Philip A. Luck

We currently have four open roles on the team: - Research Engineer (job-boards.greenhouse.io/anthropic/jobs…) - Research Economist (job-boards.greenhouse.io/anthropic/jobs…) - Transformative AI Research Economist (job-boards.greenhouse.io/anthropic/jobs…) - Research Operations (job-boards.greenhouse.io/anthropic/jobs…) 6/6

The damage this does to U.S. credibility and moral standing will far outlast whatever short-term policy gain is imagined.

"H-1B households generate substantial positive fiscal balances at every level of government... The average H-1B household contributes $30,050 net annually — 2.6 times the $11,530 contribution of a typical U.S. household" eig.org/fiscal-impacts…

US spending on offices for computers has now surpassed US spending on offices for humans: bloomberg.com/news/articles/…

This is 100% completely unsustainable as a society. Nearly 50% of all consumer spending now comes from the top 10% of earners. The bottom 80%? Their share keeps falling. This is why the economy can look strong in the data while millions of people feel like they're falling behind.

"To Address Farm Labor Shortage, Trump Administration Turns to Migrant Workers" nytimes.com/2026/03/15/us/… "As the president’s immigration policies squeeze an already tight supply of farm labor, the Trump administration is making it cheaper to hire foreign farmworkers."

The real AI slop isn't written by AI. It's written about it. Please stop.

New Immigration Detention: Current Data and Trends #Immigration#ICEDetention #TRAC See TRAC’s latest Quick Facts at tracreports.org/immigration/qu…

We hiring a Program Coordinator/Program Manager. Learn more and apply: careers.csis.org/opportunities/…

85 people have paid the $100,000 H-1B fee so far, totaling $8.5 million in revenue. But fee revenue from H-1B apps abroad is down $28 million. So the fee — justified by a paper claiming the revenue-maximizing fee was >$100,000! — appears to have lost the government $20 million.

I have never been more jealous. Amazing analogy.

"US oil intensity is twice as high as the European Union and 40% higher than China’s... That makes it harder for American consumers to shield themselves from rising gasoline prices... Americans are already spending $187 million more a day on gasoline than they were a week ago"

I’m sure our future robot overlords will reward your loyalty. I am staying on two wheels. 🚲

China used to mostly license—and steal—IP from other countries. Now other countries are paying China billions of dollars to use Chinese IP. Another example of what I call the “great tech reversal.” Really interesting analysis by @Ryanfeath: csis.org/blogs/trustee-…

Under the radar: the annual population adjustment to the household survey is a much bigger deal than I expected. The adjustment both lowered the population and shifted the population older. LFPR and EPOP were adjusted down by .4 and .5 points. Big.

China talks a lot about boosting consumption, but the new budget is conservative. Fiscal impulse is small, consumer subsidies are modest, property sector support is tepid, growth in central transfers to local governments are lower than the projected GDP growth rate, and personal income growth is targeted in line with GDP. Hard to see a major consumption push from this policy mix.

Anthropic surged in popularity and downloads after the Pentagon blacklisted it in a dispute over how its Claude chatbot could be used in war. Many tech workers, including some at rival AI firms, sided with the company in the clash. wapo.st/4lby36J

Yikes. Almost every major industry group shed jobs in February. Private sector overall: -86,000 Hospitality -27,000 Healthcare -28,000 Manufacturing -12,000 Tranport/warehouse -11,000 Construction -11,000 Information -11,000 Federal gov't -10,000 Professional/biz -5,000 Mining -2,000 Social assistance +9,000 Finance +10,000 #jobs

Are we finally seeing AI in the productivity data? A big upward revision to earlier data and strong Q4 bring us 2.2% above CBO's pre-pandemic forecast. Annual rates: 1 year: 2.8% 2 years: 2.5% 6 years: 2.2%

I've decided to leave OpenAI. I'm incredibly proud of all the work I've been part of here, from helping create the reasoning paradigm with @MillionInt, scaling up test-time compute with @polynoamial, working on RL algorithms with my fellow strawberries, shipping o1-preview (which started life as of one of my derisking runs), to post-training o1 and o3 with @ericmitchellai, @yanndubs and many others. I'm most proud of having led the post-training team here for the last year -- the team has done incredible work and shipped some really smart models, including GPT-5, 5.1, 5.2, and 5.3-Codex. OpenAI has genuinely some of the most talented researchers I have ever met, and I have learned more than I could have imagined knowing since I joined as a new grad. I want to thank @markchen90 @FidjiSimo @sama @merettm for all their support over my time here, and too many collaborators to name for the insights, ideas, and just plain fun we have had working together. After leading post-training for a year, though, I'm longing to start fresh and return to IC research work. I've been thinking about going back to technical research for quite some time, and I genuinely believe my colleagues and team here are set up to succeed going forward without me. I'm personally very excited for my next chapter -- I'm proud to be joining @AnthropicAI to get back into the weeds in RL research, and I'm looking forward supporting my friends there at this important time. Many of people I most trust and respect have joined Anthropic over the last couple of years, and I'm excited to work with them again. I have also been very impressed with Anthropic's talent, research taste and values, and I'm excited to be part of what the company does next!