Sure, I get that. Been running text to image diffusion models on local hardware for a while. This is about the accuracy, specificity, and quality. It’s not a gimmick, it’s a threat… it’s also fun as hell.