Wowed by a new paper I just read and wish I had thought to write myself. Lukas Berglund and others, led by Owain Evans, asked a simple, powerful, elegant question: can LLMs trained on A is B infer automatically that B is A? The shocking (yet, in historical context, see below, unsurprising) answer is no:

TL;DR: Artificial “Intelligence” is just as much a scam now as it has been the previous five or six times it was touted as The Next Big Thing™.

The referenced paper is available free.

  • @CrayonRosaryEnglish
    2•1 year ago

    You’re fucking pathetic if you expect a language model to reason. The example about mother and son is literally a logic puzzle that requires reasoning. Not much, but more than zero! You sound exactly like the people you are making fun of in your other comment.

    • @ttmrichterOPEnglish
      1•1 year ago

      And yet it’s pitched as Artificial Intelligence.

      Almost as if the LLM people are a bunch of grifters.

      Almost.

      • @CrayonRosaryEnglish
        1•1 year ago

        It is, but not general artificial intelligence.