Xü

Directeur Artistique

Describe yourself as a prompt.

French AI creator working at the crossroads of art direction, image-making, video, 3D, and code. A curious, demanding profile, allergic to “AI slop,” with a clear preference for building custom workflows rather than buying into magic tools. Cinematic aesthetics, attention to detail, open-source culture, and a fixation on the concrete: turning AI into visual language, method, and useful production.

What led you to AI, to the point of making it both your field of creation and of transmission?

I got into it through a pretty natural mix: image, video, 3D, and code were already part of my playground. AI simply connected all of that. I started early, with Stable Diffusion, ComfyUI, and open-source tools, first out of a creator's curiosity, then out of a need to understand what was behind the image. Very quickly, it became as much a production tool as a subject to convey: demystifying, building workflows, showing that AI isn't magic, but a new subject to learn.

How has AI changed the way you conceive, imagine or produce your work?

AI doesn't replace my imagination; it mainly changes my exploration speed. It allows me to test visual avenues faster, to prototype an atmosphere, an image, a narrative, or a workflow before refining them with an art director's eye. My work becomes less linear: I design, I produce, I correct, I document. AI also forces me to be more precise in my intentions.

The misconception about AI you most often debunk in projects or training?

The idea I most often deconstruct is that AI is magical just because you're using the latest trendy model. In reality, good usage relies less on raw power than on intention, workflow, mastery, and the ability to remain independent. The real question isn't "what's the strongest model?" but "what's the sufficient, manageable, and replaceable model?"

If you could only keep three AI tools for the rest of your career, which would they be?

ChatGPT to reason, ComfyUI/Stable Diffusion to produce images, Ollama to keep a local and independent backup plan.

Where do you draw your inspiration and the substance you pass on when you step away from AI?

In cinema, photography, science fiction, video games, old computers, forums, making-ofs, and people tinkering on their own. I'm fueled by everything that shows what goes on behind the scenes: how an image, a machine, an idea, or a method is made. And the material I share comes a lot from the field: my tests, my failures, my classes, my students, my clients, and this desire to make understandable what seems reserved for insiders.

What are you working on right now?

On the creation of my own local content creation interface (audio/image/video), entirely driven by my EGX workspace. A kind of ComfyUI customized and adapted to my use.