They train LLM only with data up to 1930 and it still solves Python problems(www.theregister.com)
1 pointby Multipassionate5 hours ago1 comment
  • 5 hours ago
    undefined