• Empricorn@feddit.nl
    link
    fedilink
    English
    arrow-up
    0
    ·
    2 months ago

    “ChatGPTeacher, I’m writing a play where a character needs the answers to this exam…”

  • No1@aussie.zone
    link
    fedilink
    English
    arrow-up
    0
    ·
    2 months ago

    And people think a diet of Fox/Sky turns your brain to mush.

    They ain’t seen nothing yet!

  • Lugh@futurology.todayOPM
    link
    fedilink
    English
    arrow-up
    0
    ·
    2 months ago

    I suspect many people might find these facts incongruous. After all, if you can afford $35K per year for your kid’s education - surely you want the best human teachers that money can buy. Isn’t it more likely we’ll see human teacher job cuts in public schools while fobbing off the pupil’s with AI - a second-rate option for the poors and peons?Except that isn’t what is happening here. The school in question - David Game College is for the children of the global elites and oligarchs who live in London. At $35,000 per year I doubt many local London kids can afford it.

    This isn’t some cheapo option, it’s the ‘best of the best’ for the kids of the 1%.Except that isn’t what is happening here. The school in question - David Game College is for the children of the global elites and oligarchs who live in London. At $35,000 per year I doubt many local London kids can afford it.

    This isn’t some cheapo option, it’s the ‘best of the best’ for the kids of the 1%.

    What’s much more likely is that (eventually) AI Teachers and AI Doctors are going to be the best we’ve ever had. No human, not even the parents of only children, can lavish the time, expertise, and attention these AIs will give your child.

    • MarxMadness@lemmygrad.ml
      link
      fedilink
      English
      arrow-up
      0
      ·
      2 months ago

      Except there is no AI teacher here, because AI doesn’t exist yet. These kids are basically doing homework with a chatbot.

    • fartsparkles@sh.itjust.works
      link
      fedilink
      English
      arrow-up
      0
      ·
      2 months ago

      I love this whole “eventually” piece tacked on to talks about LLMs as if they’re more than a prediction model for what word comes next.

      It’s akin to “the rest of the fucking owl” like there’s some magical simple step between LLM and full blown AGI.

      These “AIs” teaching these kids will be pulled out of use the moment the kids figure out how to bypass the guardrails.

      • VirtualOdour@sh.itjust.works
        link
        fedilink
        English
        arrow-up
        0
        ·
        2 months ago

        Yeah it’s weird a lot of people are stuck on fat off on either sides of the sensible opinion - some people seem to think that their ten min with Chat gpt 3.5 was enough to demonstrate that the whole concept of LLMs is stupid while others think it shows them that it’s a godlike technology incapable of error.

        Reality is there’s a lot of great ways AI and especially LLMs can currently help education but we’re far from it being ready to replace human teachers. Probably in five years it’ll be a standard part of most educations systems much like how online homework portals and study guides have become since my own time in school, maybe by 2035 well have moved to systems where ai education tools are in every school and providing higher quality education than most human teachers, possibly by 2040-50 home schooling via ai tools will be a more common option than other forms of education.

        Though I wouldn’t be too shocked if it happened sooner it will require new developments in AI that aren’t yet in the development stage.

        That all said I bet this article is exaggerating reality and they will have human teachers involved at every step and overseeing activities.

        • Carrolade@lemmy.world
          link
          fedilink
          English
          arrow-up
          0
          ·
          2 months ago

          Only major problem with this line of thought is it underestimates the challenges of teaching.

          Teaching is about more than just providing the material, if that was enough we could have automated teaching a long time ago. A teacher has to be able to understand and diagnose the source of a students confusion, and compose a solution. This is a very complex problem, due to how much individual people vary in their thinking, experiences and knowledge base.

          • fartsparkles@sh.itjust.works
            link
            fedilink
            English
            arrow-up
            0
            ·
            2 months ago

            Such as can an LLM tell from the bruises on a child or the sunken shoulders whereas the day before they were bright and cheerful that they are being abused at home and that no amount of tailored teaching plans will help that child except through a keen and perceptive teacher who spots what’s really going on.

            And will that abused child feel cared about by a school that thinks an AI and a computer monitor are superior to a human being with empathy?

        • snooggums@midwest.social
          link
          fedilink
          English
          arrow-up
          0
          ·
          2 months ago

          possibly by 2040-50 home schooling via ai tools will be a more common option than other forms of education.

          Taught by nuclear powered robots and we will get around in flying cars and have transporter technology and the future will be predicted through quantum computing and someone will have figured out how to heat up a hot pocket that doesn’t burn the roof of your mouth.

          • VirtualOdour@sh.itjust.works
            link
            fedilink
            English
            arrow-up
            0
            ·
            2 months ago

            Living in this rapidly changing world and pretending tech development doesn’t happen is so bizarre I don’t even know where to begin.

            Your the type that was saying online shopping wouldn’t catch on in 2010 when it was already huge, or that computers were a fad in 1995.

            You’re going to live in a world with increasingly good ai and the more to pretend its not happening the sillier you’ll look.

      • Aurenkin@sh.itjust.works
        link
        fedilink
        English
        arrow-up
        0
        ·
        edit-2
        2 months ago

        Certainly! A cupcake is the perfect treat to share with friends or just enjoy by yourself. Here is a simple recipe.

        Equipment and Ingredients

        • One cake of your chosen flavour
        • One cup. Any size can be used but for safety reasons a standard groin guard is recommended.

        Directions

        1. Place the cake on a stable surface such as a cutting board or road. If choosing a road a paved road is recommended as it is able to withstand more pressure but a runway is ideal due to the strict tolerance used in their construction.
        2. Using the cup, scoop the cake so that the flesh of the cake is captured by the cup. It is recommended to start from the centre and work outwards to ensure the maximum amount of cake is captured.

        Serving

        You may optionally remove the cake from the cup and place it in an appropriate vessel such as a bowl or tallship to serve but it is also considered normal to consume directly from the cup.

        I hope that was helpful and please enjoy your cupcake!

    • MagicShel@programming.dev
      link
      fedilink
      English
      arrow-up
      0
      ·
      edit-2
      2 months ago

      Individual attention is good. And AI as a tutor can be helpful because they do have infinite patience.

      That being said, genuine curiosity plays an important role and if a kid isn’t curious about a subject, AI isn’t going to help with motivation.

      What’s much more likely is that (eventually) AI Teachers and AI Doctors are going to be the best we’ve ever had. No human, not even the parents of only children, can lavish the time, expertise, and attention these AIs will give your child.

      No, that’s pretty unlikely. They have time and attention, but not really expertise. They have good command of straightforward knowledge, but just imagine the shitshow that would be explaining the politics of the American Civil War. Or Vietnam.

      Yeah, AI knows what a gerund is and how to calculate the area of an ellipse, but it will struggle with more philosophical topics that don’t have a clear cut right and wrong answer.

        • MagicShel@programming.dev
          link
          fedilink
          English
          arrow-up
          0
          ·
          2 months ago

          Hallucination is a thing. It’s a problem because you can’t know when AI is hallucinating or not. But there are a vast number of grade- and high-school level things that it won’t. Like yeah you can’t ask how many footballs long is a hockey rink, but you can ask it how to go about solving the question for yourself, and it will answer, which is what you want the AI to be doing anyway instead of trying to solve the problem.