Strategy without delivery is just a deck
Essay, 2026-05-28
The AI market is sorting people into those who can talk about it and those who can build with it. The gap is where the value sits.
The AI market is quietly sorting people into two piles: those who can talk a good game about AI, and those who can actually build with it. The gap between them is widening, and that's where most of the value now sits.
I've spent fifteen years in rooms where the decisions are commercial, and the last two writing the code that ships the product. I keep noticing how rarely those two live in the same person.
Why the split is expensive
A strategist who can't build underestimates what's now trivial and overestimates what's still hard. Without being hands-on, it is hard to know how the important AI concepts actually work, especially given how fast the field is moving. A builder who can't read the room or work with clients solves the wrong problems beautifully. Both produce expensive misfires, just in opposite directions.
The test I'd apply
If you're hiring or buying AI help, ask one question: can they show you something they shipped this month, and explain a decision inside it they'd make differently now? Strategy is cheap. Shipped judgement isn't.
That's the thesis behind how I work and behind this site. The work is the proof.
More field notes