• mlg@lemmy.world
    link
    fedilink
    English
    arrow-up
    7
    ·
    1 day ago

    There was a thread on one of the tech communities about CS horror shows, and one of them was a guy telling a story about how Amazon had to completely redesign their smart home printer function to run through some monkeychain cloud pipeline because right before launch because they realized one of the libraries related to CUPS was AGPL, which is a blacklisted license at Amazon.

    The kicker was that the library also offered a lifetime corporate license for $100.

    Amazon redesigned their entire printing functionality stack to avoid paying $100 (or following AGPL lol).

  • Paragone@lemmy.world
    link
    fedilink
    English
    arrow-up
    5
    arrow-down
    2
    ·
    2 days ago

    WHEN you use AI to create a derivative of a project, in order to be exempt from copyright of that project,

    IT IS A DERIVATIVE: it isn’t exempt.

    I hope some court upholds this principle, explicitly, soon.

    Else the “oh, it’s a copy of your work, therefore you have no right to any license on it” gaslighting of the AI-companies will become the de-facto legal-standard.

    No, IANAL, but have looked-into law a few times, through the years, to get understanding of the rules on IP-protected work.

    _ /\ _

    • brian@programming.dev
      link
      fedilink
      arrow-up
      1
      ·
      1 day ago

      that Google/Oracle lawsuit settled pretty clearly that apis can be copied and reimplemented freely. this is just that but automated

      • spartanatreyu@programming.dev
        link
        fedilink
        arrow-up
        2
        ·
        1 day ago

        Not exactly,

        They said that they APIs were copyrighted, and that the way that Google had reimplemented the API fell under fair use doctrine (the original work was still copyrighted and they were allowed to use it without authorization).


        As for using an LLM to reimplement a program, you’re going to end up with a derivative work which can only be created/authorized by the original copyright holder. If you don’t have the authorization, you are infringing on someone else’s copyright.


        If however, you used an LLM to read a program’s source code, sort each word alphabetically then output the result, you’re going to have something that while based on an original copyrighted work no longer functions as a program. It has been transformed to such a degree that it is considered a transformative work, and thus falls under the fair use doctrine.


        TL;DR: Simply getting an LLM to regurgitate source code back out to you does not remove the licence of the original source.