• sytone@lemmy.world
    link
    fedilink
    English
    arrow-up
    0
    ·
    26 days ago

    I think some of the problem is the term vibe coding. If I want a small app to help something specific I used to look at it and go is it worth writing it for the return most of the time I was just too busy.

    Now I sit down and write up a small document outlining what I want why I want it and the outcome typically I’d use AI then to review it and give any suggestions that I may not have thought of and use that for the seed for the application.

    That is the document I use with an agent to actually do the code generation. I push that to a local GIT instance and the document lives with the code and then gets iterated overtime and then I typically use the issues list to add improvements and fixes to it again iterating through the issue before I think about letting an agent do it

    To me vibecoding is just putting a sentence in and hoping for the best I like a little bit more structure than that.

    Again, most of this is for personal use so I don’t share it widely a few friends have access to some of the stuff I’ve written because it helps them. I only ever released one AI first project on GitHub and I probably won’t go any further than that.

    • Sl00k@programming.dev
      link
      fedilink
      English
      arrow-up
      1
      ·
      26 days ago

      I think some of the problem is the term vibe coding.

      Spot on. Most engineers at companies are using LLMs nowadays but when we use it in our daily lives it becomes “vibe coded”.

      I’ve started using the term agentically built or agentically driven instead.