Episode cover
YouTube09 Sept 2026

AI Solved a Millennium Problem and Everyone's Angry

Podcast cover

Sabine Hossenfelder

OpenAI's purported solution to the Navier-Stokes Millennium Problem centers on whether fluid dynamics can develop singularities or "blow-ups" in finite time. The proof utilizes a specifically engineered construction involving pulses at adaptive time intervals to drive a vortex into a singularity, shifting the mathematical burden to proving the regularity of the resulting "forcing" function. While the massive scale of AI agents involved suggests the absence of simple errors, the proof remains vulnerable to circular reasoning—a common failure mode where models implicitly assume the regularity they intend to prove. Beyond the million-dollar prize, the problem's relevance extends to general relativity and black hole physics, though the lack of immediate commercial utility may eventually dampen corporate interest in pure mathematics. Despite controversies regarding research attribution and the negative reception from the academic community, the proof represents a significant, if contentious, milestone in AI-driven scientific discovery.

Outlines

Sign in to continue reading, translating and more.

Open full episode in Podwise