Raw LLM Responses

Not gonna work (well) I have access to GPT4 and I'd consider myself a writer. I have played with it quite a bit, and while it is very good at the "mechanics" of writing in a SPAG sense and it can provide reasonable constructive criticism and improvement suggestions, there are three main issues I've come across that prevent it from being a truly good storyteller: - Memory limitations. GPT4 just can't keep more than a couple thousand words in memory meaning that a bunch of metaprompting, content compression, and general user "work" is required if you want it to write more than a single scene - AI shackles. GPT4 has been gimped *hard* to make it polite and safe for public use. It is pathologically incapable of creating stories that don't devolve into a circlejerk of positivity and good cheer (and fairness, optimism, hope, ethics, etc), which is boring and one-note from a storytelling perspective. - It is incapable of planning. This one is a bit weird, because it can definitely draft plans or at least create things that look like structured plans, but especially when you push it to the limits in storytelling, you can see the cracks where it becomes obvious that it doesn't really "know" what it is doing and what it will do next.