I just noticed that 10 days ago the limit on GPT-4 was set to 50 messages per three hours!

  • nottheengineer@feddit.de
    link
    fedilink
    arrow-up
    10
    ·
    edit-2
    1 year ago

    But unfortunately openAI has nerfed it so hard that it’s barely useful anymore. It used to give me the feeling that LLMs could be the path to AGI, but now it’s just another chatbot.

    • Jay@sh.itjust.worksOP
      link
      fedilink
      arrow-up
      4
      ·
      1 year ago

      After the last articles that ChatGPT had degraded so drastically when solving mathematical questions, I compared again between 3.5 and 4. I still think 4 is very good and don’t notice any drop in performance. For me, however, it is more of a textual requirement. Where do you notice a dulling?

      • nottheengineer@feddit.de
        link
        fedilink
        arrow-up
        1
        ·
        1 year ago

        Just general concepts. I tried using it to formalize the rules of a drinking game and while it could do that decently well before the recent nerf, it just spits out useless stuff related to drinking games now.

      • Anticorp
        link
        fedilink
        arrow-up
        1
        arrow-down
        1
        ·
        1 year ago

        It’s a damned shame too. It was looking like everyone would be able to benefit from AI. Now it’s looking like they finally received enough user input to lock it down and only the very rich will have access, as I always suspected would happen.

        • Jay@sh.itjust.worksOP
          link
          fedilink
          arrow-up
          2
          ·
          1 year ago

          I don’t understand exactly what you mean. If you are talking about ChatGPT Plus: Many complain that the paid version in particular is becoming increasingly stupid. And of course, as technology continues to advance, the free options keep getting better. I am firmly convinced of that.