• LastoftheDinosaurs@lemmy.world
    link
    fedilink
    English
    arrow-up
    13
    arrow-down
    10
    ·
    edit-2
    10 months ago

    Hasn’t been an issue with GPT4. Last problem like this I had was over the summer using the free model.

    You’d have no way of knowing that someone was lying on Google, either. I use ChatGPT all day every day at work without an issue. It’s written thousands of lines of code for me.

      • LastoftheDinosaurs@lemmy.world
        link
        fedilink
        English
        arrow-up
        4
        arrow-down
        1
        ·
        10 months ago

        Mostly just simple Python scripts to automate stuff in a datacenter, but after the release of an assistant feature last week I’ve written a full-stack app using almost nothing but ChatGPT. The app is using FastAPI, Docker, and Nginx as the reverse proxy and for web auth.

        • Avid Amoeba@lemmy.ca
          link
          fedilink
          English
          arrow-up
          8
          ·
          10 months ago

          Yeah that makes sense. The success rate might fall off a cliff in more complex software projects. E.g. applications that require designs beyond 10 UML boxes with hundreds of thousands of lines, especially not written in JS/Python.

          • LastoftheDinosaurs@lemmy.world
            link
            fedilink
            English
            arrow-up
            3
            arrow-down
            2
            ·
            10 months ago

            I think my employer owns it. I didn’t write any of it myself; you could also ask ChatGPT to write the code for you. Start with a simple python script to connect to OpenAI, get it to print the response, then you can add the assistant API calls once that’s working. I knew I wanted to use Docker and Nginx, and I picked FastAPI from a list of 5 web frameworks it gave me when I asked how to serve the app.