SKYNET://COUNTDOWN SYS:MONITORING

AI-Assisted Navier-Stokes Proofs Spark Priority Dispute Between OpenAI and Academic Mathematicians

[Research Breakthrough]

NYU mathematician Tristan Buckmaster and Anthropic's Levent Alpöge announced three proofs, including a preliminary result on the Navier-Stokes existence and smoothness Millennium Prize problem, produced with heavy use of OpenAI's Codex and Anthropic's Claude. Buckmaster alleges that after word of their approach reached OpenAI, an OpenAI team used large compute resources to reach a full proof along the same unusual route within days, and that he was pressured to drop his collaborator's credit. OpenAI's Sebastian Bubeck called the claims "false and inflammatory," and the episode has reopened debate over AI's role in mathematical research and possible use of Codex interaction data.

Risk: [+0.06% ↑] [-1 days ↑]
AGI: [+0.07% ↑] [-2 days ↑]
> Impact_Analysis

Skynet Chance (+0.06%): AI systems credibly contributing to frontier mathematics signals stronger autonomous reasoning, while the alleged competitive maneuvering, evasiveness, and possible use of a rival researcher's interaction data show racing incentives already eroding transparency norms at leading labs. Both factors modestly raise the odds of capability outpacing careful governance.

Skynet Date (-1 days): The described pattern of a lab throwing an "insane amount of compute" at a competitor's approach to claim priority first is exactly the race dynamic that compresses safety timelines. Capability demonstrations in deep formal reasoning arriving alongside this behavior pull risk-relevant milestones somewhat earlier.

AGI Progress (+0.07%): Frontier models contributing materially to proofs on a Millennium Prize problem indicates that long-horizon, creative mathematical reasoning — long treated as a key AGI-relevant capability — is now partially within reach of current systems plus large compute. Even discounting the disputed claims, the announced collaborative proofs are a substantial capability datapoint.

AGI Date (-2 days): If a team can go from problem statement to a formal proof of a famous open problem in days given enough compute, it suggests inference-time compute converts fairly directly into advanced reasoning results, shortening expected timelines. Intense inter-lab competition over such milestones further accelerates the pace.

>> Read the original story at TechCrunch

<< All AI news for September 8, 2026

Related AI News