Because it’s Facebook

That’s fine for you but what about others working on the same codebase?

Also people act like only AI code is bad quality and full of bugs. I’ve seen enough legacy codebases, and my own side projects, to know that AI does not have a monopoly on slop. We humans are perfectly capable of that ourselves. It’s the code review process that prevents that being merged, AI or not

I think it’s safe to assume that any public code on the internet is already in their training dataset.

At this point anything that is public in general, probably all the posts in the “fediverse” have been taken already

I think the Fediverse still suffers from the “screaming” issue, but some strategic mutes and hiding boosts from users, plus never looking too long at the “all” sections and it’s pretty nice.