By skyforbes From when I was a kid I had a world that I created in my head that I've been building upon all the way into adulthood. I've written stories about it myself before, I went to an art college where I learned to draw professionally and drew things from that world professionally and still do to this day. But I never had a unified story. So I used chatgpt to help me organize my thoughts and generate a proper world for everything in my head to be put down on paper. I made approximately 120 pdf's that contain approximately 1,400 pages worth of world building notes. I used my own writing, my own art, and the writings of old dead writers like Edward Gibbon and Lawrence Stern, Livy, Ovid, and many others along with my worldbuilding notes, to train the AI agent to think like me. I created an entire 10,000 year history of this universe, and managed to over 2 months get the universe to mostly work without plot holes or continuity issues, which was a pain in the ass. I tried to create a unified history of the universe, but it proved to be a monumental task, possibly being upwards of 8 million words, which didn't seem possible with the current level of AI systems anywhere. So I had it hallucinate stories within this world.
One of the stories I guided it through a set of hallucinations is this 19,000 word novella. It took the better part of the past two weeks trying to fit within a narrative blueprint of a base story I created for it, using mainly Deep Research in sections. I also had to edit and format it after in pages. I have it linked here.
It is called "The Legend of Nirav."
I am unsure if people will be interested, and its okay if no one is, but this is a serious post about the limits of AI generation, and the future of what AI may be able to do, so please refrain from posting comments that are unserious.