Have Almost Too Much Self-Belief

Self-belief is the first thing Altman tells ambitious people to build. He says it separates the people who compound into outliers from the people who quit. In How To Be Successful he puts it bluntly: conviction keeps you moving when the evidence hasn’t arrived yet, and it spreads to the talent and capital you have to recruit.

The hedge that makes it work. Altman pairs the delusion with its opposite: enough self-awareness to still update on the truth. Belief without a way to be right is just stubbornness, and he is explicit that the hard part is the balance, not the belief alone.

At OpenAI. Committing a lab to scaling neural networks in 2015, years before it was consensus and while it was widely mocked as unrealistic, is the principle applied to itself. The conviction preceded the evidence by half a decade.

Doesn’t transfer cleanly. Self-belief only reads as genius in hindsight. Survivorship bias means we mostly hear from the deluded people who happened to be right, while the same trait sank countless founders whose bets didn’t break their way. No essay can tell you which kind you are.