A Rebuttal Claiming OpenAI's Connes Disproof Is Invalid Got Shredded on HN Within Hours — Including as Likely AI-Generated Itself
A PhilArchive PDF arguing OpenAI's claimed counterexample to Connes's rigidity conjecture is invalid reached 31 points and 41 comments on August 1, but the thread turned on the rebuttal rather than the proof: commenters noted the same author has published a claimed proof of the Riemann Hypothesis, identified a false assertion at its core (that a semidirect product with an abelian factor must admit that factor in its center), and argued the rebuttal itself reads as AI-generated. An OpenAI mathematician responded directly on X. The substantive question the thread raised and did not answer is the one that matters for everything downstream of machine-generated mathematics: when validation effort is this high, how does anything get adjudicated? Included as a signal about verification norms, not as a credible refutation — the claim did not survive scrutiny.
Source
↳ Follow the thread