LLMs are getting better at character-level text manipulation from Hacker News on 2025-10-13 19:39 (#70QTN) Comments
