• angelmountain@lemy.nlEnglish
    8·
    14 hours ago

    I’m a little bit confused. Does this mean more bugs were created because of using AI while coding previous versions, or did people use AI to find bugs that were already created before AI was used to create code?

    • mholiv@lemmy.worldEnglish
      254·
      9 hours ago

      People used LLMs to find bugs that existed before LLMs created code.

      LLMs are VERY good at finding bugs in existing code. That and C as a language is prone to memory bugs. Even before LLMs there has been a call to move to memory safe languages.

      • shortwavesurfer@lemmy.zipEnglish
        4·
        5 hours ago

        Ai has made me a btter bug hunter. I point an LLM at an application I like to use and give it a prompt like please scan this codebase URL for bugs.

        If it comes back with any, I then try to confirm them for myself to make sure it’s not hallucinating any, and if that bug actually does exist, then I can report it.

        I found one critical vulnerability and several high severity issues in an app that I enjoy using and reported them to the developer and got them fixed.