The RADV driver, Mesa’s Vulkan driver for AMD Radeon GPUs, now has its own AI/LLM contribution policy on top of Mesa’s, Phoronix reports. Valve’s Samuel Pitoiset pushed it.
What it says
- LLM-assisted merge requests need a good understanding of the codebase, because “reviewing them has a real cost.” Maintainers may close them at their discretion.
- LLM-written issues are discouraged. Describe the bug in your own words.
- LLM-generated reproducers are fine if you’ve reviewed them and they actually show the bug.
Pitoiset’s reason: “We are starting to get slightly more LLM MRs by external contributors these days… it seems better to state this now before getting flooded.”
My take
This is the sane middle. Not a ban, just “you own what you submit.” Same theme as the kernel: AI is fine, wasting reviewers’ time isn’t.
God Bless, and May the Source Be with You!
