Fields medalist Duminil-Copin on OpenAI's 700-plus math papers: "I cried"; just 30 of them hold about 4,000 pages tied to his work, as much as he has written in his whole career

Why it matters

【Key points】

Hugo Duminil-Copin, a 2022 Fields medalist (Geneva and IHES), spoke on October 8 at an online meeting of the French Mathematical Society and on Swiss broadcaster RTS's Forum about the 700-plus math manuscripts OpenAI released on October 6. On RTS he called it "a monumental shock" and said "in a sense, we are the day after"; mathematicians must reinvent their work. He said he cried, thinking of young researchers. Le Monde reported him saying OpenAI had "annihilated" his field.

His numbers: he has written 3,000 to 4,000 pages of proofs in his career, yet in just 30 of the roughly 700 papers he counted about 4,000 pages directly tied to his work, and as many on questions he meant to study. He rejects a "bulldozer" approach and wants mathematicians to choose where AI helps. The same day, fellow Fields medalist Cédric Villani told France Inter it is "the most radical change in mathematics in at least four centuries, maybe 1,000 years."

【Verifiable】

RTS interview at the link; Le Figaro's Oct 11 write-up of Le Monde. Counts differ: Le Figaro says 16 of about 4,000 problems are in his area; RTS says more than 20 papers claim major results there. He signed the September 11 letter "A Severe Misalignment of AI in Mathematics" (mathandai.org); RTS says 28 of 46 living Fields medalists signed. OpenAI's IAS-hosted advisory group, AGMAI, formed in late September, advises on review and release but has no decision power. About 100 X posts named him in 24 hours, still 7 to 10 hourly.

【Our read】

The pages matter more than the tears: Tao's "proof abundance" now has a measurable case, and the bottleneck is reading and understanding, which Lean cannot do. He and Villani agree the results are significant; they disagree on pace and who sets it. The advisory group existed before October 6 and the 700-paper dump still happened. Watch whether AGMAI publishes its view and whether OpenAI's next batch is staged by field with mathematicians consulted first.