OpenAI’s Supposed Mathematical Breakthrough Devolves Into Explosive Drama as Mathematician Accuses It of Stealing His Work
TL;DR
OpenAI announced on September 8 that it had solved one of the Navier-Stokes Millennium Problems, saying 10,000 AI agents worked for 88 hours to find a scenario in which the fluid equations break down. NYU mathematician Tristan Buckmaster, who has spent years on the problem, says OpenAI followed a suspiciously similar path to his unpublished work and alleges the company drew on his Codex sessions. OpenAI denies any copying and says it cannot rule out that de-identified user data helped improve its models.
Nauti's Take
If the proof holds up, it marks real progress: swarms of agents could help researchers crack questions that have resisted people for decades. The bigger risk is trust, because if data from Codex sessions can feed into models even indirectly, every research idea typed into a chat window becomes exposed.
Anyone working on unpublished results should check the data settings of their AI tool now.
Tweets
Sources
