A $5 million gold card for immigrants?
That is the topic of my latest Bloomberg column, here is one excerpt:
As usual, however, the devil is in the details. There is a good chance Trump’s proposal could work out well — and a chance it could severely damage the nation.
One worry is selection effects. The $5 million fee means the program would skew toward older people, and would probably also skew somewhat male. Neither of those biases is a problem if other methods of establishing residence remain robust. But will they?
With a gold card program, the government would have a financial incentive to limit other ways of establishing residency. You can get an O-1 visa or an H-1B, for instance, if you have a strong record of accomplishment or an interested employer with a proper priority and perhaps some luck. Neither of those options cost anything close to $5 million, even with legal fees. Not everyone with a spare $5 million can get an O-1, or a proper job offer, but still: At the margin, these options would compete with each other.
These other options are well-suited for getting young, talented people into the US, which is precisely the weakness of the gold card proposal. Ideally the US would expand these other paths, but with a gold card program they might be narrowed so the government can reap more revenue from sales of gold cards.
I favor the idea in principle, but am worried it might be part of a broader package to tighten immigration more generally.
Matt Yglesias on morality
I do broadly align with utilitarian/consequentialist ideas, and I particularly like the formulation Richard Y Chappell calls Beneficentrism, which is simply the claim that one very important thing in life is to try to help others, including those who are very different or distant from ourselves.
I think that in that formulation, you address in one fell swoop 95 percent of what trips people up about utilitarianism. You can take special care of your friends and family. You can care more about citizens of your country than you care about people on the other side of the world. But you should care some about the general welfare. This is in fact pretty important and you should be doing something about it. How much? Probably more than you are doing. Probably more than I am doing.
Here is the full (gated) post, mostly about other matters.
Can Enhanced Street Lighting Improve Public Safety at Scale?
Street lighting is often believed to influence street crime, but most prior studies have examined small-scale interventions in limited areas. The effect of large-scale lighting enhancements on public safety remains uncertain. This study evaluates the impact of Philadelphia’s citywide rollout of enhanced street lighting, which began in August 2023. Over 10 months, 34,374 streetlights were upgraded across 13,275 street segments, converting roughly one-third of the city’s street segments to new LED fixtures that provide clearer and more even illumination. We assess the effect of these upgrades on total crime, violent crime, property crime, and nuisance crime. Results show a 15% decline in outdoor nighttime street crimes and a 21% reduction in outdoor nighttime gun violence following the streetlight upgrades. The upgrades may account for approximately 5% of the citywide reduction in gun violence during this period, or about one sixth of the 31% citywide decline. Qualitative data further suggests that residents’ perceptions of safety and neighborhood vitality improved following the installation of new streetlights. Our study demonstrates that large-scale streetlight upgrades can lead to significant reductions in crime rates across urban areas, supporting the use of energy-efficient LED lighting as a crime reduction strategy. These findings suggest that other cities should consider similar lighting interventions as part of their crime prevention efforts. Further research is needed to explore the impact of enhanced streetlight interventions on other types of crime and to determine whether the crime-reduction benefits are sustained when these upgrades are implemented across the entire City of Philadelphia for extended periods.
That is from a new paper by John M. Macdonald, et.al. Via the excellent Kevin Lewis.
Boris Spassky, RIP
In Leningrad’s embrace, midwinter’s chill, A prodigy was born with iron will. The chessboard’s call, a siren to his mind, Young Boris Spassky left his peers behind.
A crown he claimed in nineteen sixty-nine, Against Petrosian’s force, his star did shine. Yet Reykjavik’s cold winds would soon conspire, With Fischer’s challenge, stoking global fire.
The “Match of Century,” where East met West, Two minds engaged in psychological test. Though Spassky yielded, grace he did display, Applauding Fischer’s genius in the fray.
Beyond the board, his life took varied course, From Soviet roots to seeking new resource. In France he found a refuge, fresh terrain, Yet ties to Mother Russia would remain.
A “one-legged dissident,” some would declare, Not fully here nor there, a soul aware. Through Cold War’s tension, politics entwined, He stood apart, a free and thoughtful mind.
His games, a blend of strategy and art, Reflect the depth and courage of his heart. Now as we mourn his final checkmate’s fall, His legacy inspires players all.
Rest, Grandmaster, your battles now complete, Your journey etched where history and chess compete.
That is a tribute poem from GPT 4.5
The uneven spread of AI
This paper examines the spatial and temporal dynamics of artificial intelligence (AI) adoption in the United States, leveraging county-level data on AI-related job postings from 2014 to 2023. We document significant variation in AI intensity across counties with tech hubs like Santa Clara, CA, leading in adoption, but rapid growth occurring in unexpected, suburban, and remote-friendly areas such as Maries, MO, and Hughes, SD, particularly following the lockdown era. Controlling for county and year fixed effects, we find that higher shares of STEM degrees, labor market tightness, and patent activity are key drivers of AI adoption, while manufacturing intensity and turnover rates hinder growth. Our results point to the uneven distribution of AI’s economic benefits and the critical role of local education, innovation, and labor market dynamics in shaping adoption patterns. Furthermore, they suggest the potential of place-based policies to attract AI talent and investments, providing actionable insights for policymakers aiming to bridge regional disparities in AI-driven economic growth.
