No one has yet to explain to me how “You can’t stop people from submitting codegen” is any different from “You can’t stop people from submitting code they don’t have the license to.”

You can totally add to your contributing documentation that the developer’s signoff on a commit implies they did not use codegen for it nor did anyone they received it from. There is exactly as little automated test for code that can’t legally be submitted as there is for LLM code, so the “Well, you can’t tell!” argument is bogus.

It’s not even that I want projects to adopt the policy, it’s just that the constant refrain of “You can’t /stop/ people. You can’t /not/ accept LLM-generated contributions.” seems so obviously wrong that it smells like something the speakers would like the listeners to just accept without questioning than an actual thing anyone believes.