OpenAI claims major scientific proofs, pending review

OpenAI announced on October 26 that its models had produced three major scientific results, all awaiting independent verification. The company says the work resolves longstanding mathematical conjectures, refutes the Exponential Time Hypothesis and establishes existence proofs for certain interacting quantum field theories. The mathematics and computer science materials include Lean formalizations published on GitHub, while a physics review group has chosen to withhold its materials.

In mathematics, OpenAI claims a proof of the local smoothing conjecture in every dimension, which would also settle the Kakeya and Stein restriction conjectures. In computer science, it says its models disproved the Exponential Time Hypothesis, a foundational assumption about the complexity of Boolean satisfiability. Russell Impagliazzo, Ramamohan Paturi and Ryan Williams are leading verification of that result. OpenAI agreed to release raw data and fund a conference at the reviewers’ request.

The physics claims include the existence of quantum chromodynamics, a unified completion of the Standard Model and pure Yang-Mills theory. The latter addresses a subproblem of a Millennium Prize problem, not the entire challenge. A group led by David Gross plans to write up the work itself. The Advisory Group on Mathematics and AI has criticized delayed disclosure. Lean verification establishes that a formal proof follows from its formal statement, but does not establish that the statement faithfully represents the intended conjecture.