[JohnnyB]

I haven't used SD, but I'd been using diffusion models for a few months and have been using MJ pretty much since it became available. Lots of fun to work with these and to play with their potential and boundaries.

Using MJ I recently tried doing a long series of prompts based on the stem "the face of ...." and slotting in words like 'God', 'poverty', 'a CEO', 'an Indian' etc. just to see what the outputs for these would be in accordance with the content of the training data sets. It was an interesting exploration to see when the faces were male or female, or when they had different racial features etc.