How I Think

I think in patterns and forces.

I grew up in a culture where effort explains everything: if something didn’t work, you didn’t want it enough. I respect the ethic, but it isn’t true. Effort is a weak predictor of outcomes. The strong predictor is whether you are aligned with the underlying forces or fighting them. The swimmer’s stroke matters less than the current. A great deal of what happens to people and companies is explained by currents nobody inside them bothered to map.

Much of what passes for thinking is reasoning by analogy: copying what already exists with the serial numbers filed off. Analogy is cheap and usually adequate, which is why it is worth so little as an edge: anything it can see, everyone can see, and it is already priced in. Seeing clearly requires reasoning from first principles, stripping a problem down to what is physically and structurally undeniable, then building back up while ignoring what everyone else assumes. Physics keeps discovering counterintuitive things because it refuses to argue from consensus. I try to hold businesses and systems to the same standard.

The laws of nature work bottom-up, from how matter behaves to how bodies and brains work, how two people negotiate, how markets clear, and how empires drift. Convention, consensus, and received wisdom are recommendations by comparison. The rules that bind are the ones set by the structure of things, and much of the rest is negotiable.

So I look for returns, in capital and in ideas, in understanding systems rather than predicting events, and I map the forces that drive them. I think in decades, because that is where forces do their work. A force that moves something two percent a year is invisible in a quarter and undeniable in twenty years. Nearly everyone is paid to watch the quarter; the opportunity is in the twenty years.

These are the forces I keep returning to in markets, businesses, and the systems I study. Each one is a lens.

Force of Compounding

Most people understand compounding arithmetically; few understand it intuitively. Capital, knowledge, trust, and code all compound, so the useful question is less “how big is this today?” than “what is the rate, and what disrupts it?”

Reversion Gravity

Almost everything returns to base rates: margins, growth rates, sentiment. The few things that escape this gravity (network effects, regulatory capture, genuine intellectual property) are where durable value lives. Identifying what actually escapes, versus what merely appears to, is the central problem of investing.

Narrative Pull

Stories move capital before fundamentals do. A compelling narrative creates its own momentum: capital flows in, talent follows, outcomes improve, and the narrative appears validated. The skill is distinguishing narratives that are pulling reality toward them from narratives that are merely running ahead of it.

Friction of Coordination

Every organization, market, and system is constrained by the cost of getting people to act together. Businesses that structurally reduce coordination costs, with software, standards, or design, win in ways that are difficult to reverse. The moat there is what you make effortless more than what you own.

Incentive Weight

Incentive structures predict behavior more reliably than intentions, strategies, or stated values. When I study a company, I study what its people are paid to do. When I study a market, I study what its structure rewards. Show me the incentives, and I will show you the outcome.

Adaptation Tempo

Systems survive by adapting, but adaptation has a tempo. Too slow and the environment moves past you; too fast and you lose coherence. The businesses that last move fast enough to respond and slowly enough to accumulate.

Shape of Expectation

People are extraordinary at limbo. No matter where you set the bar, they will bend to get just under it. It comes from how humans allocate effort rather than from laziness: we anchor to whatever target exists and unconsciously optimize toward it. Set a pessimistic scope and people will fill the time. Set an optimistic one and they’ll compress their work to meet it.

So expectations do causal work. The four-minute mile was “impossible” until Roger Bannister ran it in 1954; John Landy followed within seven weeks, and more than a dozen runners had done it by the end of the decade. The constraint had been the expectation itself, acting as a ceiling on what anyone attempted. Marathon finish times cluster at round numbers (3:00, 3:30, 4:00) because runners anchor to them and adapt their effort accordingly. The goal warps the distribution of outcomes toward itself.

In organizations this goes wrong in a predictable way. A team estimates how long something will take and plans around the estimate, which becomes the target, and the target becomes the floor. High urgency (the site is down, the company might die, the deadline is now) tends to produce the fastest work, because there is no scope to anchor to, only the problem. Introducing a “realistic” timeline often makes things slower than having no timeline at all.

