On shipping the infrastructure you need

I shipped about thirty open-source projects in March 2026. Rust CLIs, ML models, an agent framework, iOS apps, clinical tools. People keep asking why.

The short answer is that biotech runs on infrastructure that was built for a different era. Our lab tooling, our clinical software, our literature search, our way of coordinating experiments — too much of the tooling is legacy, and legacy is not just old code; it is the wrong set of assumptions compounded over twenty years. When you need to move fast on something like epigenetic reprogramming, you cannot wait for the ecosystem to catch up. You either build around it, or you build it.

I build it. And then I open-source it — not as charity, but because research infrastructure is a commons, not a moat. Our moat is the biology. The moat is manufacturing GMP Procyanidin C1 at scale. The moat is the clinic. Giving away the search tool that made our papers easier to write does not cost me a thing; it makes my peers faster, and faster peers produce more signal for me to learn from.

This is the same logic that makes WordPress worth more than most of the CMS startups that came after it. Commons-first is not a moral argument. It is a coordination strategy.