Byzantine on Nostr: anyway to revise prompts or iterate on images generated by stable diffusion?