Paul Graham Just Published 'How Universities Should Prepare Founders.' What It Means for Your YC Application If You're Still in School

Paul Graham just published How Universities Should Prepare Founders, and it shot to the front page of Hacker News within hours. On the surface it's advice for university administrators. But read it as a YC applicant and it's something else: the clearest public statement in years of exactly what YC partners look for, from the person who built the filter.

If you're a student, a recent grad, or anyone wondering whether your background "counts" for YC, this essay is a cheat sheet. Here's what it says, and how to turn it into application answers.

What does YC actually look for in a founder?

Graham states it in one sentence: "They want people who are good at building things and have a habit of doing it."

That's it. Not a polished business plan. Not a finance background. Not a 4.0. Graham is explicit that YC partners "care a lot about the projects that applicants have worked on and not at all about their GPAs."

So when the YC application asks what you've built, it isn't small talk. That question is the whole evaluation compressed into a text box. A candidate who can point to three things they shipped because they wanted to, before anyone told them to, is answering the only question that matters. A candidate who lists coursework and a business plan competition win is answering a question YC didn't ask.

Do you need to study "entrepreneurship" to get into YC?

No, and Graham argues it may actively hurt you. He keeps "entrepreneurship" in quotes throughout the essay on purpose. His point: the hard part of a startup is product, which means "knowing what to build, and being able to build it." That knowledge comes from computer science, mechanical engineering, or molecular biology, not from management or finance.

He's blunt about the theater universities put on. Business plan competitions, he writes, are "not merely useless but positively misleading," because they train founders to believe fundraising is the essential step and that the goal is a story that appeals to investors. As an investor himself, Graham says that's backwards: "The people you need to impress are users, not investors, and the way you impress them is with prototypes, not words."

The application lesson is direct. If your YC application reads like a pitch deck aimed at investors, you've optimized for the wrong judge. Show traction, prototypes, and users. A screen recording of something working beats a paragraph about your total addressable market.

Does your major matter?

Less than you think. Graham's own example is Mark Zuckerberg, "a psychology major, not a CS major," who was nonetheless good at programming. His broader claim is that building should be "understood in a very broad sense" - he points to Steve Jobs studying calligraphy, which later shaped Apple's dominance in desktop publishing.

What matters is that you can build and that you've built. If you're a bio major who ships software on the side, or a designer who codes, that cross-domain building is an asset, not a liability. Don't apologize for a non-traditional path. Name the powerful ideas you chased and the things you made because of them.

Do you have to drop out of college to do YC?

Graham is careful here, and so should you be. He writes that "running a startup is incompatible with being a full-time student" and that "the only way to learn how to start a startup is to do it." He's not telling every 19-year-old to drop out on day one. He's saying a real startup will eventually demand your full attention, and you can't run one seriously while carrying a full course load.

The practical read for applicants: YC doesn't reward the dropout gesture itself. It rewards the commitment and the traction the gesture is supposed to signal. If you're still in school and applying, the strongest position is a project that's already pulling you out of school by force of its own momentum - users, revenue, or a product that's growing faster than you can keep up with between classes. That's the story that reads as "founder," whether or not you've formally withdrawn.

Why projects are the thing YC keeps asking about

Graham gives four reasons side projects matter, and every one maps to something the YC application probes. Projects are the best way to understand a subject deeply, because "the excitement of creating something new is a much more powerful motivator than the fear of doing badly on an exam." Projects are how cofounders find each other - he notes Apple and Microsoft "were just the last of many projects their founders had worked on together," which is why YC weighs how long cofounders have actually worked together. Projects make starting a company feel natural, because you're already used to working without a boss. And projects are where the best ideas hide: "random side projects are where the best startup ideas come from," precisely because the best ideas look too implausible for anyone hunting deliberately.

That's the answer for founders who freeze on the "why this idea" question. The strongest origin story usually isn't "we surveyed the market and found a gap." It's "we built this for ourselves and people kept asking to use it." As Graham notes: who'd expect a giant company to grow out of a student directory?

The two Harvard stories worth stealing

Graham drops a detail most applicants miss: Microsoft and Meta both got started during Harvard's reading period, the gap between classes ending and exams starting. His argument is that unstructured time with nothing due the next day is what let those projects exist at all. "Merely eliminating it for a few weeks resulted in two trillion dollar companies."

He also notes both Gates and Zuckerberg got in trouble with Harvard over those projects - Gates for bringing non-student Paul Allen into the computer lab, Zuckerberg for Facemash. The takeaway isn't to break rules for its own sake. It's that genuinely novel projects are "untidy things" that don't fit neatly into official programs. If your best work happened in the margins of your schedule, outside any class or club, that's not a weakness on your application. That's the pattern.

How to use this before you apply

Graham handed applicants a rubric. Before you submit, audit your application against it: Does it lead with things you've built, or things you've planned? Does it show users and working prototypes, or a story aimed at investors? Does it treat your projects - especially the weird, unofficial ones - as the main event?

The hard part is seeing your own application the way a YC partner will, because you're too close to it. That's the reason YC Roaster exists: it puts your application in front of founders who've been through YC, so the gap between "I built things" and "I clearly demonstrated I build things" gets caught before a partner catches it for you. Graham just told you what the filter looks for; the next step is clearing it.

YC F26 applications have closed, but the next batch is already the thing to prepare for. And if this essay is right, the best preparation isn't polishing a pitch. It's going and building something.