I’ve recently started a new job that is much more analytically oriented. I’m going to try creating some shorter posts about odd little analytical projects I’ve been pursuing. All too often I simply don’t share these with anyone.

Recently, the folks at StabilityAI released their Stable Diffusion model, which can be used to generate digital images from text prompts. I got it up and running on my personal Mac following the instructions from this post. (PS - Thank you Ben Firshman for writing the post).

I’ve had a lot of fun with it - some friends have been sending me prompts over WhatsApp to run for them. I’m still learning how it all works but (perhaps unsurprisingly!) it runs rather slowly on my machine (14 inch Macbook Pro, M1 Max, 32GB of RAM).

Below are some of the fun pictures I’ve generated: Picasso horses, 1960’s travel posters (for skiing, to Mars, to the bottom of the ocean), 1950’s magazine ads for the iPhone, impressionist paintings of dancing in a park, and Hieronymous Bosch’s rendition of the NYC skyline, the Canadian parliament, the Mos Eisley cantina, and various Pokemon.

What would you create?