https://futurism.com/the-byte/government-ai-worse-summarizing

The upshot: these AI summaries were so bad that the assessors agreed that using them could require more work down the line, because of the amount of fact-checking they require. If that’s the case, then the purported upsides of using the technology — cost-cutting and time-saving — are seriously called into question.

  • Infamousblt [any]@hexbear.netEnglish
    47·
    2 months ago

    Sure, but it’s cheaper, and so if we fire all of our employees and replace them with AI, for this next quarter our profits will go WAY up, and then I can get my bonus and retire. So it’s totally fine!

    • hexaflexagonbear [he/him]@hexbear.netEnglish
      24·
      2 months ago

      There’s a certain level of risk aversion with these decisions though. One of the justification of salaries for managers who generally don’t do shit is they take “responsibility”. Honestly even if AI was performing at or above human level, a lot of briefs would have to be done by someone you could fire anyway.

      And as much as next quarter performance is all they care about, there are still some survival instincts left. My last company put a ban on using genAI for all client facing activities because a sales guy almost presented a deck with client is going to instantly walk out levels of wrong information in it.

      • UmbraVivi [he/him, she/her]@hexbear.netEnglish
        15·
        2 months ago

        Yeah, that’s something I was thinking about. With human employees, you can always blame workers when anything goes wrong, fire some people and call it a day. AI can’t take responsibility the same way.

    • Diuretic_Materialism [he/him]@hexbear.netEnglish
      12·
      2 months ago

      They’ll fire everyone and love the short term profit boost but within a year realize it’s fucking up their production processes. But they’ll be so hooked on all that money saving that they’ll pull some sneaky ways of rehiring everyone buy for less money and benefits.

  • nat_turner_overdrive [he/him]@hexbear.netEnglish
    37·
    2 months ago

    Any time a client mentions “I asked ChatGPT” or any of the other hopped-up chatbots, what follows is always, without fail, completely ass-backwards and wrong as hell. We literally note in client files the ones who keep asking some shitty chatbot instead of us because they’re frequent fuckups and knowing that they’re a chatbot pervert helps us narrow down what stupid shit they’ve done again.

  • queermunist she/her@lemmy.mlEnglish
    28·
    2 months ago

    Yeah I’ve purged “AI” from my vocabulary, at least for now.

    These are chatbots. That’s it. “AI” is a marketing term.

    • keepcarrot [she/her]@hexbear.netEnglish
      5·
      2 months ago

      I recall my AI class discussed a bunch of different things that people call AI that don’t come anywhere near “replacement human”. For instance, the AI in red alert 2 has some basic rules about buildings and gathering a certain number of units and send them the players way.

      Obviously, RA2s “AI” isn’t being used for labour discipline and llms are massively overhyped but I think getting hung up on the word is… idk, kinda a waste of time (as I feel like a lot of this thread is)

        • keepcarrot [she/her]@hexbear.netEnglish
          3·
          2 months ago

          I think people are allowed to be annoyed, but if thats all you want to talk about i think its a waste of energy? It’s just language, we can call it flubbon if you like and move the conversation along.

          Unless we want to get bogged down talking about whether band aids “medical adhesive strips”, which is a perfectly fine conversation to have if that’s what both participants want to talk about.

            • keepcarrot [she/her]@hexbear.netEnglish
              4·
              2 months ago

              Because people call it an AI instead of a bunch of related trained predictive algorithms? If the other things were happening (labour discipline, art theft, using a gallon of water to run a bad google search) but people were using whatever term you wanted, what would actually change?

              Like, I’m not saying it’s wrong to be annoyed by these companies ad copy, and there’s absolutely people out there who think “AI” is more human than their employees, it’s just a huge amount of time and energy wasted over a relatively minor part of the whole relationship. Even this 3 reply exchange here is probably too much.

  • FnordPrefect [comrade/them, he/him]@hexbear.netEnglish
    21·
    2 months ago

    porky-happy “Pfft! That only matters if you care about factual accuracy. So let me make it real simple: Facts don’t care about your feelings, and my finances the future doesn’t care about your facts!”

