What should a twenty-year-old do to be successful?

On August 19, 2026, one of my younger friends and I ran our fastest miles of the day on the track. Mine was 7:19. His was about 6:15. I was still gasping for air when he asked, out of nowhere, what a twenty-year-old should do to be successful.

I grew up in Nepal and came to the United States at thirty. That was ten years ago. My first answer was blunt. You are in a better institution than the one I came from. Do not screw up, and you should be fine.

We ran one more lap. Then I told him I would write him a letter. This is that letter.

When someone asks what to do, most answers arrive as a list of things to do. I want to give you an ordering instead, because some decisions have to be settled before the others become worth thinking about.

The question is already a filter¶

If you are asking, you have already done part of the work. Asking implies searching, and searching implies you are screening options out. It is selection bias running in your favor.

Two rare events have to line up. First, you have to form a question worth answering. Second, you have to find the right person to answer it. Neither is common. You have done your part by asking. I am doing my part by writing this letter with care.

Survival before strategy¶

Warren Buffett once described his partner Charlie Munger’s method this way: “He likes to invert. He says, all I want to know is where I’m going to die so I’ll never go there.”^[1] Charlie Munger took the move from the mathematician Carl Gustav Jacob Jacobi, whose working rule was “Man muss immer umkehren,” which means invert, always invert.^[2] Hard problems are often easier read backward. Instead of asking what produces success, ask what reliably produces failure, then remove those things.

Some failures are known ahead of time. Driving without a seatbelt, unprotected sex, drinking without limits, treating people badly. These are known bads. You do not need new information to classify them, and nobody has to run the experiment personally. Asked at a law school commencement what a young person should avoid, Munger named sloth and unreliability.^[3] Avoid bad decisions consistently. Later I will talk about how to pursue good decisions with persistence.

What sets this class of problem apart is not the size of the loss but whether the loss can be undone. A car crash without a seatbelt does not cost you one outcome. It removes every future decision from the table. For these bad decisions, avoidance is not fear. It is leaving yourself something left to choose tomorrow.

Wild problems¶

Another class of decisions I call good problems. Russ Roberts calls them wild problems: whom to marry, whether to have children, where to move, what work to give your life to. Measurement and calculation cannot settle them. The right lens, he argues, is a life well lived, not day-to-day optimization of the kind that fills productivity blogs.^[4]

The outcome of a good problem is usually unknown, but pursuing it can still pay. This line, often credited to Socrates though the attribution does not hold, puts the wager this way: “By all means, marry. If you find a good wife, you will be happy. If not, you will become a philosopher.”^[5] Either way, you leave with something.

Brian Christian and Tom Griffiths, in Algorithms to Live By, write about choosing under uncertainty so that later regret stays small.^[6] Jeff Bezos gave the same idea a personal form he called the regret minimization framework: project yourself forward to age eighty, ask which choice you would regret not having made, then act so that each year you add fewer new regrets than the year before.^[7] BJ Miller, a palliative care physician, puts it in one line: “Loss is one thing and regret is another.”^[8] People more often regret what they left undone than what they tried. For good problems, then, do not build a pile of regrets. Be thankful you have a die to roll, while many people may never get one.

The flourishing test¶

For a wild problem, Russ Roberts says ask one simple question: will I flourish if I pursue this? Your gut usually answers before your head does.^[4]

I also find in my own life that working through a wild problem is how flourishing happened to me. Flourishing means ending up with a talent you did not have before. I am using the word the way Edmund Phelps uses it: meaningful work, self-expression, and personal growth, rather than material comfort alone.^[9] Phelps also argues that good institutions make that kind of flourishing more likely, which is another way of saying the room you stand in matters as much as the effort you bring.

To humblebrag, I did not know how to speak English, then I learned. I did not know how to write, then I learned. I could not stand in front of a room and give a long presentation, and that one was nerve-wracking, until it was not. Speaking, writing, thinking, and processing were all acquired the same way, by entering a problem whose outcome I could not see. I did not do that alone. I have been blessed by many people who nurtured and supported me, often people I met by chance, with a weak tie or no prior connection at all, whom I would otherwise never have met. Their kindness opened doors for me.

A wild problem worth entering hands you a capability even when the result disappoints, and that is what makes its downside survivable. A decision that pays nothing when it fails is not a wild problem.

