Ladybird, the from-scratch browser project led by Andreas Kling, no longer accepts public pull requests.
The core problem Kling identified:
- Effort used to signal intent. A large, well-structured patch implied someone understood the codebase and cared about the outcome. That was a reasonable filter for good faith.
- LLMs broke that filter. Anyone can now generate a substantial-looking patch with minimal understanding of what it does or why.
- Accountability is the actual requirement. Kling’s standard: the person introducing a change must be the person who decides it belongs in the project — and who answers for the consequences.
Kling’s framing is precise. He isn’t banning AI-assisted code. He’s saying the old open contribution model assumed a correlation between patch size and contributor investment. That correlation no longer holds. A 500-line PR generated in 90 seconds carries the same review burden as one written over a week, but the submitter may not be able to defend, debug, or maintain it.
This matters because Ladybird isn’t a hobby project anymore. Kling has spent 15+ years on browser engines, including work on WebKit, and Ladybird is targeting real users. Shipping a browser means every merged patch is a liability — security, performance, correctness. The project can’t absorb review costs for drive-by contributions where nobody stands behind the code.
NOTE
Kling didn’t frame this as “AI code is bad.” His argument is about responsibility: who owns the change after it lands. That distinction matters for any maintainer thinking through the same problem.
Other large open-source projects haven’t gone this far yet, but the pressure is identical. Maintainers across the ecosystem report rising volumes of low-context PRs that look plausible but require disproportionate review effort.
Ladybird’s move is the sharpest response so far to a problem every major open-source project now faces: contribution models built on human effort signals don’t survive cheap generation. Expect more projects to follow with their own versions of this policy.