What's this about?
The Rust project published an official policy on LLM use in contributions. The gist: it's fine to use LLMs to answer questions, analyze, distill, refine, check, suggest, and review — but not to *create* contributions wholesale. It also explicitly forbids harassing contributors over LLM use, banned or not.
What exactly is a project "LLM policy"?
A governance document telling contributors where AI is welcome and where it isn't. The distinction Rust draws is create vs. assist: a human must originate and own the work, with LLMs allowed as a helper (reviewer, explainer) rather than the author. It's an attempt to keep maintainer review load and quality under control as AI-generated PRs surge.
What's the community saying?
HN 109 pts, 70 comments.
- Enforceability doubt: if someone strips the em-dashes and bold text, how would anyone know it's AI-generated? The line may be unpoliceable.
- Approval of the anti-harassment clause — protecting people from LLM-use witch-hunts is seen as mature governance.
- "Rust already has AI code in its codebase" — some argue the horse has left the barn and the policy formalizes reality more than it changes it.
- The create/assist boundary itself is debated: where does "heavy autocomplete" end and "creation" begin?