How to enter a wild problem: stochastic tinkering¶

Knowing which problems to enter is not the same as knowing how to enter them. I used to call the method trial and error. After reading Nassim Nicholas Taleb, I call it stochastic tinkering.^[10]

Stochastic means involving chance. Tinkering means messing with things in a hands-on, decentralized way. Together, stochastic tinkering is trial and error done many times, with the loss on each try kept small. What makes it work is not the trying itself. It is the shape of the payoff. When the upside is larger than the downside, the average payoff is positive, and Taleb calls that exposure convex: uncertainty helps you more than it hurts you.^[10]

Even a bet with a positive expected value can ruin you if you stake everything on one try. That path is gambler’s ruin: keep betting your whole stack, and chance will eventually wipe you out.^[11] The Kelly criterion is the rule that keeps the stake a fraction of what you can afford to lose, so one bad draw does not end the game.^[12] Claude Shannon’s rebalancing idea, often called Shannon’s Demon, and what I like to call Shannon’s Angel, works on the same principle in markets: systematic rebalancing turns volatility into a source of growth rather than only a source of fear.^[13] These look like investment tricks. They are also rules for deciding under uncertainty. They cap the downside and leave the ceiling open. I teach the same ideas in my optimal-betting notes.^[14]

In practice, that means running many small experiments instead of one big bet. Take a class online to really learn, build a short project, and ask for a coffee with someone two steps ahead. The key takeaway is to keep the cost of being wrong small enough that you can try again next week.

What I like most about stochastic tinkering is that it is the same rule as a wild problem, just smaller. Avoid the known bads so you survive. Take up good problems so you flourish. Run small experiments so life stays interesting.

Do not confuse randomness with luck, but engineer the luck¶

In Fooled by Randomness, Taleb argues that mild success can come from skill and hard work, while huge success is often skill and hard work compounded with luck. So please do not read your own huge success, or someone else’s, as talent alone.^[15] I think that not being fooled by randomness gives you three things at once: a kinder reading of other people’s success, a humbler reading of your own, and more compassion for your failures.

However, you can still engineer luck, and persistence is part of that work.^[16] Luck can be worked on, but only in layers, and these layers are not equally available. James Austin sorted chance into four kinds. Chance I is blind luck, the country you were born in, the family you got, and it often outweighs everything built on top. Chance II favors those in motion, curious about many things and willing to experiment. Chance III favors the prepared mind, the knowledge that lets you notice a live opening when it appears. Chance IV comes from a distinctive path, when work finds you because of the odd combination of skills and interests that only you have.^[17] Trevor Noah puts the same map in plainer speech: lucky people create more motion in their lives, notice what shows up because they are paying attention, and become the kind of person opportunity can find.^[18] Jason Roberts gives the second layer a usable form: luck equals doing times telling, so if either side is zero the area is zero.^[19] In domains where results follow a power law, most attempts look like failure and a few carry almost all the return, so persistence that looks unreasonable from the outside is often the price of staying in the game long enough for a break to land.^[20]

When the action set is empty¶

Some problems leave you almost nothing to do. A loved one is sick, and no plan of yours can fix it. These are the third class of problems: bad problems. You are completely vulnerable because the outward set of actions is empty. I have relied on three responses, and none of them is a tactic for changing the outcome.

The first is demanding prayer. Do not be anxious about anything, but present your requests through prayer, petition, and thanksgiving, and the peace that follows outlasts your understanding of the situation.^[21] Ask as Matthew’s prayer asks, to be kept from temptation and delivered from evil.^[22] Ask as Jabez asks, to be kept from harm.^[23] Demand as a child of God, with prayer and petition.

The second is amor fati, love of fate: the practice of taking on everything life brings, including what you never planned for. A close cousin is right acceptance, what some Buddhist and Hindu traditions call samyak swikar.

The third is metanoia, a deep change of heart and mind that alters how you see your own life, with healing and growth. It arrives late, usually after the hard part is over. Be open to it.

These three do not alter the event. They set what the event does to you, which is the only variable left.

Compressed to one sentence: avoid what would end your options, enter what will teach you something even when it fails, keep each attempt small enough to survive, judge success without confusing luck for talent alone, and accept what you cannot move.