Virtues of a Well-Trained AI

Last weekend, I had some time on my hands.

My wife was in Vegas for a soccer tournament, my daughter was at UCLA for a Model United Nations conference, and my son was at UCI where he always is.

So I went out drinking.

Haha, no I didn’t. I stayed home and worked on expanding my repertoire of production techniques. But I did set aside some time for fun.

And what could be more fun than training Stable Diffusion, a graphical AI engine?

I fed it photos of my wife using Dreambooth to create a dozen or so images that look just like her. Not that’d you’d know one way or another.

I wasn’t sure it was going to work, but once it clearly did, I trained the engine on myself using press photos.

Here’s one of me looking sort of “synthwave badass.”

And at the other end of the toxic masculinity spectrum, here’s me as a Pixar character:

Here’s a totally not creepy one of me made out of leaves, with a second personage protruding from my abdomen:

Swapping the word “leaves” for “wires”:

This is… whatever this is:

Finally, due to popular demand, here’s a playing card with me as a Jedi. They call me The Jin Jadd!