• P03 Locke@lemmy.dbzer0.com
    cake
    link
    fedilink
    English
    arrow-up
    7
    ·
    2 days ago

    I generally favor local LLM usage, but I will sometimes use Kagi’s assistant tools. I trust them to throw out my data a fuckton more than I do Google.

    • UniversalBasicJustice@lemmy.blahaj.zone
      link
      fedilink
      English
      arrow-up
      2
      ·
      edit-2
      1 day ago

      A locally run LLM is a different story, one that I’m slightly more open to, but “trusting” any corp is how we got here. To me degoogling isn’t just about moving away from Google but about removing myself from any nonessential service and limiting the data I share with essential services.

      I also just plain don’t know what I’d use any LLM for.

      • P03 Locke@lemmy.dbzer0.com
        cake
        link
        fedilink
        English
        arrow-up
        3
        ·
        1 day ago

        I also just plain don’t know what I’d use any LLM for.

        PewDiePie has been on a tear lately with recommending everybody switch to Linux and recently experimenting with LLMs. He put out a legit good local LLM project that showcases the different kinds of things you can do with a model.

        My biggest LLM use cases have been:

        • A replacement for search, since you typically need multiple searches and multiple sources to get the information you need
        • Code generation, because it’s so good at producing entire fully-featured utilities or just let it spin away at implementing some new functionality
        • Helping me out with some random Linux error or difficulty, because support forums are fucking garbage
        • UniversalBasicJustice@lemmy.blahaj.zone
          link
          fedilink
          English
          arrow-up
          1
          ·
          1 day ago

          I think needing multiple searches and sources is a good thing, personally. Yes it takes more time and effort but it reinforces research skills. Performing the search yourself provides an opportunity to critically evaluate the information you’re finding and, if applicable, take into account any bias a source may have.

          Can’t comment on code generation.

          I suppose it depends on the Linux difficulty but reading through documentation and ArchWiki typically gets me enough info to make a more informed/refined search. You’re also likely to stumble across information you’ll want to know rather than just find an answer to the question you’re asking.

          • P03 Locke@lemmy.dbzer0.com
            cake
            link
            fedilink
            English
            arrow-up
            1
            ·
            12 hours ago

            Your responses are the equivalent of “I don’t use the internet, because nothing beats the experience of walking to your library and looking through the research material yourself”.

            Sure, I can do all of those things. But, sometimes I don’t want to spend hours going through the searches, bang my head against the wall with useless tech forums, and get frustrated with shifting through the results. The LLM can get it done in a minute or two, and I can still ask follow-up questions if I don’t trust anything, or want clarification.

            And yes, like Wikipedia, I still review the sources and review the code.