Raw LLM Responses

There’s a good amount of people in the comments pointing out that the story is really dry and / or sucks. This is definitely true, but keep in mind that even 6 months ago, it was basically impossible for AI to pull anything like this off. If any of you had the opportunity to use AI Dungeon’s dragon model, which was based on gpt3 (until it wasn’t) it was decent at writing, but it was horrible at managing characters and events. Pronouns would regularly change, the AI couldn’t keep track of who was speaking to who, and it would hallucinate something new that made very little sense every other paragraph. It basically read like a really bad trip. Even services like NovelAI with their own fine-tuned models struggle to create a model that can keep track of setting and characters, despite the fact that they provide features to keep important information in the model’s context window and are designed from the ground up for writing. Anyway, this was an awesome project! This is the kind of thing that brings me back to this sub. I hope human writers don’t become obsolete in six months! lol