OpenAI Cracks Navier-Stokes; Mathematicians Uneasy — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- OpenAI claimed its latest model solved the Navier-Stokes equations, one of seven Millennium Prize Problems carrying a $1m reward, deploying roughly 10,000 autonomous agents at an estimated cost of $15m after hearing rumours two Millennium Problems had already been solved.
- Mathematicians including Profs Colva Roney-Dougal (St Andrews), David Silvester (Manchester), and James Robinson (Warwick) called the field "very unstable," with Robinson dismissing the spectacle as "immature playground boasting writ large, underpinned by billions of dollars and the potential for significant environmental damage."
- The breakthrough relied heavily on prior work by Madrid-based mathematicians Diego Córdoba and Luis Martinez-Zoroa, according to Oxford philosopher of mathematics Alexander Paseau.
- Prof Tristan Buckmaster (NYU) and Levent Alpöge (Anthropic), who were running related Navier-Stokes work using OpenAI products, suspected the model had trained on their unpublished progress; OpenAI denied this after an internal investigation.
- University coursework faces immediate disruption — Silvester said standard take-home problem sets are now "dead" because authenticity cannot be verified, forcing lecturers to police which problems students may use AI on.
- Buckmaster warned that mathematics sharing norms have already collapsed: "The big story now in mathematics is that nobody wants to share anything … Mathematics is different today than it was only a few days ago."
Why it matters: The $15m spend and 10,000-agent deployment set a new scale for AI-driven mathematical proof — and mathematicians say the trust damage is immediate, with researchers now withholding work-in-progress from peers rather than sharing drafts.
Ask SkimNews
