you are viewing a single comment's thread
view the rest of the comments
[–] 2 points 2 days ago (1 child)

I guess that sounds reasonable. I just don't see how a strict no AI policy can be reasonably enforced, if somebody is hell bent on shoving AI slop into your codebase they're gonna find a way, it's not like it's particularly hard to make it look like you actually wrote the code instead of the AI.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 2 points 2 days ago (1 child)

    the "cannot be enforced" argument always felt like concern tolling to me. for things like licences, anyone can just copy code from other incompatible projects and claim it's theirs. every policy can be abused and i don't see why this is different.

    if someone is hell bent on adding ai code (or any other code, like incompatible licenced code or even backdoors) they will have to make signifficant effort to mask the signals to bypass the review. for everything else it will be a good guideline on how to act when that situation happens instead of depending on the individual maintainers opinion

  • source
  • parent
  • hideshow 2 child comments
  • [–] 2 points 2 days ago*

    they will have to make signifficant effort to mask the signals to bypass the review.

    I guess it depends how significant the effort will be, obviously if somebody just shits out out a 50k line diff with an AI it's gonna be really suspicious right off the bat but a smaller, more contained diff isn't going to immediately raise any flags. Eventually somebody is going to make it really easy to make the AI not sound like an AI at least not as obviously as it does now.

    I think eventually maintainers of high profile open source projects are just gonna have to whitelist contributors and just reject contributions from the general public, I wouldn't be surprised if we started seeing interview processes being set up for open source contributions.

  • source
  • parent