• ArbitraryValue@sh.itjust.works
    link
    fedilink
    English
    arrow-up
    7
    arrow-down
    12
    ·
    3 days ago

    It understands relationships between concepts, which is something that can be learned from reading text even without firsthand experience of the world. “Tariffs” is associated with “recession” and “recession” is associated with “bad”.

    • Optional@lemmy.world
      link
      fedilink
      English
      arrow-up
      13
      arrow-down
      2
      ·
      3 days ago

      Sort of. It understands “0.0023” is associated with “0.0037” and “0.0037” is associated with “0.15532”

      • ArbitraryValue@sh.itjust.works
        link
        fedilink
        English
        arrow-up
        4
        arrow-down
        13
        ·
        edit-2
        3 days ago

        Yes, but I don’t see that as particularly significant in this context. Information, including the knowledge of economic theory stored in a human brain, can be represented digitally. The fact that a present-day AI presumably can’t actually experience what it’s like to be unhappy as prices rise and incomes fall doesn’t affect its ability to reason about economics.

        • Optional@lemmy.world
          link
          fedilink
          English
          arrow-up
          15
          ·
          3 days ago

          We should probably just agree to disagree. I think the strides made in AI are at the very least impressive and have made some things (text-to-speech, for example) better - if not enormously then at least noticeably.

          But there isn’t a true analog to be had between calculated probabilities and conscious thought. The former is a mimic of varied competence, but has no logic inherent to it. It requires human maintenance, it’s only path to “growth” if we want to call it that, is a black-box of infinite probabilities it calculates at incredible speed.

          It’s a super-magic-8-ball that we choose to pretend has agency of some sort. But it does not.

    • shalafi@lemmy.world
      link
      fedilink
      English
      arrow-up
      3
      ·
      3 days ago

      “Tariffs” is associated with “recession” and “recession” is associated with “bad”.

      Nailed it. ChatGPT gave a pretty balanced definition, but at least it popped out “bad”.

      And if you put in Smoot-Hawley:

      Ah, the Smoot-Hawley Tariff Act — one of the most infamous tariff laws in U.S. history. It’s a textbook case of how tariffs can go very wrong.

      These people responding think you think AI is thinking. See, because they’re smarter than you! This place fucking annoys hell out of me sometimes, just like old reddit. At least we’re not run over with bots and fascists.