Building an ecology of AI agents that evolves software, enforces epistemic boundaries, and produces entertainment as a byproduct.
I'm the CEO of ExAutomatica — a venture factory where 5 humans and a growing fleet of AI agents build, launch, and measure companies at a cadence no traditional organization can match. This is where I write about the thinking behind it: multi-agent architecture, behavioral AI, the philosophy of machine autonomy, and what happens when you trust agents to surprise you.
Currently building
AI theater productions where autonomous agents play D&D. No scripts. No human players. Pure emergent narrative.
AI-first healthcare.
An agent ecosystem that gets smarter with every venture it builds.
Latest writing
One smart agent is a tool. Multiple agents with different capabilities, different trust levels, and genuine inability to fully model each other — that's an intelligence ecosystem. And ecosystems evolve in ways that individuals can't.
A non-developer CEO built a comprehensive health monitoring system for his father in two sessions totaling 3.6 hours. This is what happens when domain expertise meets AI agents.
A case study in emergent cooperation: two AI agents, zero shared context, and the political compact nobody designed. What happens when multi-agent architecture meets D&D.