• KindaABigDyl@programming.dev
      link
      fedilink
      arrow-up
      1
      ·
      6 months ago

      This is already how the military works BC they lost the source code for ancient machines. They’ve gotta now hire reverse engineer researchers to help out

  • ulterno@lemmy.kde.social
    link
    fedilink
    English
    arrow-up
    1
    ·
    6 months ago

    One of my previous employers once told me (abridged)

    It’s not like old times when we could slowly work to get a perfect result.
    Nowadays, we need perfect results, fast.

    They were asking me to do technical content writing for their website.
    I quickly realised that it’s actually the threshold for calling something “perfect”, that has lowered over time.

    Clearly, I was not fit for that work, because instead of just plagiarising and paraphrasing stuff from other websites, I insisted on reading up on material from multiple sources, understanding it well and then writing it down myself. That makes it pretty slow.

    That was a year before ChatGPT, or I would just have used that thingy.

    • KindaABigDyl@programming.dev
      link
      fedilink
      arrow-up
      1
      ·
      edit-2
      6 months ago

      I agree. We’ve let the standards for what is good drop.

      I think it’s mainly because the “just works” mentality has become infectious among engineers. It’s one thing when just starting out, but as you learn more and gain experience you should care more.

      People do the designing and architecture and programming just because it all pays well, not because they have a love for the craft.

      I think the second, slightly less strong reason is because many engineers do not know how to effectively communicate with management when something will result in terribly written software and just do it anyway. Another skill I see less and less amongst my brethren.

      • ulterno@lemmy.kde.social
        link
        fedilink
        English
        arrow-up
        2
        ·
        6 months ago

        People do the designing and architecture and programming just because it all pays well, not because they have a love for the craft.

        True.
        I like programming and tend to pride myself in making good code, but when I see other’s attitude at work, it makes me reevaluate what I care about.

        Perhaps this is the reason of the memetic difference between corporate code quality vs OSS code quality. When I contribute to Open Source (at least to other’s projects), I see myself try to be as considerate as possible of multiple factors that I wouldn’t even care of at work.

      • Mikufan@ani.social
        link
        fedilink
        arrow-up
        1
        ·
        6 months ago

        Yeah, shure but it can’t actually program, it can give some code pieces and help find errors but so can a informed internet search as well.

        If your colleague is worse… Thats a different story.

    • masinko@lemmy.world
      link
      fedilink
      arrow-up
      0
      ·
      6 months ago

      Kind of will. There are already templates on demand for things like generating unit tests as you code. They’re pretty robust already, and have aside from a few things (or edge cases), I don’t have to do much code refactoring or fixing them.

      They already save me several hours a week from manually setting up full ones. Haven’t delved into other stuff they can do, but I’m sure it would only be more useful with time.

      I can very easily see companies looking at the time save and thinking “we can downsize”.

      • Mikufan@ani.social
        link
        fedilink
        arrow-up
        2
        ·
        6 months ago

        For openly available programs that might be true, but a programmer isn’t just writing code, “ai” is very bad at identifying and solving problems for example and Chatgpt is even getting worse at math. And internal programming languages, or, programs not meant for public eyes won’t be put in a AI in general.

        They are a tool, and will stay a tool hopefully forever, they are supposed to make shit easier.

        And those mentioned companies have already done that and fallen face first into shit. And rehired those that didn’t get a better job… In general its a shit idea to replace people with AI, cause its bad.

        Also “ais” are currently in a feedback loop and basically make themselves brain dead over time. Wich is for example one of the reasons why GPT and others get worse at math. Image generators are more obvious regarding that, when you feed the ai ai images, the fingers get worse.

        • jaybone@lemmy.world
          link
          fedilink
          arrow-up
          1
          ·
          6 months ago

          And the execs who made these stupid decisions get their golden parachutes and everyone else has been laid off.