Fast Company is delighted to make this article available to any student for free. Please request a copy by email.
I was sitting on the steps of Duke Chapel at 2 a.m. in December 2023, the Gothic towers looming above me, a 210-foot reminder of everything I was about to walk away from. My phone was exploding with notifications: Y Combinator had just accepted us. ChatGPT had hit 100 million users in two months—faster than TikTok, faster than Instagram, faster than anything in human history.
And I was about to break my single mother’s heart.
The chapel bells rang twice, echoing across the empty quad. In six hours, I’d be dropping out of one of America’s best universities. The same university my mother had sacrificed everything to get me into. The same university whose acceptance letter made her cry tears of joy in our cramped apartment kitchen in Dhaka, Bangladesh, a place most Duke students couldn’t find on a map, where American university acceptance letters arrive like answered prayers, meant to lift entire families. The only thing harder than getting into Duke from there was explaining why I was leaving.
Now I’d make her cry again.
I threw up twice before leaving Duke that morning, once in the dorm bathroom, once behind the student center. Not from the previous night’s parties; I’d stopped going to those months earlier while my hallmates were doing keg stands. I was too busy watching the world change at warp speed. While my classmates wrote papers about AI’s potential, we were teaching it to think in practice. While they debugged thesis statements, we debugged systems that would touch millions. They worried about grades. We worried about scale.
The decision wasn’t romantic. It was terrifying. But sitting on those chapel steps, watching my classmates live their normal college lives while history was being written in real time 3,000 miles away, I knew.
My cofounder, Md Abdul Halim Rafi, and I were accepted into Y Combinator’s W24 batch with Octolane AI to build the next AI Salesforce. We were one of 260 companies selected from more than 27,000 applications, less than 1% acceptance rate. Duke’s is 8%. Harvard’s is 5%. I’m not great at math, but even I could figure out which achievement was statistically more improbable.
Some trains only come once. And this one was leaving the station with or without me.
Why This Moment Is Different
History rarely offers moments when the ground shifts beneath our feet. I’ve studied them obsessively: The internet in the ’90s, when two Stanford PhD dropouts named Larry Page and Sergey Brin saw search differently. Social media in the 2000s, when a college dropout named Mark Zuckerberg connected the world from his dorm room.
Today, it’s AI, and it’s moving faster than any wave before.
Consider this: ChatGPT hit 100 million users in two months. TikTok took nine months. Instagram took 2.5 years. The telephone took 75 years. We’re watching revolution at warp speed. The AI market is projected to reach $1.81 trillion by 2030, growing at 35.9% annually, faster than the cloud computing boom, faster than mobile, faster than the internet itself.
Here’s what kept me up at Duke: By 2028, AI will autonomously make 15% of all work decisions. Not assist—decide. We’re not talking about a tool anymore. The question for me was: “Do I want to build the next GPT or learn about it in a textbook?”
Actually, scratch that. When BackRub started as a Stanford research project in 1996, Page and Brin had years to develop it before formally launching Google in 1998—and even then, it took until their 2004 IPO (six years post-launch) before the world fully understood just how profitable and dominant they’d become. Zuckerberg had seven years after Facebook’s 2004 launch before Google even tried to compete with Google+ in 2011.
AI founders? Look at what happened to Character AI: They built something revolutionary, then Google just hired the team. Look at Inflection, which got $1.5 billion in funding, then Microsoft basically bought them for parts. We don’t have years. We have months. Maybe 24 if we’re lucky.
Every week I stayed in that Duke classroom, another AI startup was getting a $100 million Series B. Every test I took, OpenAI was releasing another model that made last month’s breakthrough obsolete. Every night I spent in the library, someone my age in San Francisco was defining how humanity would interact with AI forever.
The companies that win the AI race in 2026 and 2027 will dominate for decades. The rest will be Wikipedia entries nobody reads.
The Framework That Made Me Jump
Dropping out sounds romantic. In reality, it’s a calculated risk that requires brutal honesty. Here’s the framework that helped me decide:
1. Conviction: Were we solving a problem that mattered? Sales teams waste 30% to 40% of their time manually updating CRMs. The question was simple: Could we eliminate a trillion-dollar inefficiency?
