Physix FrontierPhysix Frontier · News Briefing Card (QbitAI · Oct 11, 2026)
OpenAI Posts Two Kakeya Papers, Lifting 3D Result to 4D
KEY FACTS
- On October 7, OpenAI published 722 mathematical manuscripts on GitHub, with group 074 containing two Kakeya conjecture papers.
- One 97-page paper claims to prove the three-dimensional Kakeya maximal function conjecture, stronger than the set-version result of Wang Hong and Zahl.
- Another 175-page paper claims to prove the Hausdorff dimension conjecture for four-dimensional Kakeya sets, where the previous best result was about 3.059.
- The author field of both papers lists only OpenAI; they have not undergone peer review, nor have they been formalized in Lean.
- Terence Tao led the Human Mathematics Association in issuing a statement urging mathematicians to stop collaborating with OpenAI.
KEY DATA
722Total public math manuscripts
97 pages3D paper page count
175 pages4D paper page count
about 42%Lean formalization ratio
PHYSIX OBSERVATION
AI mass-produces high-dimensional mathematical proofs, turning peer review and authorship attribution into bottlenecks. Terence Tao's boycott statement points to the core contradiction: problem-solving speed far exceeds human capacity to digest, and the mathematical community may lose its say over which problems to choose. If the proofs hold, AI will shift from tool to competitor, forcing a reconstruction of the academic evaluation system.
Source: QbitAI report