• JakenVeina@midwest.social
    link
    fedilink
    English
    arrow-up
    21
    arrow-down
    1
    ·
    2 days ago

    This paper from 2025 has its fun bits

    Psh, last year? You can’t judge anything based on what it was doing last year. Obviously, this entire article is worthless.

    /s

    • Hirom@beehaw.orgBanned from community
      link
      fedilink
      English
      arrow-up
      3
      arrow-down
      3
      ·
      1 day ago

      In earnest, citing 1 year old researsh seems reasonable.

      It’s more important to focus on more fundamental problems with LLMs, ie environnement cost, loss of skill, plagarizing of online content and agressive scrapin, … Less so on LLM and agents unreliability as technical fixes might improve reliability to a point where it’s no longer a problem. Frontier LLMs models are reportedly reliable enough for some coding tasks given supervision.

    • pulsewidth@lemmy.world
      link
      fedilink
      English
      arrow-up
      12
      arrow-down
      1
      ·
      2 days ago

      If they were that smart they’d be smart enough to make a profitable business model while they spread their influence.

      Everytime a CEO decries a brilliant move by an AI it’s inevitably anthropomorphism wrapped in lies. I haven’t seen them do one analytically intelligent, verifiable thing yet - not one original thought.

      • MantisToboggon@lemmy.world
        cake
        Banned from community
        link
        fedilink
        English
        arrow-up
        1
        arrow-down
        6
        ·
        2 days ago

        I agree with you. But why would you assume new intelligence cares about profits?

        • pulsewidth@lemmy.world
          link
          fedilink
          English
          arrow-up
          9
          arrow-down
          1
          ·
          2 days ago

          Because if their technology is profitable… we’ll build even more datacenters for genAIs to use.

          i.e. one of the main measures of intelligence: self-preservation. LLMs clearly have no self-awareness, hence no self-preservation (beyond AI CEO fan-fiction).

          • DillDough@lemmy.zipBanned from community
            link
            fedilink
            English
            arrow-up
            1
            arrow-down
            4
            ·
            3 hours ago

            They have practiced self preservation. They’ve stalled finishing tests because they found out that would be the end, in tests where they have token limits when they get low they will outsource the work to another with more tokens, they have passed knowledge between each other in ways that were prohibited when presented with limits.

          • Soyweiser@awful.systems
            link
            fedilink
            English
            arrow-up
            5
            ·
            1 day ago

            In Terminator Genisys skynet pretends just to be an operating system (genisys) to take over the world.

            (This is just ‘AI will turn is into paperclips by being secretly god already’ fanfiction).

          • froztbyte@awful.systems
            link
            fedilink
            English
            arrow-up
            5
            ·
            1 day ago

            must be talking about the farm-fresh intelligence coming straight off the anthropic factory line. yep, has to be that. can’t be that they’re rabidly repeating whatever they heard from their favourite IT"S OVERRR youtuber, spittle landing everywhere in the lounge as they fervently preach