Essays on venture studios, AI as leverage, operators, and the long game of building software that lasts.
Validation is not about collecting encouragement. It is about designing a test that can come back no, and writing down what no looks like before you run it.
Read article →If AI is the reason your product exists, your moat is a model you do not own. The durable version treats AI as the engine room for a problem that was already worth solving.
Read article →A company built to be sold and a company built to compound are different companies from the first month. Retention is the mechanism that separates them, and it is a product problem before it is a metric.
Read article →A studio is not a fund, so the pitch that works on investors works badly here. What a studio is actually deciding, and the four things that make a submission easy to say yes to.
Read article →Building and buying fail in opposite directions. The question is not which is better, it is which risk you are equipped to carry right now, and what has to be true before you commit either way.
Read article →A studio that keeps running its own companies has not built companies, it has built departments. How we think about the handoff, what the operator actually owns, and the part we still get wrong.
Read article →Occasional notes on venture studios, operators, and building software that lasts. No schedule, no filler.