If your agents are executing on assumptions anyone can Google, you haven't found your secret yet — and without one, you're just automating the median.
Zero to One is Peter Thiel's argument that competition is a kind of intellectual surrender. The real move is to find a secret — something true that almost nobody believes — and build a monopoly around it before the world catches up. He distinguishes copying (going from one to n) from genuine creation (zero to one), and insists that durable companies start by owning a small market completely rather than fighting for scraps in a large one. The book is a provocation as much as a framework: most people, he argues, have been trained out of believing secrets still exist.
The translation for an AI-native founder is uncomfortable and clarifying at once. Agents make execution cheap, which means your moat cannot live in execution anymore — it lives entirely in the secret. If your core insight is something a sharp competitor can reproduce with the same foundation models and six weeks of runway, you don't have a zero-to-one company, you have a feature. The question Thiel would ask today is sharper than ever: what do you believe about how a domain actually works that your market has not yet accepted, and can your agents operationalize that belief faster than anyone else can think to copy it?
- Cheap execution raises the bar for the insight underneath it, not the other way around
- own a narrow problem completely before agents help you sprawl into adjacent ones
- the secret is the product spec — everything else is implementation.
