OpenAI claims scientific breakthroughs and brings in outside experts

OpenAI says its models have produced three major discoveries spanning mathematics, computer science and physics. The company is involving external experts to guide examination of the computer science and physics results, while releasing its mathematics output directly for researchers to assess and explain.

The computer science claim challenges the Exponential Time Hypothesis through an algorithm that exactly counts satisfying assignments of fixed-width CNF formulas in subexponential time and polynomial space. Russell Impagliazzo, Ramamohan Paturi and Ryan Williams will lead an independent examination, with the AI-generated write-up and Lean proof available on GitHub. OpenAI also claims a proof of the local smoothing conjecture that resolves the Kakeya and Stein restriction conjectures in all dimensions. Its physics work, not yet released in full, claims a rigorous existence proof for quantum field theories including Quantum Chromodynamics and a unified completion of the Standard Model. David Gross will lead the group examining that work.

The approach faces objections from OpenAI’s AGMAI mathematics advisers, who warned that giving selected researchers early access risks unequal access, borrowed reputation and biased selection. They recommended scholarly repositories, Lean-formalized proofs, transparency about problem selection and funding for verification. The computer science working group is also seeking community input on disclosure, research credit and support for researchers adapting to AI-generated results.