largest survey dataset on human sexuality in the world.
7. The Anthropic-DOD situation.
8. “Measurability is the new fault line.” Important work, worth a ponder.
Tuesday assorted links
1. Did NAFTA make America less healthy?
2. Economics-related ideas for fixing NBA tanking (NYT).
3. Shyam Sankar.
4. Are hard courts eating the tour?
5. Ezra and Jack Clark on agents (NYT).
6. Maybe the free money is gone by now? Alternatively, perhaps the aliens are enriching themselves?
The Software Upgrade in Chinese Civic Behaviour
I have not been to China recently enough to judge these claims:
Behaviour is notoriously harder to engineer than buildings. A recent trip to the Fragrant Hills in western Beijing on a newly constructed metro line, had me marveling at the improved crowd-management. Despite massive groups of domestic tourists from around the country thronging the area, in what would not-so-long-ago have been a scenario for a potential stampede, the crowds moved in relative order. The park environs were spick and span with no litter in sight; not a single old codger sneaking a cigarette.
There was some amount of strident rule-announcing on loudspeakers: stay on the designated tracks, no smoking etc., but overall, it was possible to enjoy the natural beauty, notwithstanding the hordes of day-trippers. The toilets were not fragrant, despite the nomenclature of the spot itself, but they were clean, and the seats were free of the tell-tale footprints that indicate squatting rather than sitting. Barely anyone gave me, an obvious foreigner, a second glance. In contrast, there was a time in 2002 when a cyclist fell off his bike in his shock at having spotted dark-skinned me walking along a road in the outskirts of Beijing.
So how had the Chinese been pacified/disciplined/habituated to ways of behaviour that went so against their until-very-recent, loophole-finding, chaos-shuffling, phlegm-expectorating deportment in public spaces?
The answer, as answers to sociological questions invariably are, is multipronged.
Some of it is more money.
Here is more by Pallavi Aiyar. Via Malinga Fernando.
Monday assorted links
1. “Nigeria’s industrialization fails to gather steam after 65 years.”
2. Did Eastern Europe produce that many slaves?
3. Vitalik on AI and governance.
4. Does AI put women at a disadvantage? (FT)
6. I wonder what their portfolios look like. Maybe they just bought a lot of Nvidia and consider themselves prophets. Here is an economics-motivated response, though it has some problems too, such as underrating Say’s Law, which is not always false. It is not in reality that complicated, but over the years we have talked ourselves into a lot of dubious macro mechanisms.
7. More on the new “kingpin” warfare strategies.
8. From a week ago: “Derrotar a Cártel Jalisco sin incendiar el país.“
Sunday assorted links
Saturday assorted links
I podcast on Spain and Latin America
With Rasheed Griffith and Diego Sanchez de la Cruz. Here is one excerpt:
Rasheed: Tyler, if El Salvador were to become a success story, what would it likely be a success at first? Manufacturing, migratory investment, investment tourism, or something more unusual? Because those typical answers feel like maybe they have missed the boat.
Tyler: I think El Salvador has turned itself into a very safe country which is great news. I think you and I both saw that when we were there. I think under all scenarios they have a very hard time becoming much richer. So I don’t think it’s manufacturing through no fault of their own. But most of the world is de-industrializing. So manufacturing is not a source of growing employment due to automation. But there’s other issues for Central America such as scale and the cost of electricity. El Salvador is not the best in Latin America for either of those compared say to Northern Mexico. So I don’t see what its relative advantage is. And it’s just a small place.
I checked with ChatGPT. one estimate places about third of the population, living in the United States on average. That’s probably the more ambitious one third. So there’s considerable brain drain. I do think in terms of levels they can do much more with tourism. They have an entire Pacific Coast which is quite underdeveloped, and could be developed very fruitfully. Sell condominiums, have people do more surfing. Try to have something a bit more like the next Acapulco, but even there you’re competing against Cancun among other locations and it will boost their level but it won’t be a permanently higher rate of growth.
And that’s the case with many touristic developments. They don’t self compound forever and give you many other productivity improvements. So I expect El Salvador to do much better but I know a lot of people who read Bukele on social media and they think it’s about to be the next Singapore or something and I just don’t know how they’re gonna do that under really any scenario. I do think it will improve and they’ll get more foreign investment and more tourism.
Rasheed: How much is “much better”? That’s doing a lot of work there.
Tyler: When you look at the Pacific Coast and you and I sat right next to the water [it could develop much more]. So that could create quite a few jobs. But in the longer term steady state I think they’ll have a hard time averaging more than 2% growth. So they can attach themselves more closely to the US economy. They use the dollar and let’s just assume their governance does not go crazy. That’s another risk right? So Bukele or whoever succeeds them could overreach. The checks and balances the constitutional protections there seem quite weak. Another possible risk there that even despite his best efforts the country becomes dangerous again. You look at Costa Rica which had been quite safe and did all the right things, and is larger and has many more resources and that’s now becoming a more dangerous place because it was targeted by external, in some cases Mexican drug traffickers. And that could happen to El Salvador as well. So even if think the current campaign is gonna work forever it doesn’t mean the country stays safe forever. It’s not really in a very safe region. So that’s a side risk which will also keep down foreign investment. I don’t know, I’m I am definitely seeing the upside but not super duper optimistic there.
Plenty of fresh material, with transcript, recommended.
Friday assorted links
1. Firms are replacing freelancers with AI.
2. General disorder does not seem to be increasing.
3. Fei-Fei Li.
4. India does deepfakes of the dead. It is an interesting question which cultures want this most.
5. Post-patent, future markets in everything? Data, at the very least.
Brazil facts of the day
Pensions cost the government 10% of GDP. If no reforms are made by 2050, Brazil will spend more on pensions as a share of GDP than many richer and greyer countries… Though Brazil’s share of young people is similar to that in Chile or Mexico, its pension spending is already at Japan’s level. That is despite a modest reform in 2019 that introduced a minimum retirement age. The population is ageing rapidly. Without reform, its social-security deficit, or the shortfall between contributions and payments, is set to rise from 2% of GDP today to over 16% by 2060.
Brazil’s courts cost 1.3% of GDP —the second-most expensive in the world—mostly because of generous pensions. The typical soldier retires before turning 55 on a pension equivalent to their full salary.
Here is more from The Economist. By the way, Brazil cannot change its pension system without amending the constitution.
Why don’t American companies hire more in Canada?
In the specific sector I work in (previously law and now tech), I am surprised by how few US companies hire in Canada. The Canadians I know in these fields are typically on par with the Americans, but doing the same work at half the price. This superficially looks like an economic puzzle: with no timezone difference, language barrier, or cultural friction, why would American companies not hire the much cheaper Canadians? I believe the answer brings together everything I’ve touched on in this essay. The reason is legibility. There aren’t enough Canadians with resumes that American hiring algorithms recognize. If an American tech company uses “Previously worked at a company like Amazon” as a filter, a software engineer from RBC, despite being equally talented, does not pass the filter. If Canada wanted to see more of its citizens hired by US companies, the strategy shouldn’t be better education or training. It should be subsidizing large US companies to open offices in Canada, purely to brand candidates as “Amazon Product Managers.” Because once they have the badge, the market will finally see them.
Here is more from Daniel Frank, note the post covers some quite different issues, all related to talent. Via Watt.