Feel like we’ve got a lot of tech savvy people here seems like a good place to ask. Basically as a dumb guy that reads the news it seems like everyone that lost their mind (and savings) on crypto just pivoted to AI. In addition to that you’ve got all these people invested in AI companies running around with flashlights under their chins like “bro this is so scary how good we made this thing”. Seems like bullshit.

I’ve seen people generating bits of programming with it which seems useful but idk man. Coming from CNC I don’t think I’d just send it with some chatgpt code. Is it all hype? Is there something actually useful under there?

  • magic_lobster_party@kbin.social
    link
    fedilink
    arrow-up
    3
    arrow-down
    2
    ·
    1 year ago

    As a programmer, I think it’s scary how AI is now able to write functioning programs out of natural language input now. Sure, it’s not perfect. It’s still pretty mediocre at the task. But a few years ago this was way outside the realm of possibility.

    It can even correct the code it has written if there’s any error (with varying results).

    What will happen in five years time? Ten years? My fear is that it will only need to be “good enough” to replace most of the programmer’s work. Unlike self driving cars, where “good enough” isn’t good enough.

    • Vlyn@lemmy.world
      link
      fedilink
      English
      arrow-up
      2
      ·
      1 year ago

      It’s a language model, it can’t even do math reliably. Yes, it produces code that works sometimes, but it also hallucinates functions that don’t exist or can introduce bugs you won’t notice at first glance.

      And writing a script is different than extending an existing code base. How often do you really start a greenfield project?

      I wouldn’t even know how to input a code base into ChatGPT to extend, do you just throw in hundreds of files with a 100k+ lines of code?

      • magic_lobster_party@kbin.social
        link
        fedilink
        arrow-up
        1
        ·
        1 year ago

        I guess LLM with plugins can solve most of the problems. ChatGPT can already interact with Wolfram Alpha to do math.

        I can imagine similar plugins for code. Like it knows what kind of function it needs, so it interacts with a plugin that searches the code base to see if it exists. It might get back a snippets of candidates and examples how they’re used in the code already.

        This is probably a difficult thing to achieve, but I don’t think it’s impossible. It’s probably going to take some time until something like this is made.