Mathematician Points Out Flaw in OpenAI Paper: Core Proof Relies on False Theorem

fkasummer · x · 2026-08-05

A mathematics researcher pointed out a theoretical flaw in OpenAI's paper regarding the existence of nonsofic groups. The paper heavily relies on Theorem 1 from Kun's 2019 paper as a black box to prove its core point.

However, Theorem 1 in Kun's paper is proved as a consequence of Theorem 3, which has been shown to be false as stated. This discovery challenges the mathematical rigor of OpenAI's publication.

Related event: OpenAI's Math Proof Challenged by Mathematicians(3 posts)→

Original post →

More from Research

Research channel →