• AGI you mean? IDK It’s arguably here. They’re senile, though. Working with a senile genius is the best way I can put it, I think. Is it “people?” IDK fuck if I know. How could we ever really know?

        • vala@lemmy.dbzer0.com
          link
          fedilink
          arrow-up
          5
          ·
          11 hours ago

          M not trying to be super rude here but the truth is, It’s arguable if you argue with people who don’t know any better. So in that sense, sure.

          When it comes to people that take these things seriously, it’s not a really a very hot topic. We’re clearly nowhere near AGI.

          Really good LLM != AGI

          • I don’t really know what AGI means, I guess. If we’re talking about something that benchmarks as high as the average human, we’re way, way past that. If we’re talking about something that benchmarks as high as experts in all areas, we’re not there, but we’re not that far off. When connected to the internet? It’s basically there.

            https://www.smithsonianmag.com/smart-news/ai-disproves-a-decades-old-mathematical-idea-the-biggest-conjecture-that-the-tech-has-played-a-role-in-yet-180989189/

            They’re literally solving problems that humans haven’t been able to. They ARE capable of novel concepts, to whatever degree. No offense but y’all seem like your knowledge is like three years old.

            • CileTheSane@lemmy.ca
              link
              fedilink
              arrow-up
              3
              ·
              8 hours ago

              If we’re talking about something that benchmarks as high as the average human

              Well my calculator can solve a random nine digit number to the power of another random nine digit number faster than any human can, so I guess calculators have been AGI for decades!

            • vala@lemmy.dbzer0.com
              link
              fedilink
              arrow-up
              4
              ·
              11 hours ago

              They’re literally solving problems that humans haven’t been able to.

              If you actually read instead of continuing to spam bullshit you would know this isn’t true.

              You say “haven’t been able to”. This is fucking false my dude. At best it’s solved problems humans haven’t been bothered to solve yet. The solutions already existed and it assembled them into a proof.

              I don’t really know what AGI means

              You don’t know what any of this means dawg. You are literally just saying shit that you heard other people say.

              • OK bro. I guess you can solve 130 year old equations? Or maybe this is a weird way of measuring intelligence anyways and I was just demonstrating that they’re capable of executing novel concepts, which they clearly are if they’re solving things no one has solved before. Like do you not get how that is the important part? Not that humans can’t solve it, it’s that we HAVEN’T. The AI, on it’s own, managed to solve a novel problem. We did not teach it to solve that problem. We taught it HOW to solve that problem, and then it did. And I was just using math because people say that it’s bad at math when it no longer is. It’s good at reasoning, it can come up with novel concepts. IDK. This feels like, “AI is whatever isn’t.”

                I’m honestly asking you what your benchmark would be, because I’m sitting here saying maybe and IDK and I wonder and you’re here stating things very rigidly, but somehow I’m the one being accused of having ideological blinders? Mate I asked you what they were bad at now and you just ranted instead. But I’m the unreasonable one unwilling to listen. Okay. Have a nice night, homie.

      • TheLegendaryAssholeOfJushinLiger@sh.itjust.works
        link
        fedilink
        arrow-up
        2
        ·
        edit-2
        12 hours ago

        Have you interacted with Fable? It scored on ARC AGI 3 and is pretty dang impressive at everything I’ve thrown at it. They’re solving problems humans haven’t. They’re just senile. Also is calling something “arguable” really an assertion? More of a headscratcher than a real question, I suppose.

        • CileTheSane@lemmy.ca
          link
          fedilink
          arrow-up
          2
          ·
          8 hours ago

          is calling something “arguable” really an assertion?

          Yes. “Arguably unicorns exist” is me asserting there exist legitimate arguments for unicorns existing. By saying it this way I get to imply they exist without having to provide any real arguments myself, and if anyone attacks my position I can fall back to “I’m not saying they do exist, just implying there are other people who think they exist without providing any sources or arguments actually supporting that claim.”

        • vala@lemmy.dbzer0.com
          link
          fedilink
          arrow-up
          5
          ·
          11 hours ago

          They’re solving problems humans haven’t

          Citation needed.

          I’m guessing we have different definitions of “solving” and “humans haven’t”.

          I’ve never seen any evidence that LLMs can extrapolate into truly novel problem spaces. Can problems be “solved” though interpolation? Sure but not likely novel ones.