One of the most fun and unexpected explorations I took on with generative AI was music. I became curious — what would an endless stream of AI-generated music look and feel like? Could it be interesting enough for people to actually listen to? More importantly, could it feel intentional — not just noise for the sake of novelty?

That curiosity led to Quasi Music, an experiment in building a simple, ambient listening experience powered by generative models. The interface was minimal by design — tap to listen, zone out, or stay engaged. It lived as a standalone page but integrated directly with Quasi Market, where users could also generate their own niche music styles.

The response surprised me. Within a month of launch, over 11,000 users tuned in, and more than 17,000 minutes of AI-generated music were streamed. Users described the experience as calming, weirdly addicting, and oddly helpful for focus. The experiment validated a key insight: even abstract AI use cases can connect if the interface feels lightweight, curious, and friction-free.