OpenAI Quietly Weakened Its Math Breakthrough Claims

OpenAI revised its mathematics announcement from problems showing "no progress on the main result for at least a decade" to softer language about "substantial progress on long-standing open problems," with no correction note. Cambridge mathematician Francesco Fournier-Facio noted the original text minimized groundwork by researchers Andreas Thom and Gábor Kun on the non-sofic groups result.