What OpenAI’s latest controversy tells us about the future of math

Details

OpenAI announced its AI agents solved the Navier-Stokes existence and smoothness problem, one of seven Millennium Prize Problems set by the Clay Mathematics Institute in 2000, each carrying a $1 million reward. Only one Millennium Prize Problem had been solved in the previous 26 years. The Navier-Stokes equations model fluid flow; the specific question was whether the equations could produce physically impossible results — like infinite velocity — under any conditions. OpenAI's proof shows that the full equations can indeed break down. The announcement was immediately shadowed by an attribution dispute. NYU mathematician Tristan Buckmaster posted a Mastodon proof just the day before, showing that a simplified version of Navier-Stokes can break down — after nearly a year of AI-assisted work using models from both OpenAI and Anthropic, in collaboration with Levent Alpöge, an Anthropic employee. OpenAI's Sébastien Bubeck acknowledged the team was inspired by rumors of Buckmaster and Alpöge's work; the company denies using their research directly. Buckmaster and Alpöge dispute this. Technically, OpenAI used an internal model described as dramatically outperforming the Astra model released just last week — suggesting a significant capability jump that has not yet been made public. OpenAI says it will not claim the million-dollar prize. The episode raises structural questions about AI's role in mathematics: if the hardest open problems now require frontier AI capabilities and the compute resources only a handful of companies control, independent academic mathematicians become dependent on those companies' choices about access and attribution.