DeepDream Vision Quest began with a strange new development in machine learning—the surreal imagery produced by Google’s DeepDream algorithm. What fascinated me most wasn’t just the final output, but how the images slowly emerged from noise, like watching a dream take shape. Each time I tweaked the parameters, it felt like playing hide and seek with the algorithm. I began to wonder: What if these hallucinations came from a live camera instead?

I built a kind of magic mirror that hallucinated wildly until it detected motion. I turned this system into an interactive art installation running on a custom hardware rig powered by the Python/OpenCV code I developed specifically for the task.

I was invited to exhibit at the LAST Festival, the California Academy of Sciences, the Stanford Linear Accelerator and a dozen other venues. The project led me to teach a series of popular workshops on generative art in San Francisco.