I build companies and I fly aircraft.

The through-line is narrower than it looks. I studied physics and math at MIT because I wanted the layer underneath the systems, not the systems. Then I spent a decade proving I could learn any domain by hand: Amazon twice, the Exchange team at Microsoft, a real-estate brokerage I ran with two salespeople that closed north of $80 million a year, a title-data business built on a graph of who routes dealflow to whom. At Opendoor I designed the AI coding interview, which sounds like a small thing until you ask what the old one was measuring.

Today I am co-founder and CEO of Closin. We build vertical AI for revenue, starting with RevOps for cybersecurity companies. The bet is one sentence: AI has collapsed the cost of knowing things, so every moat made of information is draining, and what is left is the human layer: judgment, trust, and who gets held to the number. Revenue is where that layer concentrates.

The flying is the same instinct. Commercial ratings in helicopters, multi-engine airplanes, seaplanes, and gliders. A private rating in a hot air balloon, which has no directional control at all: you choose an altitude and take the wind it gives you. I hold instrument ratings and I am a certificated ground instructor, which mostly means I have been the student recently enough to remember what the student gets wrong.

I read history compulsively and I play grand strategy games, which is the same hobby.

I write here about AI, markets, and history.

I am not trying to know the most. I am trying to be fastest to the next domain.