1. 10
    GitHub Copilot is Getting Sued ai build.typogram.co
  1.  

  2. 7

    I see there was discussion on this topic 2 weeks ago: https://lobste.rs/s/ukgknn/github_copilot_litigation

    1. 2

      Yes. This feels like a stunt.

      A real suit would sue a person or company for using code generated by CoPilot that copied their own code. MicroSoft would need to decide to jump into the fray (taking risk) or not (abandoning users of their tools).

      1. 3

        I think the lawsuit is meant well to regulate AI - it is uncharted law teritory. I don’t think Microsoft would just stop the product even if they lose. There are middle ground to reach an agreement on. Maybe add something to the terms and conditions that user consent on the usage when they submit code to Github public repo.

        1. 2

          I’m going to ask that you please read the prior thread, or any of the other multi-hundred-comment threads on other popular programming forums, when this lawsuit was previously announced. There’s not really anything new or useful to be said that hasn’t been said already. For example, GitHub’s terms of service have already been discussed extensively, along with what rights they require you to grant and how they require indemnification.

          1. 1

            Maybe add something to the terms and conditions that user consent on the usage when they submit code to Github public repo.

            That’s very tricky. If Copilot’s use is fair use then it’s fine to do it for anything but if not then it requires an explicit license. If a particular F/OSS license permits it, then it’s fine, but if not then you need additional rights. All copyright holders need to agree to this, so if you have accepted a PR from someone else then you can’t tick the box legally. If you tick it but don’t have the rights to do so and someone sues Microsoft, there may be something in the GitHub T&Cs that lets Microsoft sue you (you probably agree to indemnify Microsoft if you don’t have the rights that you claim you have) but actually doing this would kill GitHub as a viable product.

        2. 1

          Stop anthropmorphizing computers. Calling it “machine learning” does not mean it is learning in the human sense. Computers do not have semantic understanding, they are machines of pure syntax. Stop making these lazy comparisons of fundamentally different things.