GitHub denkt über einen Kill-Schalter für Pull-Requests nach, um KI-Abweichungen zu stoppen
https://www.theregister.com/2026/02/03/github_kill_switch_pull_requests_ai/
GitHub denkt über einen Kill-Schalter für Pull-Requests nach, um KI-Abweichungen zu stoppen
https://www.theregister.com/2026/02/03/github_kill_switch_pull_requests_ai/
14 Kommentare
I’m somewhat skeptical that Microsoft will allow them to do anything effective.
Giving the repo owners more control can’t hurt. But given Copilot is GitHub’s entire business model at this point… I can’t imagine Microslop letting this get implemented. Implementing this is bad optics for AI given they are trying to shove it in to every conceivable product
Github? The Microslop company? That Github?
Microsoft – Solving problems we created!
Let’s put some more AI into it, namely to check if the contribution is low quality or not
I will create a merge request with Copilot to implement this feature 👍
Oh man. This is 100% a precursor to Microsoft offering a paid AI agent that specializes in AI Slop and “code review review” that you attach to your repo.
Soon, someone will vibecode a Git repo host just for OpenClaw agents to make their own Open Source software
Here is a list of problems the article lists associated with AI. As I read this, none of this is specific to AI and all of these can be addressed by having a robust test suite. If you are relying on humans to understand the whole codebase to make changes, whether they be reviewers or submitters, you’re already screwed.
The solution here is going to be a combination of better automated testing of code as well as probably using AI tools to do some review of PRs and flag potential issues for the submitter before they submit, as well as to help the person merging PRs do so a better job.
>Review trust model is broken: reviewers can no longer assume authors understand or wrote the code they submit.
>AI-generated PRs can look structurally „fine“ but be logically wrong, unsafe, or interact with systems the reviewer doesn’t fully know.
>Line-by-line review is still mandatory for shipped code, but does not scale with large AI-assisted or agentic PRs.
>Maintainers are uncomfortable approving PRs they don’t fully understand, yet AI makes it easy to submit large changes without deep understanding.
>Increased cognitive load: reviewers must now evaluate both the code and whether the author understands it.
>Review burden is higher than pre-AI, not lower.
So they will stop competitors from scraping their data while still using it to train their own models.
Soooo, are they stopping windows 11 pull requests?
Bit late no?
No need actions are already on fire so nothing can merge anyway
even Microslop knows that source control tools and the open source initiative are the only things keeping it in the good graces of a whole lot of parties.