Hoskinson Responds to OpenAI Announcement

Cardano founder Charles Hoskinson stated on September 9 that artificial intelligence had advanced in formal mathematics far beyond what he previously anticipated. His remarks came during a YouTube broadcast shortly after OpenAI reported that an internal system had produced a proposed solution to the Navier-Stokes Millennium Prize Problem.

Hoskinson described the reported capabilities as pretty remarkable. He explained that he had expected formal systems mainly to help larger teams of mathematicians collaborate and verify human-written proofs. He did not foresee large language models generating complete proofs on their own so soon.

Hoskinson has longstanding ties to formal mathematics research. In 2021 he donated $20 million to Carnegie Mellon University to establish the Hoskinson Center for Formal Mathematics.

Details of OpenAI’s Reported Result

OpenAI published its findings on September 8. The company said an internal model coordinated roughly 10,000 agents and generated a proposed solution after 88 hours of work. GPT-6 Astra then spent an additional 17 hours formalizing and checking the argument in the Lean proof assistant.

The construction aims to show that an initially smooth, stationary fluid can develop a singularity in finite time when subjected to a smooth external force. OpenAI stated that this satisfies statements C and D in the official Millennium Prize formulation. The company released an analytical paper along with Lean code and said it does not plan to claim the associated $1 million prize.

A Lean formalization supplies machine-checkable verification that the encoded steps follow from the stated assumptions. It does not by itself confirm that every definition and assumption matches the intended mathematical problem.

Clay Institute Status and Review Process

The Clay Mathematics Institute continues to list the Navier-Stokes problem as unsolved. Its rules require a proposed solution to appear in a qualifying publication, followed by a waiting period of at least two years and general acceptance by the global mathematics community. OpenAI’s announcement and formalization therefore do not constitute immediate institutional recognition.

Mathematicians must still examine whether the construction meets the precise problem statement and whether the use of external forcing addresses the question as commonly understood.

Provenance Questions and Research Privacy

The announcement prompted scrutiny involving New York University mathematician Tristan Buckmaster and Anthropic researcher Levent Alpöge. The pair had been developing a related result on the Euler equations that also employed a forcing approach. Buckmaster questioned whether private work entered into OpenAI’s Codex system might have contributed to the company’s output, while stopping short of alleging proven misconduct.

OpenAI denied accessing their specific work. The company added that it could not completely rule out the possibility that de-identified data from product use had helped improve its models. OpenAI maintained that its proof was developed independently and differed from the researchers’ efforts.

Hoskinson argued that the episode should prompt researchers to reconsider how they handle unpublished ideas. He said scholars who place prompts, notes, and research logs into centralized cloud AI services need to weigh whether that material remains confidential.

He further noted that sufficiently advanced models can take existing work, improve it, iterate on it, and formalize it to the point of addressing hard problems. Hoskinson presented private AI environments as one practical response for protecting intellectual property while still using powerful models.

Public examination of OpenAI’s paper and Lean formalization will determine the next steps. Until specialists complete their review and Clay’s formal conditions are satisfied, the result remains a claimed solution rather than a recognized resolution of the Millennium Prize Problem.