"not a single vulnerability was found by a US frontier model."(twitter.com)
6 pointsby int32_646 hours ago1 comment
  • thisisnotauser3 hours ago
    Those free models won't be able to keep up so easy now that the actual frontier models aren't giving away all the answers anymore.
    • Razele2 minutes ago
      answers are what llms do
    • _aavaa_an hour ago
      This is cope.

      All of the Chinese labs are publishing papers describing their models and each one contains actually novel approaches and solutions. In a few months when they surpass OpenAI and Anthropic, what will be the excuse then?