They train LLM only with data up to 1930 and it still solves Python problems(theregister.com)4 points by Multipassionate 112 days ago | 0 comments