Optimism, then, is a strategic instrument. Set expectations that are genuinely ambitious, rather than performatively so, and people will organize their effort around meeting them. Compounded across thousands of tasks over years, the gap between what a pessimistic scope produces and what an optimistic one produces is much of the difference between an ordinary organization and an extraordinary one.

Seeing the forces comes first, and acting on them second. This is the process I follow, and the order matters: most mistakes come from doing the steps out of order.

The Process

Question every assumption. Every thesis, every model, every “obvious” conclusion should come with a name attached: the person who made the claim and the data behind it. The most dangerous assumptions come from smart people, because nobody thinks to challenge them. Challenge them anyway. Then make the assumptions less wrong.
Delete what doesn’t matter. Most data, analysis, and commentary doesn’t change the decision, so strip it away. If you’re not willing to cut at least a tenth of what you thought was essential, you haven’t cut enough. You can always add it back. In practice, you rarely need to.
Simplify before you optimize. Don’t refine a model for a business you shouldn’t own. Don’t polish an analysis that shouldn’t exist. Simplification comes after deletion, never before. The common mistake is optimizing a part of the process that should have been removed entirely.
Accelerate what remains. Once you’ve identified what actually matters, increase the tempo. Speed of learning compounds. Every cycle of observation, hypothesis, and correction makes the next cycle faster. But only accelerate after you’ve questioned, deleted, and simplified; otherwise you’re just moving faster in the wrong direction.
Systematize last. Build checklists, models, and repeatable processes only after the thinking is right. Systematizing bad thinking doesn’t make it efficient; it makes you wrong at scale.

Corollaries

Seek disconfirming evidence. The most valuable input is the kind that tells you you’re wrong. Solicit it, especially from people you respect. It sounds like simple advice, and it is rarely followed.
Be wrong, not confidently wrong. Conviction without humility is the most expensive trait in investing. It’s fine to be wrong, and you will be, often. Just don’t be wrong and certain at the same time.
Stay hands-on. You cannot evaluate what you cannot feel. Anyone making capital allocation decisions should understand the operations at a level deeper than the dashboard. Otherwise, you’re a cavalry leader who can’t ride a horse.
Urgency is an operating principle. That means a bias toward action and learning, toward closing the gap between what you know and what you need to know, rather than panic. Time is the one resource you cannot allocate more of.

Convictions

The best investments come from understanding systems, not predicting events.
Incentive structures are more predictive than financial statements.
Most of what looks like genius is actually good positioning: riding the right wave at the right time.
Intelligence is becoming abundant and judgment is not, so the premium is migrating to taste: knowing what is worth wanting.
The next decade of value creation will be determined by who controls the best training data.
Durability is rarer and more valuable than growth.

The Long Arc

The defining force of my lifetime is that thinking itself is becoming cheap. Intelligence, scarce for the whole of human history and the constraint behind most other constraints, is being manufactured at falling cost. Every system I study is being rewritten by that one fact, mostly by people who haven’t noticed yet. This is a substrate change rather than a market cycle, the kind of shift a civilization sees a handful of times, and much of what sits downstream of it is still mispriced.

Moments like this reward seeing the forces early, positioning quietly, and compounding capital, knowledge, trust, and taste without interruption. The people who matter across a century are rarely the ones who predicted it loudest; more often they built things the forces kept pushing forward long after everyone stopped paying attention.

I am playing for the arc rather than the quarter or the cycle.

Sources

Marathon finish times clustering at round numbers: Allen, Dechow, Pope & Wu, “Reference-Dependent Preferences: Evidence from Marathon Runners” (NBER Working Paper 20343, later in Management Science).
Roger Bannister broke four minutes in May 1954; John Landy followed within seven weeks, and by decade’s end the “impossible” barrier had been cleared by more than a dozen runners.

← Back