This was a doodle I did in GIMP back when my laptop was brand new, in 2021(I dunno why I made it so dark). I put it into Stable Diffusion as a starter image, and played around with the settings.
Just as an FYI, AI art is insanely image heavy. I have a number of pictures related to this experiment, but only picked my favorites to share.
So … this is insanely revolutionary. I really want to nerd out over how this is going to totally change our digital landscape and expressions of creativity.
As I’ve said before, I think heavily in words (I am, at my core, a writer after all). With effort, I can conjure up basic images in my mind, but all of those little details escape me — my doodle is about as close to realistic as I can manage. This might not make much sense, but AI art lets me describe the shape of the words in my head, and translates them over into a picture for me.
It’s kind of cool to realize that I’m not as vague with writing descriptions as I thought I was.
We have Stable Diffusion installed at home, but my husband is still ironing out the technical wrinkles, so to speak. Despite that, I generated over 300 pictures in 24 hours, lol. The huge majority get discarded for coherency issues — missing limbs, deformed faces, etc.
Once we get it up and properly running, I’ll write up a tutorial.
For reference, this is what the same settings used in the last picture produce without the starting image: