Preparing founders: Paul Graham's questions read from the region
Graham argues a university should not teach entrepreneurship but teach building well. His answer assumes one kind of institution. The region has at least three, and the criterion fits best the one nobody counts.
In August, Paul Graham published How Universities Should Prepare Founders. The thesis is short: what Y Combinator looks for in a founder is someone who is good at building things and has a habit of doing it. Which leads somewhere uncomfortable for a lot of vice-rectorates: the way to prepare founders is not to build a curriculum on “entrepreneurship”, it is to teach computer science, engineering or molecular biology well. And only two changes are needed, he says: students have to believe that starting something is a real option, and they need time to work on their own projects.
The argument holds. It also assumes a fairly specific kind of institution: a research university, strong departments, full-time students, and that gap in the calendar Harvard calls reading period. Read from Latin America and the Caribbean, the essay does not have one answer but three, depending on who is being asked.
Three institutions, three roles
Research universities. These are the closest to Graham’s assumption, and the role they have played in the region is producing most of its research and its doctoral programmes. Their limit is not belief: the role models exist and the departments are there. It is the calendar. The free time Graham asks for competes with a dense curriculum and with the logic of accreditation, which rewards what can be documented.
Mass-access universities. Their role has been access and social mobility at scale, and it is an enormous role that conversations about innovation tend to skip. Here Graham’s point about free time runs straight into a wall, because a significant share of their students study and work. Giving time back is not an administrative decision: it is a decision about who can afford it.
Professional institutes and technical training centres. Their role has been employability and the link with industry. And here is the inversion worth looking at: these are the institutions where things actually get built. Workshops, labs, final projects with a real counterpart who will use the result. That is precisely Graham’s criterion, and these are the institutions almost never counted as a pipeline of founders.
Three questions
What do we teach well, and are we using it for this? Graham is blunt: you do not need an innovation centre or a dean of entrepreneurship, and if you create them they will most likely subtract. The question that comes before any new programme is what the institution already does well, and whether that is connected to the people who want to build something.
Whose free time is it? It is the scarce input in the argument and the one that can least be taken for granted in the region. If a student works, reading period does not appear on its own: it has to be designed, and someone has to decide who it reaches.
Who does the student get to see? Graham notes that Harvard alumni apply to YC at about twice the rate of Yale and Princeton alumni, and that the difference is not in the students but in what is customary. His remedy is that they see real founders, and he adds a detail that matters more in the region than in Boston: what inspires is not the richest person but the closest in age. The useful role model is not the one on the magazine cover, it is the graduate three years ahead.
Graham’s answer costs no money. It costs institutional discipline, which is harder to come by. Here it costs the same, with one question first: which of the three institutions are you, and what is it that you already do well?