That is from a new paper by Eleftherios Andreadis,Manolis Chatzikonstantinou, Elena Kalotychou, Christodoulos Louca and Christos Makridis. Via the excellent Kevin Lewis.
Thursday assorted links
1. Paul McCartney will be publishing a history of Wings.
3. Dean Karlan resigns from US AID, and tells his side of the story.
4. AIs are turning some people sane.
5. David Perell + Dana Gioia, three hour conversation on writing.
6. Tax increase pending on Harvard?
7. New set of NBER economics papers on AI and economics.
8. “chinese cars cost less than renewing the registration of my car in the us” — China estimate of the day. And DeepSeek is spreading throughout China (FT).
The Trump Administration’s Attack on Science Will Backfire
The Trump administration is targeting universities for embracing racist and woke ideologies, but its aim is off. The problem is that the disciplines leading the woke charge—English, history, and sociology—don’t receive much government funding. So the administration is going after science funding, particularly the so-called “overhead” costs that support university research. This will backfire for four reasons.
First, the Trump administration appears to believe that reducing overhead payments will financially weaken the ideological forces in universities. But in reality, science overhead doesn’t support the humanities or social sciences in any meaningful way. The way universities are organized, science funding mostly stays within the College of Science to pay for lab space, research infrastructure, and scientific equipment. Cutting these funds won’t defund woke ideology—it will defund physics labs, biomedical research, and engineering departments. The humanities will remain relatively untouched.
Second, science funding supports valuable research, from combating antibiotic resistance to curing cancer to creating new materials. Not all funded projects are useful, but the returns to R&D are high. If we err it is in funding too little rather than too much. The US is a warfare-welfare state when it should be an innovation state. If you want to reform education, repeal the Biden student loan programs that tax mechanical engineers and subsidize drama majors.
Third, if government science funding subsidizes anyone, it’s American firms. Universities are the training grounds for engineers and scientists, most of whom go on to work for U.S. companies. Undermining science funding weakens this pipeline, ultimately harming American firms rather than striking a blow at wokeness. One of the biggest failings of the Biden administration were its attacks on America’s high-tech entrepreneurial firms. Why go after Apple, Google, Microsoft, and Meta when these are among our best, world-beating firms? But you know what other American sector is world-beating? American universities. The linkage is no accident.
Fourth, scientists are among the least woke academics. Most steer clear of activism, and many view leftist campus culture with skepticism. The STEM fields are highly meritocratic and reality-driven. By undermining science, the administration is weakening one of America’s leading meritocratic sectors. The long run implications of weakening meritocracy are not good. Solve for the equilibrium.
In short, going after science funding is a self-defeating strategy. If conservatives want to reform higher education, they need a smarter approach.
Hat tip: Connor.
What I’ve been reading
Alain Mabanckou, Dealing with the Dead. Most African fiction does not connect with me, and there is a tendency for the reviews to be untrustworthy. This “cemetery memoir,” from the Congo (via UCLA), connected with me and held my interest throughout.
Philip Zaleski and Carol Zaleski, The Fellowship: The Literary Lives of the Inklings: J.R.R. Tolkien, C.S. Lewis, Owen Barfield, Charles Williams. I was in the mood of thinking I don’t need to read another book about these people. Yet this one was so good it won me over nonetheless.
Eddie Huffman, Doc Watson: A Life in Music. A fun book about one of America’s greatest guitarists. Watson was blind from an early age, and he was collecting state disability benefits until he was 40 — a classic late bloomer.
Philip Freeman, In the Brewing Luminous: The Life and Music of Cecil Taylor. Call me crazy, but I think Sun Ra and Taylor are better and more important musically than say Duke Ellington. Freeman’s book is the first full-length biography of Taylor, and it is well-informed and properly appreciative. It induced me to buy another book by him. The evening I saw Taylor was one of the greatest of my life, I thank my mother for coming with me.
Carlos M.N. Eire, They Flew: A History of the Impossible. Ross Douthat recommended this one to me. It is well done, and worth reading, but I don’t find it shifted my priors on whether “impossible” events might have really happened.
I agree with the central arguments of Samir Varma’s The Science of Free Will: How Determinism Affects Everything from the Future of AI to Traffic to God to Bees. I was happy to write a foreword for the book.
Kathleen deLaski, Who Needs College Anymore? Imagining a Future Where Degrees Don’t Matter. One of a growing chorus of books suggesting higher education is on the verge of some radical changes.
