OpenAI says it's announcing three new findings in mathematics, computer science and physics on harmonic analysis, complexity theory and quantum field theory, adds its system has recently proved that exponential time hypothesis is false

OpenAI says it's announcing three new findings in mathematics, computer science and physics on harmonic analysis, complexity theory and quantum field theory, adds its system has recently proved that exponential time hypothesis is false.

Vendor claims of foundational scientific breakthroughs from AI labs have a mixed track record: some have held up, while a substantial share have later required correction, been found to restate known results, or had formal proofs contain gaps once scrutinised by domain specialists. The claim that the exponential time hypothesis is false would, if verified, carry consequences well beyond the company, since that hypothesis underpins large parts of complexity theory's established hierarchy of hardness assumptions; results at that level of consequence historically face months of peer review before acceptance, and self-announced proofs of famous open problems have far more often been retracted or revised than confirmed. The distinction worth drawing is between genuinely novel mathematics and an AI-assisted rederivation presented as new, as prior episodes of this kind have blurred the two. Follow-ons are whether a formal, checkable proof is published, whether independent mathematicians in complexity theory validate it, and whether the company's own framing shifts from certainty to qualification on follow-up. For the equity read-through, moves in the AI-adjacent complex on such announcements have tended to be sentiment-driven and to fade if expert scrutiny undercuts the claim; the tell is the reaction of the specialist community rather than the initial tape.