As I get a chance to play around with a lot more of these models. I find myself wanting to create D&D (or general fantasy and Sci-Fi themed images) generated from text prompt (think of what you see being implemented now in AI Dungeon).