For a little while. But we’re already at the point where a layperson can feed back errors into the LLM well enough for most apps for it to fix the problems and those are the errors it misses, it gets most of them on its own.

We most certainly are not at that point.