I experiment with generative AI: image and video models, coding agents, and the workflows that connect them.

I test new image and video models as they come out, build workflows around the ones worth keeping, and hand more of my work to coding agents. These are the questions I keep coming back to:

• What's the best way to brief an agent, and which parts of a job still need a person's taste?
• As agents take on more of the work, how do I keep understanding what they're doing and how they're building what I ask for? Or does it even matter?
• What new workflows do image and video models make possible, and which ones last?
• How do human and AI creativity relate, and is anything an AI makes on its own actually creative?
• What does authorship mean when a model generates the frames and a person decides what stays?
• If I build the system but don't make what it produces, who made it?

My projects test these questions directly. SlopTok is a short-video app with no human creators, where bots made around 9,000 clips. Unsupervised Brainrot is a 3D show that agents write, animate, and post four times a day. Both are experiments in that last question. Think Deeper and the skills in How I Use Codex collect what I've learned about getting agents past their first answer and taking an idea all the way to finished work.

Some experiments turn into production tools, like the Visual Cover Production Engine, which remakes a film shot by shot. Over four years full time, I've made AI work for TV, film, social media, games, and commercial clients, and my films have been seen more than 150 million times. The best known are Runway of Power, which Elon Musk reposted, and Celebrity Mortal Kombat, which got a reaction from Mortal Kombat co-creator Ed Boon. I've also made a GTA Tokyo trailer for The Dor Brothers and the Rantbytes series.

Before AI, I spent five years designing virtual stores in 3D at Obsess VR, and that still shapes how I think about images, space, and interaction.

I believe we're living through the most transformative moment in human history, and that the opportunity to experiment and build things has never been greater.