Museum of Prompt.

Since early 2024, I’ve treated this series as a daily studio practice. Each day I return to a portrait framework—not to document a single subject, but to search for a new face of art in the AI age. The “subject” is both human and machine, and the portrait becomes a site where intent and interpretation collide.

I’m drawn to glitches and prompt-induced errors. They introduce unexpected seams, misregistrations, and warped anatomies that behave like a new kind of texture—a digital pigment. What looks like a mistake often reads as feeling; these artifacts open pockets of meaning the same way brushstrokes once did. In this work, error isn’t a failure of the system; it’s my material.

Recently, I began coding a custom “texture machine” with ChatGPT-5—an evolving tool that lets me alter surfaces with a more personal touch. It extends the prompt into a tactile phase, where I can nudge edges, regrain skin, or re-time patterns. My aim is to invent an organic way of painting through a digital medium: to let code feel hand-made.

The workflow shifts between two entry points. Sometimes I start with my own photograph, layering prompts on top to reimagine the image as a new portrait. Other times I generate from scratch in Midjourney, then pass the result through my code machine to repaint and retouch—pushing the image until it holds tension between clarity and distortion. Across these passes, meme-like fragments, half-remembered faces, and algorithmic “slips” become narrative cues.

Museum of Prompt is a record of that evolving dialogue. Each piece is a snapshot of negotiation—between what I ask, what the model imagines, and what the code lets me feel with. It’s portraiture as co-authorship, where ambiguity carries the expression and the prompt becomes a brushstroke.