Recent Content:
One of my passions as an artist is to embrace all the “happy little accidents”: The seemingly random accidents that end up creating something uniquely attractive. Every piece of mine has these, whether we’re talking about my modern AI-influenced works or my ancient pieces from so many years ago that they have been lost (like all my film photography in the destruction of the photo building at college).
I also enjoy Brutalism (the art style).
When crafting the current design I’m blogging on (duality), I utilized more architecturally-themed Brutalism (and some post-Brutalism for the typography/colors). I also created a number of 3d models when determining which direction I wanted to go with the 3d animation on the homepage.
I took one of the rejected 3d models, adjusted the camera in the scene, and exported a still image to use as a seed in an AI series on Brutalism.
The above image is what I used as my “seed” for my Stable Diffusion server. I then fed it, plus a handful of prompts, into the server setting the strength to .7; .7 is actually a good strength setting for this particular study just because it gives the AI just enough freedom to experiment, while still clearly taking it’s influence from the source image in nearly every result. Here are a handful of choices from the study.
Here is where you’ll see the reasoning behind choosing a .7 on the strength scale for this: Overall, the source image is very dark.. Aside from that rectangle in the upper-left.
The AI has no choices but to integrate that difference in luminosity into the composition. Everything else is more free-form, essentially..
..Except for one subtle key feature.
A specific rotation of one axis of the camera in the scene is consistent as well. Now, of course, we’re seeing some variations on lens distortion, perspective, and more, but those can be chalked up to the prompt itself (yes, I did use “Escher” as part of one prompt, but that only covers one of the above images).
This study on Brutalism in AI can also be considered a study in controlled chaos; throwing spaghetti into a wall, but only one noodle at a time and with a specific arm and wrist position, only subtly varying certain aspects of the throw itself.
Like what you see? I’ll share more of these pieces on my Patreon later, and will eventually select individual pieces for printing as part of my “An AI Dream in Analog” series.