Growth — learning, social, and material require a stimulus.
Risk is a critical stimulus.
Habits are powerful. The most important types of learning is both unlearning and behavior change. Bad habits impede both. Risk and fear are powerful catalysts to break habits.
Gurwinder: Failure is a mark of courage. It means you risked failure.
Naval Ravikant identifies 3 requirements for building wealth: specific knowledge, leverage, and accountability. Accountability is code for "risk". For commitment. He also said "It's not white collar versus blue collar. It's unleveraged versus leveraged."
Mike Bloomberg:
"Did I want to risk an embarrassing and costly failure? Absolutely. Happiness for me has always been the thrill of the unknown, trying something that everyone says can't be done, feeling that gnawing pit in my stomach that says danger ahead. I want action. I want a challenge. Even today, after toiling for 50 years, I wake up looking forward to getting in early, practicing my profession." I really appreciate that he used the word 'practice' there, "Creating something and competing against the best. It is a real high to be a participant rather than a spectator."
Optimization is efficient—but the wrong-kind of risky. Especially when it's premature. If your objective function doesn’t reflect a rich model of the world—and it rarely does outside narrow domains—you’re effectively selling options. You reduce variance on both the downside and the upside. Most optimization is anchored in downside understanding, but this creates an asymmetry: misspecifying the upside means you’re selling options too cheaply.
Zero to One lays out two distinct forms of progress:
These mechanisms are synergistic, but also at odds with each other. If everyone adopts the current best practice, we risk crowding out new experiments. Consider a small country that gets “invaded” by Starbucks, and doesn’t get a chance to develop local coffee culture. In some ways this is good, and shows the system working as intended. Starbucks has created a fantastic product, they’ve heavily optimized it, those benefits should be spread across as many people as possible. On the other hand, these “best practices” have to come from somewhere, and spreading them too widely carries the risk of premature optimization.
*Dwarkesh interviewing Tony Blair on his work advising world leaders:*
For any one leader you’re probably giving very sensible advice for their country. It’s a positive expected value. To the extent that limits variance and experimentation across countries in different ways to govern or different policies, are we losing the ability to discover a new Singapore?
It’s tempting to homogenize in order to spread good ideas, but we need to maintain enough heterogeneity to keep pockets of innovation alive so that there are new and even better ideas in the future.