A few weeks ago, Dawn Barber, the industry liaison for Columbia's Technology Management Program, invited me to present to a class of grad students from Columbia about what it means to build a business in the AI age, and share advice for students. As part of Columbia's School of Professional Studies with a tech focus, nearly everyone in attendance had a strong intent to begin or continue working with technology after graduation.

I always love getting the opportunity to tease out some of the emergent trends happening on the bleeding edge of the industry with students who are looking for places to apply their new skills. In years past, I've covered topics like how to work in web3 or crypto companies, and how startup jobs differ quite a lot from more traditional corporate settings.

This year, since I've been pursuing my own entrepreneurial pursuits and also training diverse groups of people on AI adoption, I decided to share some of the AI Fluency best practices I've been developing, in the context of how I've been building my own startup, MuseKat.

I was expecting people to get excited to learn more about an AI-native way of starting up. I wasn't expecting to also get a free lesson on unit economics and real-time user feedback.

(But maybe that's kind of the point.)

As an early-stage founder, I spend more days than not making lots of micro-decisions on my own (or with AI). Since I don't have colleagues or a board to report out progress to, it can be hard to benchmark my progress. This is why I find showing your work – literally going through the progress of creating slides, or doing a presentation for other people – is so helpful.

The Columbia event was the second group of grad students I've met with this year. The first was at NYU's Leslie eLab, a vibrant studio and coworking space for student entrepreneurs.

As part of NYU's grad programs, they invite speakers from all stages of businesses to join for Startup School activations, demos, or workshops. Since students with non-technical backgrounds had begun asking what it takes to "vibe code" an app and turn it into a business, I thought it'd be interesting to walk them through the process.

My session in April was the first time I pulled together all of the work I'd done in the first three months of the year in my startup, MuseKat. During that time, I had successfully vibe coded a prototype on my own, reworked the character and brand design with the help of high schoolers, and launched a beta mobile app in TestFlight for iOS users.

Over 60 minutes, I walked them through the whole thing, from the beginning. My goal was to demystify parts of the founding experience, and also to show how effective use of AI in every incremental step of building can compound over time to unfathomably high productivity gains.

I shared not only how much I used AI to help me on the technical bits of building, but how I inject it as a third thought partner in nearly every step of my workflow, across every area of the business.

One of the best parts about working with AI is that you can move faster than a human on most of your tasks and priorities. But one of the worst parts about working with AI is that it will quite literally tell you what you want to hear (even if that's the AI trying to convince you that you're a religious deity, or living in a simulation).

Needless to say, human input is essential for building robust, critically vetted products. I don't have a business degree (or a graduate degree at all), so I always find it incredibly helpful to share out a view of the world that's uninitiated to see what come back my way.

This is where the grad students really stepped it up. At the Columbia event recently, students peppered me with questions in real-time as I toggled back and forth between tabs, existing artifacts, a pre-planned deck, and live demos. Of course, the questions were the best part.

At that last one, I hesitated slightly, then decided to deviate from the slides a bit.

"I'm not sure, I'm working on it..."

"Based on what you shared just now, I'd pay for it for my kid!" injected one of the student attendees.

"Ha, okay thanks," I said laughing. "Come find me after class."

"No, seriously," interjected another. "I would too."

"You would?" I asked, genuinely curious now. "Why?"

I glanced nervously to my friend Dawn who had invited me into the class to make sure this sidebar conversation wasn't too off-topic. I got the thumbs up to keep going.

We ended up deviating into a 10-minute sidebar about some of the business implications, go-to-market opportunities, and trickiness of funding anything in the AI age. I'm incredibly grateful for their help, and their smart questions. They seemed equally enthused to put their business skills to the test.

In the end, it's hard to say who took away the bigger value from the day: Me as a presenting entrepreneur with a captive audience of newly minted master's students with a much more academically aligned pedagogy, or them, as students who got a chance to engage as active participants into some of the real-time thinking alongside a founder in the trenches.

These two experiences in teaching about startup business building (even while I'm still figuring out many of these questions myself) validated the urgency to build with the public in the AI age. In fact, it mirrors a lot of what I experienced in real time in an earlier era of the Internet.

When I worked at Union Square Ventures from 2016 - 2020, I oversaw the network of 100 startup founders and helped curate small-group events, conversations, and content that directly addressed many of the real-time questions that were surfaced from builders in the trenches.

This network packed a powerful punch because it came before any playbooks or postmortems. Every event we hosted featured real builders, sharing in real time, about the messy stuff in the middle. We couldn't have canonized this content into a neat little list of best practices or how-to's if we tried. (And trust me, we tried.) In fact, any time we did, the startup playbooks so fast that our tidy content libraries went outdated almost as soon as the ink dried. We learned that the quickest path to action was inviting in people doing the work, in real time, to talk about it.

In other words:

That experience shaped how I work today. As an independent operator and now as a founder, I’ve found that teaching while I’m still learning isn’t just helpful, it’s often the only way to move fast enough to build with conviction in emerging tech. Teaching forces clarity. Sharing invites accountability. And the feedback that comes back? That’s where real learning happens.

In the AI age, speed without public scrutiny is a liability. That’s why I've realized it's not good enough to build in public; we need to build with the public. I'm so grateful to Columbia and NYU for inviting me in to share some of these real-time stories with students at the peak of their business learning journeys.

Thank you! Special thank you to Dawn Barber, Industry Liaison for the Technology Management Program at Columbia and Frank Rimalovski, the Executive Director of the Entrepreneurial Institute at NYU, for inviting me to share these stories with students this semester. And to all of the students who challenged me with such great questions!

For my next post in this series, I'll share how elementary-aged students surprised me with their assessments of AI generated content through a real-time activation at a school Career Fair.