New Fields Medalist Jacob Tsimerman Joins OpenAI for AI Safety on the Same Day

Jacob Tsimerman, fresh off the stage after receiving the Fields Medal at the International Congress of Mathematicians in Philadelphia, announced at a press conference on July 23 that he will soon join OpenAI to work on AI safety.

The 38-year-old mathematician was born in Kazan, Russia, and moved with his family to Israel and then Canada. He won back-to-back gold medals at the International Mathematical Olympiad, with a perfect score in 2004. He entered the University of Toronto at 16, completed his undergraduate degree in two years, earned a PhD from Princeton under Peter Sarnak, and after a Harvard Society of Fellows stint became the youngest tenured full professor in the University of Toronto's mathematics department.

Tsimerman was honored for his proof of the André-Oort conjecture, a core problem in arithmetic geometry that had remained unsolved for nearly 40 years. He is the first mathematician to win the Fields Medal while employed at a Canadian university. This year's Fields cycle was unusually strong — three century-old problems were resolved in the same four-year period: the Kakeya conjecture (1917), the kinetic theory portion of Hilbert's sixth problem (1900), and the André-Oort conjecture.

In a striking paradox, Tsimerman has long been vocal about AI extinction risks. His only X post ever predicted that AI would surpass human mathematicians within two years. He co-authored a paper titled "A Taxonomy of Extinction-Futures Involving AI" that systematically categorizes scenarios in which AI could eliminate humanity. He has explained that while he supports pausing AI development, he believes prevention is unrealistic, so direct engagement is the more responsible path.

Tsimerman's move to OpenAI mirrors the trajectory of his former student, Levent Alpöge, now a mathematician at Anthropic. Just three days ago, Alpöge used Anthropic's Claude Fable 5 to disprove the Jacobian conjecture — a problem open for 87 years — in a single night. His counterexample was just 216 characters long and could be verified by hand. Fields Medalist Timothy Gowers called it the first time he had seen an LLM solve a long-standing problem outside its own domain.

Alpöge's undergraduate thesis advisor at Harvard was Tsimerman, then a Junior Fellow. Today, the former student works at Anthropic using AI to crack century-old math problems while his teacher joins OpenAI to ensure the technology stays safe. Teacher and student, one lab each.

The broader cohort of four Fields Medalists offers a glimpse into the AI industry's talent war. Beyond Tsimerman, Wang Hong (Kakeya conjecture) just secured a position at IHES in her academic prime; Deng Yu (Hilbert's sixth problem) works in mathematical physics far from the reasoning model frontier; and the youngest, John Pardon (symplectic topology), has the most potential but the least immediate overlap with AI. Analysts expect the list to shift as OpenAI, DeepMind, and Anthropic continue their mathematician arms race.