Archived story — belongs to the Tuesday, September 15, 2026 sweep.
P1 · Tectonic · Frontier AI · sweep Tuesday, September 15, 2026
Frontier models solve open mathematical problems and credit dispute
OpenAI and Anthropic push the frontier of mathematical discovery with unreleased models (Astra, Navier-Stokes, advances on Riemann), but the milestone is tainted by accusations of appropriation of unpublished work and by opacity regarding training data derived from researchers. The strategic consequence is twofold: proprietary compute capacity becomes infrastructure of scientific production, and credit attribution —legal and reputational— becomes the regulatory battleground of the next phase.
Title, thesis and gravity assessment written by an AI model, without prior human review.
Gravity85/100 · Tectonic
Why this gravity
OpenAI announces that an unreleased internal model (Astra) solves ten open problems in mathematics, including the Navier-Stokes millennium problem; Anthropic appears in the credit dispute. It does not rise to 90+ because the result lacks formal verification and is contaminated by accusations of appropriation of unpublished work.
- Independent sources
- 10
- Cross verticals
- Frontier AI, Tech Industry, Geopolitics, Investing
- Tier 1 actors
- OpenAI, Anthropic
- Gravity vs. the previous edition
- no change (85)
Lead source
On the Navier–Stokes Millennium Prize Problem · Simon WillisonMore: LessWrong (karma ≥25), Techmeme, MIT Technology Review, Decrypt, Hacker News, OpenAI Blog, Zvi Mowshowitz, El País América and 1 more
Timeline · Madrid time
- 02:00
- 16:35
- 22:34Ten advances in mathematics and theoretical computer science · Simon Willison
- 01:04OpenAI’s Unreleased Model Astra Solves Ten Major Open Mathematics Problems · LessWrong (karma ≥25)
- 18:25
- 07:42Navier-Stokes – Tristan Buckmaster [pdf] · Hacker News
- 19:13On the Navier–Stokes Millennium Prize Problem · Hacker News
- 21:00
- 21:05
- 23:38
- 01:55On the Navier–Stokes Millennium Prize Problem · Simon Willison
- 05:10What OpenAI’s latest controversy tells us about the future of math · MIT Technology Review
- 12:08
- 23:22
- 17:48Brand New AI Solves a Millennium Prize · Zvi Mowshowitz
- 20:02Brand New AI Solves a Millennium Prize · LessWrong (karma ≥25)
ShareWhatsApp