artificial DeepSeek discovered their new model having an "aha" moment where it developed an advanced reasoning technique – entirely on its own /u/MetaKnowing January 21, 2025 January 21, 2025 submitted by /u/MetaKnowing [link] [comments]