Ventures

The Money Program is the distillation of twenty-plus years of thinking about capital allocation, expressed as a set of rules, playbooks and agent-driven research pipelines. It is how I run my own book, and it is the lens I bring to every portfolio conversation.

The thesis is simple: frontier AI lets a small team read more of the market, more accurately, more often, than a traditional research desk. The advantage is not speed — everyone has speed now — the advantage is the quality of the questions the agents are told to ask.

Reachproof turns the messy reality of creator reach into something a brand team can actually trust. Built with my daughter Eleanor, who runs the product and brought the insight that the market has been measuring the wrong thing.

It also happens to be a near-perfect case study for the solo-billionaire thesis: two operators, a Claude Code session, and a very clean agent stack, shipping faster than most VC-backed teams of twenty.

Inteligex was the long-running experiment where I learned, in expensive detail, that the edge in algorithmic trading is not the model — it is the combination of judgement, architecture, and honest accounting of your own mistakes.

It is not a live product today, but everything I do now with enterprise AI is shaped by it. It was machine learning based — and now, it has been superseded by a GenAI based thesis. The investor-operator instinct I bring to SOP++ Factory, to Maersk's wider AI roadmap, and to the ventures above — all of it has roots here.