

Pushing as much of the costs of generated AI onto the contributor is 100% the right way to do this. There has been a substantial shift towards generated code. Which might be temporary depending on your belief system but we are in the world we are in right now.
Contributors using LLMs should have a burden of proof higher than that of traditional contributors. This includes tests, recordings, proofs, discussion, etc.


https://goreleaser.com/ does Rust now… But it’s binary focused but probably what the guy is doing…
For go Debian has: https://go-team.pages.debian.net/ I would imagine there is something similar for rust?
Being open source helps a lot, however he should probably source help for each distro. He is right that it’s a pain but I’m not sure his reasons are.