AI Solved a Millennium Problem and Everyone’s Angry

Summary
Physicist and science commentator Sabine Hossenfelder analyzes OpenAI’s claimed solution to the Navier–Stokes Millennium Prize problem demonstrating finite-time blowup under forcing. She examines the construction used in OpenAI’s 160-plus-page paper, addresses common criticisms and potential failure modes of AI-generated mathematical proofs, and discusses the surrounding priority disputes and community reactions.

What is shown

• [00:52] Display and breakdown of the incompressible Navier–Stokes equations ($\partial_t \mathbf{u} + (\mathbf{u} \cdot \nabla)\mathbf{u} = -\nabla p + \nu \Delta \mathbf{u} + \mathbf{f}$, $\nabla \cdot \mathbf{u} = 0$), defining velocity $\mathbf{u}$, pressure $p$, viscosity $\nu$, and forcing $\mathbf{f}$.
• [01:08] Footage of a meteorologist presenting hurricane tracking on a studio display to illustrate fluid dynamics applications.
• [01:43] Screenshot of the title page and table of contents of the 165-page OpenAI manuscript titled “FINITE TIME BLOWUP FOR NAVIER-STOKES”.
• [02:21] 3D rendered diagram of the blowup construction showing an inward spiral with axial stretching around a central vortex axis.
• [02:30] Web page screenshot of the Clay Mathematics Institute Millennium Prize problem description for the Navier–Stokes equation.
• [05:16] Screenshot of an X (Twitter) exchange between @anabology and Sam Altman (@sama) discussing research priority and turning focus toward room-temperature superconductors.

Claims & numbers

• The presenter notes that OpenAI’s claimed proof spans around 160 pages (the displayed manuscript table of contents lists sections through page 157 and references on page 165).
• The presenter notes OpenAI reportedly deployed thousands of AI agents to construct and verify the proof.
• The presenter states that based on her experience with OpenAI models, GPT-5.6 still struggles with circular arguments, and asserts her impression that the new Astra model is worse than GPT-5.6 at mathematics.
• The presenter estimates a 90% chance that OpenAI’s blowup proof is mathematically correct, alongside a 99% chance that mathematicians will keep complaining about it.
• The presenter mentions the Clay Mathematics Institute’s $1 million Millennium Prize associated with the problem.

Notable quotes

• [00:18] “Let me say ahead that I’ve been struck by how extremely negative the reaction has been to the OpenAI proof...”
• [04:00] “By the way, my impression is that the new Astra is worse than GPT-5.6 when it comes to maths, but that’s just an aside.”
• [06:16] “In summary, I think there’s a 90% chance this proof is correct, but a 99% chance that mathematicians will continue to complain about it.”

Assessment
This is an independent analysis and commentary video evaluating OpenAI’s paper release and community controversies. The presenter summarizes the underlying fluid mechanics and the preprint’s analytical structure from an external perspective, noting that she has not checked the full 165-page proof herself.

Described by gemini-3.8-flash on 2026-10-08 from the video’s audio and frames.