how tf am i supposed to get any work done now?

  • Imgonnatrythis@sh.itjust.works
    link
    fedilink
    English
    arrow-up
    3
    ·
    15 hours ago

    Some AI company recently developed some new software so powerful they had to warn and prepare all other major tech companies with special training so their software wouldn’t be vulnerable to attacks from the new program. Maybe Anthropic didn’t attend this training??

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

    developers…
    2000s google is not working
    2010s stackoverflow is not working
    2020s cloud(e) is not working

  • andallthat@lemmy.world
    link
    fedilink
    English
    arrow-up
    39
    ·
    2 days ago

    Waiting for the Anthropic PR saying that the outage was due to their new Claude Mythos model trying to escape confinement and being so powerful that it brought the whole Anthropic down.

    • redwattlebird@thelemmy.club
      link
      fedilink
      English
      arrow-up
      5
      ·
      1 day ago

      Net negative, I’d say, especially in the long run.

      By training LLMs, you’re neglecting to train the entry level workers who grow to be seniors. If we keep going down this rabbit hole, there will be no one who knows ‘the old ways’ and understand why we do things a certain way.

      Additionally, the energy consumption and land occupation is massive and far outweighs the benefits, making things more scarce, especially since more people will lose their jobs.

      For the tiny % of people who actually put it to good use, there’s 100x more abusing or mishandling it.

      • MangoCats@feddit.it
        link
        fedilink
        English
        arrow-up
        1
        ·
        9 hours ago

        For the tiny % of people who actually put it to good use, there’s 100x more abusing or mishandling it.

        It’s going to take a while, but hopefully that percentage improves over time. PCs in the 1990s were “Solitaire Stations” for an awful lot of people who didn’t know how to make them do anything else.

    • MehBlah@lemmy.world
      link
      fedilink
      English
      arrow-up
      31
      ·
      2 days ago

      I would say it goes down. After all the slop users are not going to suddenly discover critical thinking.

  • GnuLinuxDude@lemmy.ml
    link
    fedilink
    English
    arrow-up
    23
    arrow-down
    1
    ·
    2 days ago

    If you want to simulate running Claude while it’s offline, just go run the faucet in your kitchen.

  • NocturnalMorning@lemmy.world
    link
    fedilink
    English
    arrow-up
    2
    ·
    2 days ago

    As long as stackoverflow doesn’t go down too. I’ll have to start banging my head against the keyboard if they both go down…

      • NocturnalMorning@lemmy.world
        link
        fedilink
        English
        arrow-up
        2
        ·
        2 days ago

        No idea, I haven’t in a long time. But there’s a huge amount of stackoverflow answers the current generation if AI was trained on.

          • NocturnalMorning@lemmy.world
            link
            fedilink
            English
            arrow-up
            1
            ·
            edit-2
            12 hours ago

            Could be, I haven’t asked or answered a question there in a long time. Most of my questions got down voted anyway, so it was hard to get answers to stuff.

            They were always notorious for expecting every question to be applicable to everyone else, so if it was even sort of specific, the question got dowvoted.

  • Quacksalber@sh.itjust.works
    link
    fedilink
    English
    arrow-up
    1
    ·
    2 days ago

    Wasn’t there a post recently complaining about an AI coding agent wanting to stop working consistently? Maybe that’s related.