
OpenAI Quietly Softened Its Math Breakthrough Claims With No Correction Note
OpenAI changed its mathematics announcement from sharing results on problems with "no progress on the main result for at least a decade" to language about "substantial progress on long-standing open problems." No correction note or explanation accompanied the revision. Cambridge mathematician Francesco Fournier-Facio said the original text minimized groundwork by researchers Andreas Thom and Gábor Kun on the non-sofic groups result.
Published