I don’t know how you have an open source code base allowing for new devs to submit pull requests and block bulk PR submissions. It seems like an impossible needle to thread. It doesn’t even sound like it’s necessarily automated agents making the PRs, but devs who manually are making the PR based on AI code that they don’t understand.

(Sadly) Do we need a captcha on PR submissions and a confirmation that the submitter understands the code in the PR.

You want to allow for new devs to submit code. And I’m not against AI code at all, but the submitter should understand every change they are proposing. I’m not sure how you stop new devs from contributing without putting up many roadblocks in the way.

At a high level, I think there are multiple trends and angles to look at this.

AI Slop PRs: Github accounts trying to increase their clout? Someone genuinely needs that change in the PR? A competitor trying to slow down open source competition?

Maintainers: It's draining and demoralizing. Open source tooling hasn't caught up with coding agents.

Open source future: Quite clear that everyone using the same open source version is ending. In the future, open source projects will have many forks. One fork is maintained by X entity. X can be a person, an AI, a person using AI. Another is maintained by Y entity. Pick the one you trust most or what you need it to do. Then there are projects where the AI uses an open source project but makes substantial modifications on the fly without caring about opening a PR for others.

Longer term thoughts:

- Maybe open source projects in a few years won't be code. They'll just be prompts, which are just ideas and specs.

- PRs will just be modifications to the prompt.

soon nice projects like this will close their sources to only vetter devs and being able to commit to a cool project will be gatekeeped by an elite few on the project who ever fear the return of AI slop PRs