I think it makes perfect sense for Zig to have their stand against LLM contributions while consumers of the compiler/Zig project overall use whatever code aids they like. Building a language is not a matter of churning out as much greenfield code as possible, but in careful consideration of whether or not some feature and its implementation fits coherently into the entire overall language. It's upstream of so much, and we now have decades and decades of examples where just letting rip with new additions renders a language schizoid and unergonomic. An LLM's tendency to "yes, of course, and," to any suggestion is not what a healthy language project needs, but it can be tremendously useful for someone employing a balanced and ergonomic language to generate products. I'm glad to see Mitchell keeping a cool head as the unfortunate tendency in so many devs to take sides and get dogmatic plays out yet again.