Login
You're viewing the functional.cafe public feed.
  • Nov 12, 2026, 7:18 PM
    retooted Cunt Beanface 🧄

    Here's a hot take:

    I think that, for a hobby open source software project, rejecting AI contributions is the right thing to do on basically all axes, including the purely pragmatic ones.

    Sure it's better ethically and should be better legally, but even just in terms of better software made by happier people rejecting LLMs wins.

    They aren't going to improve code quality.

    Slop PRs suck up a lot of maintainer morale, so if you can easily reject them without going into a vibecoding/"responsible AI use" discussion that's a win.

    And, frankly, the crowd you attract by allowing AI is much worse than the crowd you reject by it. A lot of those who would not contribute without AI are pure sloppists, and those who aren't can still go down the vibe bobsled.

    You lose people as users by allowing AI, while I suspect barely anyone would stop using a thing because it has no AI involvement.

    What you give up is a dubious short-term gain in development speed, but even if that exists it's offset by maintainability, and if it's a hobby thing you don't need to move fast anyway.

    💬 3🔄 25⭐ 55