Are They Still Teaching Air Guitar? Thinking About University Entrepreneurship in the Age of AI

Sixteen years ago, I wrote that, “Business plan competitions are the air guitar championships of the startup world.”

I meant that you can mimic all the movements of a startup without any of the real risk-taking or building. My criticism then was that we were teaching students to pitch without teaching them to build.

Fifteen years later, that gap matters even more. In a world where AI can help anyone ship a working product in a weekend, and where entry-level jobs are disappearing, the ability to actually build something—even something small—isn't just a nice skill to have.

It might be the whole ballgame—and I’m curious how many universities are thinking this way about their entrepreneurship programs.

Entrepreneurship Is No Longer a Track, It's a Survival Skill

I've taught or spoken in entrepreneurship related classes at Yale, Harvard, NYU, Fordham and countless other schools. I've been a judge at tons of student pitch competitions.

When I first started teaching, the technology students were most curious about was Web 2.0—blogging, digital media, what it all meant for marketing and communications. I was handed what was essentially a digital media class and turned it into something about the entrepreneurial mindset, but the audience for that framing was narrow. Learning these tools was mostly about getting ahead inside a company—the kind of thing that might help you get promoted at L'Oreal or P&G for leveraging blogs and social media for earned brand awareness.

After 2008, when banks stopped hiring, that changed. Tech went from a career accelerant to a viable career alternative to finance and MBAs sought out venture dollars instead of Goldman signing bonuses.

Now we're in a third moment.

With layoffs, AI absorbing entry-level work, and stable career ladders disappearing, being a builder—or an orchestrator of AI agents who build on your behalf—might actually be the only path college graduates have to a stable economic future. The idea that entrepreneurship courses are for the small segment of students who think they might start a company someday is just wrong.

Every student coming out of a university right now—especially in this job market—should be learning to think and act entrepreneurially, and that means trying to figure out how to generate income now.

Part of that means thinking about income sources differently. AI tools make cottage industries newly viable. You find a niche problem that isn't large enough to attract venture funding but matters enough to a small audience willing to pay. A student who builds something that earns even partial income while they're in school has changed their relationship to financial stability and professional risk in a meaningful way.

Fall in Love With Problems

One issue I see with that is students aren't fluent enough in the problems they're trying to solve. This isn't their fault—it's mostly a function of not yet having enough life and professional experience. You can't deeply understand the pain of managing a sales team or navigating a broken supply chain if you've never done it.

Second, entrepreneurship programs tend to attract people who are idea-oriented rather than problem-oriented, and they’re rarely builders. It’s still a bit of a novelty to teach entrepreneurship to computer science grads like I do at NYU Courant, and there aren’t many vibe requirements when you get a business degree.

Most of the time, when students self-select for being interested in entrepreneurship classes, you get a room full of aspiring founders who couldn't ship anything.

AI has now made that inexcusable—but has it changed how entrepreneurship is taught?

Vibe coding, no-code tools, agents—a student today can go from idea to not just a working demo in a weekend, but the actual product. The technical gap that kept entrepreneurially-minded students on the sidelines is largely gone. But the first problem—the fluency problem—remains. In some ways it's getting worse, because now students can build faster than they can validate. They can have a polished demo for a problem nobody actually has.

The fix isn't more hackathons. It's more conversations with people outside the university.

The Skill Nobody Grades

This is the piece that's consistently missing from entrepreneurship curricula: networking and relationship building. Not the cringe, transactional version—genuine curiosity about what real people face at work.

Students can't get input on what paying customers actually struggle with unless they're actively talking to them. And yet I still meet students who don't know the basics of staying in touch with people. They freeze up around "what would I even have to offer?" They haven't been shown how to expand their curiosity outward.

After I graduated from Fordham, I started an alumni mentoring program—52 pairs of recently graduated students connecting to juniors and seniors about to start their careers. The program has grown tremendously since then and continues to this day, but shouldn’t something like this be mandatory at all schools, for every single student?

Building professional relationships to inform your entrepreneurial interests shouldn’t be optional. It should be an assignment. Each semester, you should send an update to your network—to all the speakers that came into your classrooms and the alumni you met at sporting events. You should get credit for doing the thing that actually builds a career.

How else are you going to get to know an industry well enough to build a startup around it?

What AI Actually Enables (And What It Doesn't)

Here's how I'd tell students to use AI right now:

Use it for individualized coaching. Career offices can't scale for everyone, but AI can. An agent can check in, ask who you've been reaching out to, help you craft a follow-up email, and prep you for an informational interview. It can help figure out which alumni you should be reaching out to and how to approach them—new areas to explore, newsletters and podcasts to track, etc. The coaching infrastructure that used to require a full staff can now be partially automated—which means it can actually reach every student.

Use it as a writing companion, not a ghostwriter. Do the informational interview with alumni yourself. Talk through your ideas in your own words. Then use AI to shape the outline, tighten the draft, sharpen the argument. Your voice stays in it.

Ask it what entrepreneurial ideas might come out of your conversations—not necessarily venture backed ones, but what kinds of tools and services people would pay for now, so that you’re not cash flow negative the day you graduate.

Use it to test faster. No-code tools and AI-assisted development mean you can put something real in front of potential users quickly. That's powerful—but only if you're willing to hear no. Students' ideas don't get challenged enough. The bar for "is this actually worth building" needs to stay high even when building gets easy.

The opportunity here isn't that AI makes entrepreneurship easier. It's that AI removes excuses. The tools are available. The coaching can scale. The only thing left is the willingness to talk to real people, stay curious about their problems, and keep a high bar for what's truly worth working on.

That's what I'd want every entrepreneurship program to teach.

I wrote a book called Founder Unfriendly: What Investors Won't Tell You About Getting Funded, coming out April 28th. It's aimed squarely at the founders who don't have easy access to capital or well-connected networks—which is most of the students in entrepreneurship programs. If you run an entrepreneurship program and want to discuss bringing me in to speak, or getting the book in front of your students, I'd love to hear from you.