History LLMs: Models trained exclusively on pre-1913 texts by from Hacker News on 2025-12-18 22:39 (#7294H) Comments