Daily Wisdom

On November 2, 2026, Anthropic opened a program that pays up to $500,000 a year and doesn't ask if you know machine learning. It's called the Anthropic Residency. A paid, fixed term research program for academics, technical founders and staff level engineers who want to make AI safe. Placements run from six weeks to six months. The lab building some of the most advanced AI on the planet is now recruiting people who have never trained a model. The program was first laid out on Anthropic's alignment blog on October 30. The reasoning was simple. Machine learning is a young field. The skills that make great safety researchers already exist in senior people across quantitative fields. What they lack is time inside frontier AI. So Anthropic is giving them that time. Residents get dedicated self study hours, talks and guidance to get up to speed on modern ML and AI safety research. They work directly with frontier models and get their own compute budget for their projects. The pay matches the seniority. In the US, $350,000 to $500,000 a year, pro rated for the length of the placement. In London, £265,000 to £375,000. The mentors are the people running the work. Ethan Perez leads Alignment Science. Evan Hubinger leads Alignment Stress Testing and came from MIRI. Sara Price leads Alignment Training, and her background is in data science, not machine learning. That last detail is the whole thesis in one person. The Residency grows out of the Anthropic Fellows Program, which proved new researchers could contribute fast. Over 80% of past fellows produced papers. One fellowship project had AI agents uncover $4.6 million in smart contract exploits. Then came the part almost nobody is talking about. Anthropic's stated goal is for every Resident to finish ready to work on AI research full time, at Anthropic or elsewhere. Elsewhere. A frontier lab is spending its own compute and its own senior researchers to train people who might walk across the street. That is not a hiring funnel. That is a bet that the field has too few hands, and that whoever adds them first shapes where it goes. It has happened before. After World War II, physicists poured into biology. Francis Crick trained in physics, switched fields in 1947, and six years later helped work out the structure of DNA. Molecular biology was built by people who arrived without the expected background. Wall Street did the same in the 1980s, hiring physicists and mathematicians who had never priced a bond. They became the quants, and within a decade they ran the trading floors. Applications close November 22. Offers go out in January 2027. Finance got its quants from physics. AI safety is getting its researchers from everywhere else.