• @hikaru755@feddit.de
      link
      fedilink
      English
      610 months ago

      Training and fine tuning happens offline for LLMs, it’s not like they continuously learn by interacting with users. Sure, the company behind it might record conversations and use them to further tune the model, but it’s not like these models inherently need that

    • niftyOP
      link
      fedilink
      English
      210 months ago

      You can train models of all kinds without disclosing anything personal about a user. Also see differential privacy

  • Beaver [she/her]
    link
    fedilink
    English
    6
    edit-2
    10 months ago

    I could use that!

    Update: it works fantastic and lets you switch easily to different AI models

  • Autonomous User
    link
    fedilink
    English
    -4
    edit-2
    10 months ago

    I don’t see how we can prove this. Paying them to also spy on us is bad but allowing them replace our software c/localllama with their service is even worse. My funds are better spent on local AI development or device upgrade.