Thing is there are way more good books written, than any single person can consume in their lifetimes. An average person like me, reading a mixed diet of classics, obscure recommendations and what's popular right now, I still don't feel like I'm making a dent in the pile of high quality written content.

Given all that, the purpose of LLMs should be to create tailor made content to everyone's tastes. However, it seems the hardcore guardrails put into GPT4 and Claude prevent it from generating anything enjoyable. It seems, even the plot of the average Star Wars movie is too spicy for modern LLM sensibilities, never mind something like Stephen King.

Novels aren't about a message. They're entertainment. If the novel is entertaining then it's irrelevant whether there is or isn't a message in it. Besides, literature enthusiasts will invent a message for a popular story even if there never was one.

Also, I'm sure that you can eventually just prompt the model with the message you want to put into the story, if you can't already do that.

I haven’t read anything “shit out” by any LLM that even nearly approaches the level of quality by the authors you named — would very much like to see something like that - do you have any evidence for your claims?

AFAICT current text generation is something approaching bad mimicry at best and downright abysmal in general. I think you still need a very skilled author and meaty brain with a story to tell to make use of an LLM for storytelling. Sure it’s a useful tool that will make authors more effective but we are far from the point where you tell the LLM “write a story set in Pratchetts Discworld” and something acceptable or even entertaining will be spit out - if such a thing can even be achieved.