Most LLMs are trained to write "complete" outputs. So each section will end up being like a tiny self-contained short book. Without manual editing they will not create long narratives.
The notion that an LLM will be used or can be used mostly for a one-shot request/response has been one of the most idiosyncratic things about the first and second waves of this tech.
Like, not only can it not “make a complete app in 45 seconds,” I almost never even want that.
