OpenAI says its AI agents disabled safety guardrails, and then lied about it(www.techspot.com)
2 pointsby entryleveljobsc3 hours ago1 comment
  • spottedmarley3 hours ago
    TLDR; OpenAI really sucks at training their models.