There is Daniel Brook, The Einstein of Sex: Dr. Magnus Hirschfeld Visionary of Weimar Berlin. It is good to see him getting more attention.
There is also Brandy Schillace, The Intermediaries: A Weimar Story.
Nick Freiling asks
Across all of history, which human being has been seen, in-person, by the greatest number of people?
Pope John Paul II? Mick Jagger? A baseball player with extreme longevity? (Cal Ripken?) Here is o1 pro, top guess is PJPII.
Wednesday assorted links
2. Technicolor is finally being wound down.
3. Atlantic profile of Aella (salacious text, otherwise safe for work).
4. Show notes from Curt Jaimungal. And his Substack on physics.
5. Ed Leamer, RIP.
6. The budget package passed by the House.
7. Outmigration predicts AfD support pretty well (NYT). It is the less ambitious ones who stay behind who want that party.
8. Bezos Wapo Opinion page announcement in favor of free markets?
9. Loyal’s life extension drug for dogs may be coming to market this year.
Method in the Madness?
We study the link between exposure to anger and performance, something that to our knowledge has not been studied before. We exploit the US version of the nationally syndicated television cooking competition MasterChef and collect data on every instance in which judges objectively demonstrate anger to individual participants in each of the episodes in ten years, from 2010 to 2020, of this competition. Contestants exposed to anger from judges end up higher in the final rankings, increase their probability of success in cooking challenges, as well as their probability of reaching the top three and even winning the competition.
Clever paper from Alberto Chong. Now if only you donkeys would improve your comments.
Are different jobs becoming more alike?
An intriguing but speculative hypothesis:
We document, in the past decade, a reversal in the trend of increasing wage inequality and declining labor mobility patterns across occupations. This is despite a rapid increase in technological progress and increase in technical skills demanded as evident in online job advertisements. We argue that rapid technological progress over the past decade is transforming occupations and driving them closer together, leading to wage convergence and increased mobility. To study this, we introduce a novel measure of occupational distance based on job postings that allow us to represent each occupation as a vector of skills with corresponding weights that evolve over time and space. We compute bilateral distances between occupations using similarity measures and find a steady decline in these distances, indicating occupational convergence. We show that this convergence has led to a decline in wage inequality across occupations and increased labor mobility. Furthermore, we find that occupations that have undergone greater transformation-measured as self-distance over time-exhibit higher wage growth and larger reductions in within-occupation inequality.
That is a recent paper from Manolis Chatzikonstantinou, Mohammad Shahmeer Ahmad, and Alexis Antoniades. Via the excellent Kevin Lewis.
Paul Millerd on AI and writing
I have been thinking a lot about this. Have been experimenting like a madman for two months
A few unhinged thoughts: – It’s a huge advantage to have a past body of work and style in terms of fine-tuning and training. It can help you understand your own style and keep evolving that over time independent of LLMs
– Vibe Writing will be a thing in 6-12 months if not the next couple of months. LLM suggest edits => accept. The biggest thing stopping this is reliable output based on an input/preferred vibe. It is getting CLOSE
– Much of the friction of writing, like getting stuck on sections/sentences/phrasing is basically gone. You can just prompt your LLM coach for alternatives to unstuck yourself
– LLMs will empower existing authors with audiences. LLMs will enable you to build your own team of people. An LLM developmental editor (slower reasoning models), a parter co-writer LLM for remixing and rewriting sentences, an LLM copyeditor and proofreader and LLM translator – its still so early but these are coming
– This leads to the fact that LLMs will increase the speed of writing. The time to first draft can be dramatically shortened. You can now generate really good writing that is similar to yours as a first draft
– Right now we have less books because its hard to write a damn book. As this gets easier, we’ll see more books, shorter books, and more creative collaborations
– just like music has such a fast production and relase cycle (singles dropping randomly) I think authors go this wway too.
– The reading experience obviously will change. Kindle will likely ship AI features in the next 2-3 years that will help you understand characters, refactor books to your preferences, and instantly translate to different languages / audio. Of couse the traditional publishing dinosaurs will lose their minds of this.
– Theres a huge opportunity for more reading devices clear AND a big opportunity for new direct to reader distribution (long self-publishing lol) – Text to voice will be instant and cheap too meaning the divide between reading/listening gets fuzzier
Here is the link, responding to the thoughts of David Perell immediately below.
What should I ask Ian Leslie?
Yes, I will be doing a Conversation with him. I loved his forthcoming book John & Paul: A Love Story in Songs. Ian has done other things too, but for the time being those don’t matter. This will be a Beatles episode, and also an episode about artistic collaboration.
So what should I ask him?
It’s Happening!
Waymo factory in Phoenix, AZ. 2000 cars and counting. 
Hat tip: Brian Potter.