Naval
Learn to sell. Learn to build. Master both and you will be unstoppable.
A calm mind, a fit body, a home full of love. Cannot be bought. Must be earned.
Specific Knowledge
Knowledge you can't be trained for. If society can train you, it can train someone cheaper.
Found by pursuing genuine curiosity — the overlap of what fascinates you, what you'd do for free, and what the world needs. Arm yourself with it, then compound through reputation.
Leverage
Three forms. Only the last doesn't need permission. Find yours.
| Type | Permission | Scale | Example |
|---|---|---|---|
| Labour | Yes | Linear | Traditional business |
| Capital | Yes | Nonlinear | Finance |
| Code + Media | No | Infinite | Software, content, agents |
Play long-term games with long-term people. Compound interest applies to relationships, knowledge, and trust — not just capital.
Build Then Sell
Silicon Valley's model: pair a builder with a seller. Wozniak/Jobs. Allen/Gates. The rare individual who masters both creates the category.
Engineers can learn marketing easier than marketers learn engineering. Building is harder to acquire later. Selling compounds over time.
The principal-agent problem: when the builder doesn't sell, someone else decides what gets built. Act like an owner.
Judgment
Wisdom applied to external problems. Can't be automated — requires accountability and skin in the game.
Engineer lessons into systems that enforce and improve standards through clearer thinking and better decisions.
Context
- Science — First principles thinking
- Potential — Leverage what's possible
- Memes — Ideas that spread without permission
- Smart Agents — Code as infinite leverage
- Build and Sell — The original essay
- Systems vs Goals — Conversation with Scott Adams
Questions
If specific knowledge must be discovered through curiosity rather than trained, how do you distinguish genuine curiosity from manufactured interest designed to game the system?
- Code and media are the only forms of leverage that require no permission — what breaks if that permission assumption changes (e.g. AI generates code faster than you can)?
- Naval argues judgment can't be automated because it requires accountability and skin in the game — does that hold when AI agents can hold capital and face real consequences?
- If build and sell are the two skills worth mastering, which comes first when you have nothing to compound — and does the answer change depending on the type of specific knowledge you have?