Full disclosure: I L❤️VE art (especially architecture), but I’m not really a creative person… …I’m a systems person. The extent of my creativity is that when I see something I think “how can this be improved?” (Not about Art. I’m so much NoT an artsy person that I wouldn’t presume to guess how to improve it). So my “innovative” ideas are never creative, never about creating something new, they’re about improving efficiency. And even then, it’s a matter of just doing the same thing in a slightly different ways (processes) and sometimes formal frameworks (of others). 👀 Ok, so today I working on a new cover for the group (which, btw, is about the extent of my artistic AI integration, and even this took me a while to come around to). This friggen tablet was almost the death of me. It seamed like every time the image was almost perfect, chatGPT would switch the tablet out to an angle I’d already told it (10 times) to *not* do. I was specific about what it SHOULD look like, but still Chat GPT would rotate it for a final rendering. Finally, when it was right to a point where it was almost right and returns were far past the point of diminishing returns. I asked Chat GPT: - To produce a list of all of my requirements and specifications before trying again - How can I make it so (it) automatically carry them forward? - To Give me a prompt that I can use at the beginning of any image generation project that makes this standard practice. Because I know this happens fairly often, I’ll do a quick little guide about it, but for now I wanted to share the prompt (it’ll be in the first comment). ——— Tell me below in the comments, do you have tips you use to keep image generation from wandering? I have the most recent advice and a quick cheat sheet I’ll post this week, but I’m open to additional tips :)