NEW ESSAY: "Dispatches from India". A year ago I was living in New Delhi, India 🇮🇳. I traveled across 10 states and 30 cities+villages writing for The Economist, including a cover story on the economy. Since then I’ve been chewing on how to capture what I learned. It still

RT @laura_buhler: 🧠 25 economists studying transformative AI at the world's leading institutions .💙 2.5 inspiring days @stripe .💵 new gran….

RT @Afinetheorem: Excellent report (h/t marginal revolution) from Arjun Ramani on India. He's a PhD student at MIT but previously was The E….

RT @Noahpinion: Awesome.

RT @DuttaAnirudha: Should be very interesting.

My Economist special report on India: Subscribe for future writing:

Essay:

fantastic (and funny) read to help calibrate your AI timelines. the story of the fox, the hedgehog, and the rhino (guess who the rhino represents) was especially delightful!.

RT @ben_j_todd: Which tasks are going to take longest for AI to do? .4 categories:. 1. The data poor. The internet has trillions of words,….

Benchmarks/evals suffer from selection bias -- the tasks for which you can construct a usable benchmark are precisely the tasks RL is good for. An implication is we should EXPECT there to be a growing gap between benchmarks and real-world impact.

RT @lucy_rohden: protesting is cool

h/t @davidautor for teaching this paper in labor class!. paper here: and writeup by @ReviseNRetweet and coauthors here:

Worth noting Trade Adjustment Assistance was phased out in 2022 but @SenGaryPeters is trying to revive it. would love to see a trial of the wage insurance provision for other age groups. especially pertinent if rapid automation really does happen! 6/.

Caveats:.-doesn't lead to more industry-switching or movement.-impacts on employment probabilities fade after 4yrs.-if automation is sufficiently broad this idea won't work. that said the policy is limited to ages 50+. you might expect stronger effects for young workers. 5/.

The paper finds:.-shorter unemployment durations (8-17 pp increase in employment prob during the 2 yrs post job loss).-$18k more cumultative earnings in the 4 yrs post job loss.- the policy pays for itself due to reduced welfare payments/greater tax revenues (infinite MVPF) 4/

Enter wage insurance. The authors look at a special provision of US Trade Adjustment Assistance that applies to workers aged 50+. On top of the standard job training and extended UI, the provision pays up to half the diff between a worker's old and new wages upon re-employment 3/.

A major cost of job loss is "scarring". The longer a worker remains unemployed, the worse their job prospects are. This reduces long-term earnings. Sectoral shocks worsen this because a worker's outside options are thinner than usual (eg the China shock) 2/.

suppose @DarioAmodei is even partly right and an automation shock is imminent. how should policy respond?. one idea-partially restore a displaced worker's wages if they find a new job with lower pay. Hyman/Kovak/Leive find this leads to faster job-finding and pays for itself! 1/

RT @ben_golub: Report from the teaching trenches:. I teach an advanced elective (Social and economic networks) which is difficult for top u….