I made 96 of these images tonight with Disco Diffusion, and all of them were... silly. Or weird.
It's a good example, to show that AIs can not do living things, humans, or animals. The faces are impossible, the bodies, silly.
But an artist could use this, and I uploaded 3 phases in this portfolio.
The 3 pictures are finished. The following 3 are the "partials", the 67%. The last 3 are the 36%. And they are really OK to find lines and lights, right? You just don't need all the pixels, just draw over it!
If you want to try:
Tutorial for Disco Diffusion
https://youtu.be/FA2MNG8D5x0
Then, for digital artists:
https://youtu.be/A0SbvjKw9sI