  • 7bicycles [he/him]@hexbear.netEnglish
    20·
    2 months ago

    The upshot: these AI summaries were so bad that the assessors agreed that using them could require more work down the line

    Oh man, this’d be really bad if we structured our society in such a way that instead of taking a holistic approach of looking at things it was all random KPIs in an excel file that measure one very narrow field of view of things like how fast I am at my job

  • Tommasi [she/her]@hexbear.netEnglish
    17·
    2 months ago

    Pretty sure most people who’ve used Ai in their work know the results kinda sucks, and only use it because writing a prompt for an LLM is way faster than writing anything yourself.

    • invalidusernamelol [he/him]@hexbear.netEnglish
      3·
      2 months ago

      I still think in development environments, limited LLM systems can be used in tandem with other systems like linters and OG snippets to help maintain style and simplify boilerplate.

      I use Co-Pilot at work because I do development on the side and need something to help me bash out simple scripts really fast that use our apis. The codebase we have is big enough now (50,000 ish lines and hundreds of files) so it tends to pick up primarily on the context of the codebase. It does still fallback to the general context pretty often though and that’s a fucking pain.

      Having the benefits of an LLM trained on your own code and examples without the drawbacks of it occasionally just injecting random bullshit from its training data would be great.

  • CyborgMarx [any, any]@hexbear.netEnglish
    16·
    2 months ago

    Maybe because it’s not genuine AI

    I love how all the corporate bootlickers for over three years now have just assumed some real breakthrough in emergent general intelligence took place and now humanity can build rudimentary consciousness

    What world are these dipshits living in, it’s just marketing for data aggregators not a replacement flesh and blood humans

  • WafflesTasteGood [he/him]@hexbear.netEnglish
    15·
    2 months ago

    I’ve kinda seen this in manufacturing for the last few years. Not explicitly “AI” but newer equipment designed around being smarter and not requiring skilled operators. Think like WordPress but for industrial machines; it might do basic stuff pretty well but fails at complex operations, and it’s an atrocity if you ever look behind the scenes to do some troubleshooting.

    • btfod [he/him, comrade/them]@hexbear.netEnglish
      17·
      2 months ago

      Hell yeah, smart machine? That’s gonna cost a premium. Oh, and because these machines are so sophisticated, you’ll need a higher tier support contract, that’s another premium… I mean it’s not like you have skilled technicians on staff anymore, they all retired and all your new guys just know how to press “play,” since we made the machines so easy to use… you’re not fixing anything yourself anymore.

      Back to your support contract, now we have the Bronze tier which gets you one of our field techs out there within 48 hours, but if your business can’t handle that kind of downtime we could upgrade you to Silver or Gold…

      • impartial_fanboy [he/him]@hexbear.netEnglish
        6·
        2 months ago

        To expand on that for people who think it’s all just smoke and mirrors. I think, just like the assembly line, work places will be reorganized to facilitate the usefulness/capabilities of LLM’s and, perhaps more importantly, designed to obviate their weaknesses.

        It’s just that people are still figuring out what that new organization will look like. There hasn’t been a Henry Ford type for LLM’s yet (and hopefully won’t be a Nazi this time). Obviously there’s no guarantee there will be such a person/organization but I don’t think it super unlikely either.

          • impartial_fanboy [he/him]@hexbear.netEnglish
            8·
            2 months ago

            I do think people here have a tendency to just hate all of it out of hand, which I get to some extent.

            Yeah the hype cycle is certainly annoying. As is the accompanying fire/re-hire at lower pay cycle that follows any automation.

            ignoring the fact that it can render pretty amazing looking videos in such a short time span.

            I actually think the generative aspect of neural networks is the least interesting/useful/innovative/etc. Though it will admittedly be more interesting when an LLM can say, use blender to make a video rather than just wholesale generating it. Or at least generate the files/3d models necessary to have it be edited by a person just like they would anything else. I suspect there will have to be a pretty significant architecture change for them to be able to make convincing/coherent movie-length videos.

            Chaotic system control, like they’re doing with nuclear fusion plasma is the most interesting, to me anyway.