I don't know if or how you've been using AI for coding, but Claude Code is almost certainly being used to develop itself, and I have a hard time imagining their devs not having the best setup possible to make it as easy as possible for Claude Code to iterate on itself with minimal oversight. Once a codebase is well structured and you supply a well thought out meta prompt with something like CLAUDE.md, then there's no reason to check the code output/changes beyond a quick skim.