[stingraycharles]

Yes, but as I said, it’s in a way the ultimate form of dogfooding: ideally they’ll be able to get the LLM smart enough to keep the codebase working well long-term.

Now whether that’s actually possible is a second topic.