He Chased Every Trend and Missed an Entire Era: YC President's $2 Billion Lesson
Deep Thoughts on AI and Aspirations —— ByteThink Circle
Here’s a counterintuitive truth: those who make the worst judgments at the worst times often become the wisest people later. YC President Garry Tan is a living example, and he’s willing to show his work.
When he graduated from Stanford in 2003, he had built solid hands-on experience in web programming. But the internet bubble had just burst, the NASDAQ had crashed, everyone said the web was dead, and he believed it. So he pivoted to chase the hotter Windows Mobile. Around the same time, Zuckerberg was building Facemash, and Facebook hadn’t been born yet. He later summarized it himself: wrong direction, right timing—that’s how ironic it was.
Then came an even more expensive miss. When Palantir was founded, Peter Thiel flew to Seattle specifically to take him to dinner, offering a $70,000 annual salary to bring him aboard. He was fixated on the title he might reach elsewhere and declined. By his own estimate in interviews, that refusal was worth somewhere between $2 billion and $4 billion.
But his most pointed observation came after: the root cause of both misjudgments was identical. He was reading the map, not the terrain. He cared about what looked cool, what good investors would approve of, but didn’t open his eyes to see what the smartest people around him were actually doing.
Why Chasing Trends Is Slow Suicide
Unpacking the “map versus terrain” metaphor reveals a vicious cycle.
Chasing trends delivers immediate, positive feedback. The media says it’s the hot sector, peers praise your foresight, investors show satisfaction—every signal tells you you’re on the right track. The problem is, by the time everyone can see it on the map, it’s no longer an opportunity map but a crowded one. When you catch up, a hundred other people just as “smart” as you are already crammed into that trend.
Looking at the terrain is the opposite. Trusting your direct experience, doing what you genuinely understand—the feedback is slow and lonely. No one praises you, there’s no short-term applause, you have to carry it alone. Garry Tan calls this capacity earnestness—the courage to be sincere. After you become an adult, believing in the reality you’ve touched with your own hands while everyone tells you otherwise becomes paradoxically harder.
He also drew a marker for good things: a real idea has no visible boundaries. The deeper you go, the more you discover the intellectual edge is open, infinite. Conversely, opportunities that can be crisply defined at a glance are often places where competition has already maxed out. Nearly every genuinely great thing in Silicon Valley started out looking strange, fringe, non-mainstream. “A computer on every desk in every home” was a complete weirdo dream at the time.
AI Has Amplified the Cost of Chasing Trends
This wave of AI has made an old problem more lethal than before.
According to Garry Tan, with vibe coding and agent programming, a founder today can do what four hundred people did two years ago—one person can build what used to require an entire engineering team. Code is no longer scarce; the cost of making software has compressed to near zero.
What does this mean for judgment? The leverage of wrong decisions gets amplified in parallel. Before, chasing the wrong direction cost you a pile of useless code and a few months. Now, chasing the wrong direction lets you extremely quickly build something that actually runs, actually burns money, actually consumes your entire year. The faster you build, the more thoroughly you’re wrong.
The trouble is, in the AI era, chasing trends is especially easy to rationalize. Oversupply makes trend windows shorter than ever, and the urgency of “if you don’t get in now it’ll be too late” is stronger. It’s precisely this urgency that causes more people to abandon the terrain, rush toward the brightest point on the map, and collectively miss the position that truly belongs to them. He has another judgment that corroborates this: the golden age of pure per-seat SaaS is ending. Software without data moats or network effects may no longer be viable in five to ten years. Software itself is no longer the moat—deciding what to build is.
One Line You Can Take With You
Compress this whole framework into a single decision rule, and only one question remains: is the decision I’m making right now based on the map or the terrain?
The distinction is straightforward. If your reasoning is “everyone’s doing it,” “articles say this is the trend,” “investors will like it,” that’s map-based decision-making—the gravity of trends has taken over. If your reasoning is that you’ve seen it yourself, built it yourself, have a feel for a real problem that others don’t, that’s terrain.
Don’t misunderstand “not chasing trends” as refusing all new fields. The web was once a trend too. What Garry Tan lost was something he’d built hands-on experience with. Trends themselves aren’t guilty; going in empty-handed is. The standard has only one criterion: do you have direct experience in this direction? If no, don’t go. If yes, the trend is just a place you pass through, not the destination.
Garry Tan’s own path proves something else: making mistakes isn’t scary—introspection and iteration are what matter. From chasing Windows Mobile to missing Palantir to leading YC, the through-line of this path isn’t never making mistakes, but being willing to admit them, willing to adjust, still doing the work when everyone says no. The AI era has given people like this the biggest leverage in history, provided you’re standing on the terrain first.
Key points: Chasing trends delivers fast, positive feedback; looking at terrain delivers slow, lonely feedback, so humans naturally chase maps. Good ideas have no boundaries; great things start at the margins. AI amplifies the leverage of wrong direction by 400x. Decision mantra: does the reasoning for this decision come from “everyone’s doing it,” or from reality you’ve touched with your own hands.