artificial OpenAI researcher indicates they have an AI recursively self-improving in an "unhackable" box /u/MetaKnowing January 16, 2025 January 16, 2025 submitted by /u/MetaKnowing [link] [comments]