Overview
- OpenAI announced earlier this month that an internal multi‑agent system produced a proposed solution to the Navier–Stokes existence and smoothness problem and published a 166‑page paper plus Lean proof files for machine checking.
- The company says it verified the argument in Lean and has submitted the work for the Clay Mathematics Institute’s formal review, which is the established, slower process for deciding Millennium Prize validity.
- Mathematicians Tristan Buckmaster and Levent Alpöge have alleged that OpenAI may have been exposed to their private AI-driven work, an accusation OpenAI disputes and which remains unresolved after the company said it did not use user inputs past early July.
- More than a thousand mathematicians and 25 Fields Medal winners have protested OpenAI’s announcement process, arguing AI complicates attribution, risks appropriating unpublished ideas, and that machine proofs may lack human insight.
- Unconfirmed reporting says OpenAI is pursuing another Millennium problem, often linked to the Hodge Conjecture, but that claim is single‑sourced and the company is said to be moving more cautiously after the backlash.