Do Not Start a Startup in College
Waiting lets you explore in your 20s, and makes success more likely later.
YC Root Access · Lecture 3 - Before the Startup (Paul Graham)Paul Graham, the programmer and essayist who ran Y Combinator, was standing in front of a room of college students to give them the advice he said he would give his own kids. He had just compared starting a successful company to having children, a button that changes life irrevocably, and he said "in rich countries, most people delay pushing the button for a while."
Then he gave them his version for companies, in plain terms: "Do not start a startup in college."
Graham argued that a startup takes over an entire life, so you cannot be a full student and a full founder at once. If you start a startup as a student, he said, you are not really a student anymore, and you may not even nominally be one for much longer. He joked that universities hand out contraceptives to prevent one life changing decision while opening incubators and entrepreneurship programs to encourage the other, partly because applicants expect it.
He added that a class like his can only teach about startups, not what a founder actually needs to know. To explain the gap, he offered an analogy: the class was like linguistics, which teaches how languages work in general, while what you need is French. For a founder, he said, that means what Graham called "the needs of your own users," which you can only learn by starting the company itself.
What students lose by starting now, Graham said, is what they can only do in their early 20s. That means, in his words, "plunge deeply into projects on a whim that seem like they have no payoff" or "travel super cheaply with no sense of a deadline." For unambitious people, he said, that looks like failure to launch, but for ambitious people it is valuable exploration that gives them more to choose from later.
His example was Mark Zuckerberg, the Facebook founder. Graham said Zuckerberg will never get to backpack around Thailand, because any trip is either a de facto state visit or hiding out incognito in Paris. Success, he said, "has taken a lot of the serendipity out of his life," and "Facebook is running him as much as he is running Facebook." There is no tradeoff in waiting, Graham concluded, because "you'll be more likely to succeed if you wait."
Questions people ask
What should I do in college if I want to start a company later?
Work on things that genuinely interest you, learn broadly, and build side projects with people you like and respect. Graham's broader advice was to use college for open ended learning and exploration, since skill with users and technology comes from curiosity, not from entrepreneurship courses.
What if my dorm project is already taking off?
Graham allowed one exception. If you are 20 and a side project unexpectedly takes off the way Facebook did, it may be reasonable to run with it. His point was that you should not force a startup into existence at 20, not that you must kill something that is already working.
Why would waiting make me more likely to succeed?
Yes, in his view. Older founders have more technical skill, more knowledge of real user problems, and more friends who can become cofounders and hires. Waiting also lets you learn what matters by doing interesting work, rather than guessing at startup ideas.