Caleb Gannon has been a professional artist for almost a decade, working in a variety of media ranging from acrylic on canvas to large scale procedural art etched into metal. With Algomancy, he ventured into the burgeoning medium of AI image generation to capture what the dream world of a computer might look like.
Each piece begins with a composition sketch to set the vision, tone and color of the image before undergoing an iterative process of rendering and refinement through inpainting, overpainting and photoshop. A typical workflow for one of the cards is shown below, and can take anywhere from 1-10+ hours depending on how specific the art needs to be for that individual piece. In total, Caleb spent about 3,000 hours creating the art for this game.
Here is a Video of the art process
We understand that with any new technology comes uncertainty, but we are doing everything we can to remain as ethical as possible while still exploring the exciting new potential that this technology offers.
To this effect, Caleb has been working with a custom built version of the Stable Diffusion 2.1 model, which was chosen because it is completely open source and most of its training data is in the public domain, like old museum scans, which he feels adds to the strangeness of the art in a unique way.
If you have any questions about the art processes used in creating this game, don't hesitate to reach out!