Paul GrahamVerified account

Projects that yield great results often wouldn't seem promising to a rational person. Fortunately some people have a compensating irrationality: they're so interested in the project that they'll work on it for its own sake.

One reason big companies can't do what startups can is that people in big companies aren't allowed to have this kind of irrationality. You can't work on something just because it seems cool. You have to explain what the benefits will be. A reasonable request. All too reasonable.

Y Combinator itself was an instance of this pattern. We had no idea if it would make money. But it seemed like a cool thing to try. And being an instance of this pattern helped us to see it in startups that applied.

Airbnb for example. What rational person would work on such an unpromising idea? But we could see that they were like us in being really into their apparently unpromising idea, so we suspended our disbelief.

wasn't their original idea really bad and they eventually changed it? I always thought that Ycombinator bet on the team figuring out a good business model than on the idea itself.

They expanded it (originally it was just Airbeds for conferences) but the old idea was always a proper subset of the new one. Which continues to this day incidentally.