2. Timing: Was this the right moment? AI adoption was exploding, but it was still early enough for startups to move faster than incumbents stuck in quarterly earnings calls.
3. Traction: Did we have proof? Early customers kept asking the same question: “How fast can you onboard us?” One customer pulled out credit cards mid-demo.
4. Partnership: When you’re about to jump off a cliff, you better trust who’s jumping with you.
Building a startup isn’t a solo sport. When you hit the inevitable walls (and you will), you need someone who believes in the vision as deeply as you do. Someone who can carry the weight when you can’t. Someone whose skills complement yours, who challenges your assumptions, and who won’t let you quit when things get dark.
My cofounder, Rafi, my best friend since high school, left his comfortable job, said goodbye to his family, and boarded a plane to San Francisco with his wife after one phone call. He did this based on nothing but my conviction and our shared dream.
Some nights we’d code in complete silence for hours, the only sound being keyboard clicks, because we knew we didn’t have the luxury of giving up.
When someone trusts you with every fiber of their being, failure stops being an option.
5. Runway: Could we survive? YC and early investors gave us just enough capital to go full time. Not comfortable, but enough. But when you’re building with your best friend who crossed oceans for this dream, you make it work on breadcrumbs.
Without all five elements of this framework, I would have stayed in school. That’s why most people shouldn’t drop out, because unless those factors align perfectly, you’re not making a calculated leap. You’re gambling with your future.
Why Most People Shouldn’t Do This
Let me be brutally honest: Dropping out isn’t a badge of honor. It’s a tactical decision that’s wrong for 99% of people.
Stay in school if:
- You’re running from something (hard classes, social pressure, your roommate’s terrible music taste) rather than toward something specific
- You don’t have a problem that keeps you up at night and wakes you up excited (insomnia doesn’t count unless it’s productive insomnia)
- You think being a “dropout founder” sounds cool on X/LinkedIn
- You don’t have at least 12 months of runway secured (No, your parents’ basement doesn’t count as runway.)
- You’re alone without a cofounder or strong support system (Chatting with Cursor AI or GitHub copilot is not a cofounder.)
- You haven’t talked to at least 50 potential customers
- The opportunity will still be there in two years (Spoiler: If it will, it’s not urgent enough.)
The world needs doctors who finish medical school and engineers who master their craft. We can’t all be dropout founders, and thank God for that. Someone needs to actually know what they’re doing!
What I’d Tell My Past Self
If I could go back to that terrified kid holding his Duke ID at 2:12 a.m., here’s what I’d say (yes, I remember the exact time; anxiety has a way of burning timestamps into your brain):
The fear never goes away. You just get better at moving forward despite it. Every founder you admire—Zuckerberg, Gates, Jobs—was once exactly where you are: terrified, uncertain, but unable to ignore the pull of what could be.
Your parents will understand. Maybe not today, maybe not this year, but when they see you building something that matters, they’ll understand that you honored their sacrifices in a different way.
Failure isn’t falling, it’s not trying. You can always go back to school. You can always get a job. You can always become a consultant and use phrases like “leverage synergies” for the rest of your life. But you can’t always catch a wave that’s already crashed.
Find your tribe. The loneliest part isn’t leaving school; it’s the months when you’re building in obscurity while your friends are at parties you’re no longer invited to. Find other builders. Share your struggles. The founders who seem to have it all together are often one bad day away from quitting too.
A Final Thought
Last week I got coffee with a Duke student who’s considering dropping out. He reminded me of myself: brilliant, ambitious, and absolutely terrified. I told him what I’m telling you:
Don’t drop out to follow my path. Drop out only if staying would be a betrayal of the fire inside you. Drop out only if the problem you’re solving matters more than your comfort. Drop out only if you have something to build that the world desperately needs. And drop out only if you have someone like Rafi, someone who’d cross oceans for your shared dream.
But if you have that fire, that problem, that desperate need to build, then maybe, just maybe, the biggest risk isn’t leaving.
It’s staying.
