OpenAI announced that its internal AI models have refuted the Exponential Time Hypothesis (ETH), which was formulated in 1999. The hypothesis suggests that the 3-SAT problem cannot be solved in sub-exponential time. The company has engaged independent research groups, including the original authors of the hypothesis and Nobel laureate David Gross, to verify these findings.
In mathematics, the models reportedly proved the local smoothing conjecture proposed by Christopher Sogge in 1991. According to OpenAI, this result further implies proofs for the Kakeya conjecture and the Stein restriction conjecture across all spatial dimensions. The formalization of these proofs in Lean was completed on October 19.
Regarding theoretical physics, the company claims its models constructed proofs for the existence of interacting quantum field theories in four-dimensional spacetime, including quantum chromodynamics and the Yang-Mills theory. These results were formalized in Lean by October 16. The findings are being reviewed by a group led by David Gross.
OpenAI has introduced new protocols for sharing scientific results following criticism regarding its publication approach. The company stated that participation in research groups is voluntary and unpaid, with no non-disclosure agreements required. OpenAI noted it does not claim scientific credit for discoveries generated by its models.
