• FractalsInfinite@sh.itjust.works
      link
      fedilink
      arrow-up
      2
      arrow-down
      2
      ·
      edit-2
      9 months ago

      The tech is great at pretending to be human. It is simply a next “word” (or phrase) predictor. It is not good at answering obscure questions, writing code or making a logical argument. It is good at simulating someone.

      It is my experience that it approximates a human well, but it doesn’t get the details right (like truthness or reflecting objective reality), making it useless for essay writing, but great for stuff like character AI and other human simulations.

      If you are right, give an actual Iogical response only capable by a human, as opposed to a generic ad hominem. I repeat my question, Have you actually used any of the GPT3 era models?