Đy Mokomi
@dymokomi Published May 5, 2023

What’s your background and how did you get into AI art?

My background is in visual effects. I have been working in the film industry since 2000, starting in the software and research side of the business at Alias|Wavefront. Before I left the industry, I worked as a visual effects supervisor for several years. My exit was prompted by the onset of artificial intelligence and my general interest in how people think. I studied neuroscience for several years before joining a startup that was working on a version of AGI. This experience gave me a glimpse of a different kind of creativity. Ultimately, I switched to the industrial design industry, where I have been for the past four years. I am responsible for numerous visualizations and for devising novel approaches to present ideas, which led me to diffusion models early last year. I began with Disco Diffusion, having missed the entire GAN era, and now I am heavily involved with custom tools that I write, such as MJ, SD, and more.

Petals of Luna” by Đy Mokomi

Do you have a specific project you’re currently working on? What is it?

I am currently working on a significant project at the intersection of generative art and AI. I have been immersed in this work for several months, which feels like years now. The convergence of these two computationally driven yet distinct areas of art is what fuels much of my technical interests at present.

What drives you to create?

There are a few factors, two of which are major and closely interconnected. I am fascinated by all types of communication. The fact that we can communicate at all is incredible; however, the bandwidth and fidelity are quite primitive. Visual mediums help address some of the shortcomings of communication, but in my opinion, the combination of written and visual has the greatest impact, provided that people are willing to absorb and process information. Ultimately, this fuels my passion for understanding intelligence. Communication is just a part of it, but in my view, it is a very important aspect.

The Quintessential” by Đy Mokomi

What does your workflow look like?

I can roughly divide all my work into two categories. One originates from a surprise, where I accidentally stumble upon a workflow, a striking image that works well as part of the prompt, or something similar. This becomes the seed of the final artwork. All thinking stems from this initial discovery and drives the technology I use forward. The other workflow (which I enjoy more) occurs when I have some external constraint, such as a Foundation World with a theme or a thematic challenge. This lack of control over the subject sends the design part of my brain into overdrive, and I feel it allows me to create my best work. For some reason, this constraint has to be external. I’ve tried to write out themes for myself, but it never worked, because I had control to change it — there was no external judge to tell me that I went off track.

What is your favourite prompt when creating art?

I know that prompts come to mind when people discuss AI art, but I’m afraid I might disappoint you, as I don’t really use many. I mean, I do because I have to obtain an initial embedding, but 99% of what I do is purely visual. In MJ, I use 3-4 images at a time (mostly self-painted textures, shapes, or collages), while in SD, I do a lot of work using aesthetic gradients and control nets with sketches. I have only two prompts (one for MJ and one for SD), and the rest is purely technical work.

Left is a raw output from SD which has nothing to do with my intention. Right is the output after using a custom trained aesthetic gradient.

How do you imagine AI (art) will be impacting society in the near future?

AI art, specifically, will undoubtedly facilitate personalized entertainment as the most apparent commercial outcome. Ultimately, I hope it will help to enhance our collective creativity. It enables many more people to ask the “what if” question.

Who is your favourite artist?

There are many artists I admire for various reasons. Having worked in design for some time, I can appreciate problem-solving as a form of art. Naturally, people like Dieter Rams and Naoto Fukasawa come to mind as prominent industrial designers. I love Jean Girard and Katsuhiro Otomo in the more illustrative genre. Reading about artists such as Richard Diebenkorn, Marc Chagall, and Mark Rothko inspires me greatly.

“Tentacles of Memory” by Mark Rothko

Anything else you would like to share?

One thing I keep coming back to is the fact that AI art is a tool. It all comes down to ideas. The important part is the story or message you want to convey. With generative tools, it becomes easier to explore ideas and assess their merit. However, without ideas and direction, the tools remain just that - lifeless pieces of code.

by @dreamingtulpa