Most people overestimate what they can do in one year and underestimate what they can do in ten. This is not a motivational poster. It’s a description of how compounding actually works — and why it’s so hard to trust while it’s happening.
The problem with sprint culture
We live in a moment that celebrates speed. Launch fast, iterate faster, fail faster. The advice is real and useful for certain problems. But it creates a distorted picture of how durable things get built.
The companies with the longest half-lives — Berkshire Hathaway, LVMH, the great architecture firms — weren’t built in sprints. They were built through decades of consistent, unglamorous accumulation. Each year slightly better than the last. Each investment slightly better positioned than before.
The sprint model works when the landscape changes faster than you can compound. But most landscapes change slower than the hype suggests. In those cases, the compounder wins.
What compounding actually requires
Compounding is not about working harder. It’s about three things:
Starting. Every year you delay, the curve starts later. The best time to begin was ten years ago. The second best time is now — and I mean now, not after you’ve figured everything out.
Not stopping. Compounding gets destroyed by interruptions. A 30% annual gain for five years gives you 3.7x. A 30% gain for four years, a bad year, then another good year gives you far less. The math is unforgiving. The most dangerous thing isn’t a bad year — it’s quitting during the bad year.
Reinvesting. The gains have to go back in. This is the hardest part psychologically. When things are going well, the temptation is to cash out — to take the win, declare victory, move on. But that’s extracting from the compound, not feeding it.
The visibility problem
Here’s what no one tells you: compounding is invisible while it’s happening.
For the first several years, you look exactly like someone who isn’t compounding. The gains are small. The progress feels slow. The people sprinting past you look more successful by almost every visible metric.
This is the part where most people quit. Not because the strategy is wrong, but because there’s no feedback confirming it’s right. You’re operating on faith in a model, not on evidence of results.
The evidence comes later. Suddenly and all at once. That’s the nature of exponential curves — they look flat until they don’t.
My own delayed start
I have a structural lag that I used to think of as a disadvantage: roughly ten years behind where I would have been without the constraints of my starting position — five years from family circumstances, five from country and context.
For a long time, I measured myself against people who started a decade earlier with a decade fewer constraints. That comparison was useless and demoralizing.
The correct frame is this: I am in the early part of a long curve. The curve is real. The compounding is happening. The people I’m measuring myself against are at a different point on a shorter or steeper curve — but we’re not running the same race.
大后期 — the late game — is the strategy. Not the consolation prize.
What I’m compounding toward
The clearest asset I’m building is cross-domain pattern recognition. Every domain I enter adds a new lens. Architecture. Philosophy. Technology. Chinese intellectual tradition. Each one alone is unremarkable. The combination becomes increasingly rare as the years go on.
AI is accelerating this. The value of depth in any single domain is declining fast. The value of connection — of seeing that a governance problem and an attention mechanism problem share the same underlying structure — is increasing. I’ve spent years, unintentionally, building exactly the asset that’s about to matter most.
That’s compounding. Not always by design. But compounding nonetheless.