Mathematician's statement alleges OpenAI moved against researchers about to beat it to a proof

IgorCarron · x · 2026-09-08

Tristan Buckmaster published a statement described as one of the ugliest things to come out of a frontier lab — not the math, but what happened after OpenAI learned two people were about to beat them. Per the account: Buckmaster and Levent Alpöge spent a year on the Córdoba / Martínez-Zoroa smooth forcing program, got blowup results for Boussinesq and 3D Euler on Aug 15, Lean-verified Aug 22, paying OpenAI out of pocket with drafts in private Codex sessions on an obscure program. Then on Sept 3 Buckmaster emailed OpenAI privately… (thread truncated; details unverified).

Related event: Mathematicians accuse OpenAI of rushing to claim Navier-Stokes breakthrough(3 posts)→

Original post →

More from Companies & People

Companies & People channel →