Hello LJV Cohort #7. My name is Terry. My four years at Northwestern happened to overlap almost exactly with the AI wave — from ChatGPT launching my freshman year to now, when every company is trying to figure out what to do with it. I’ve been a founder, an operator at TikTok and Microsoft, and an international student through all of it.

Things changed a lot while I was here. They’re still changing. I want to talk about three things I picked up about how to use that to your advantage.

Where you put yourself matters

I grew up in Ajax, a quiet town just north of Toronto. My primary school had 20 kids per grade. Those same 20 kids went to middle school two blocks north, then high school one block further. Most of my friends ended up at U of T or U of Ottawa. My life was basically mapped out at age 11.

When I was going into sixth grade, my sister went to Cornell, and I had a choice: private school in Canada, or go back to China — a place I’d only ever visited for summers. I knew how to speak Chinese — my parents spoke it at home — but I’d never gone to Chinese school. I decided to pack up, say goodbye to everyone, and go.

I got into one of the best middle schools in Beijing. It was so prestigious that kids would wear the school uniform on weekends just to brag. When I was interning at Microsoft last year in Seattle, I saw someone still wearing the same pants from that school. It’s become this cult where you spot alumni wearing that uniform across Europe and the US.

I had the most miserable three years of my life there. I got a 28 on my Chinese exam. I barely got 60 on math. There was this thing where you’d recite poems in class, and kids would hint to the teacher to call on me because they knew I’d never know the poem. I made a fool of myself regularly. But I stuck through it.

By high school, I’d become good enough that I had guaranteed admissions to the top two universities in China. And then I gave that up — for Northwestern, for The Garage. I actually wrote about The Garage in my admissions essay. I wanted an environment where entrepreneurship was the norm, not a fading trend.

Here’s the thing I didn’t realize until later: every hard thing I put myself through turned into something useful.

Those three miserable years in Beijing? Now I can read Chinese AI social media and know exactly what companies like DeepSeek and Moonshot are working on before it hits English-language media. I can communicate with hardware teams in Shenzhen. I have a cultural bridge that most American founders don’t have.

Not being the most technical coder? That constraint forced me into AI-assisted coding earlier than most people. I learned to use these tools not as a novelty but as a necessity. Now I ship faster at higher quality than many engineers who are technically stronger than me.

Being an international student with work authorization limitations? That forced me to understand the immigration system inside and out. It’s been a non-issue for me while friends with multiple offers have gotten them rescinded because companies couldn’t sponsor.

The places you put yourself — physically, socially — shape everything. Not just the city, but what the people around you consider normal. At The Garage, I honestly learned more from conversations with other people building startups than from actually building startups. The environment is the curriculum. And the limitations that come with your choices aren’t obstacles. They’re the things that make your path uniquely yours.

Being early matters more than being good

Thanksgiving of freshman year, I was at Mudd. ChatGPT launched. I was so shaken by it that I stayed way too long — by the time I went back, the building had closed. I basically ran home without my laptop. But that night I tried it on my CS 111 Racket homework. It generated C++ instead of Racket. It was terrible. But something fundamental had shifted.

I started building with it, and that became Cogno — a multi-agent sales assistant. Back then, nobody called anything “agents.” It was so early that we got real investor interest. We got term sheets. One investor offered a couple million dollars if I dropped out. I didn’t. I’m very glad I didn’t. But the fact that the conversation was even happening shows how much access you get when you’re early. We were able to recruit people who eventually went on to work at Google, Amazon, Sierra, Intuit. Being early meant big companies wanted to talk to us just because we were building in the space.

Then TikTok. I was talking to someone there who was interested in the design approaches we’d taken at Cogno — particularly intent-based routing. They wanted me on the team. I found a sublet within two weeks, cut my trip short from Puerto Rico, and flew directly to Silicon Valley.

I worked on TikTok’s first copilot product for six months. A core feature was avatar generation — letting you create a clone of yourself to record videos without being on camera. It was so new that CNN reported on it, except they used a case of Hitler reciting terrible things to make the point that it could be dangerous. That led to the product being deprioritized permanently.

Then my manager left. For about a month, I had no manager. My direct manager was in Beijing and I think they forgot they had me in San Jose. Three meals a day, best office space, smart people to talk to, and nobody telling me what to do. That was the most productive month I had there. I worked on content understanding — developing a way of understanding TikTok videos that planted the seeds for what I’m building now.

In 2023, a proof of concept was enough. You could attract talent, get investor meetings, and have big companies take your call just because you were building in AI. The bar was showing that something was possible.

That’s not where we are anymore. AI has matured. It’s already making tangible impact in real businesses. The bar now isn’t “can you demo something cool” — it’s “can you make this work in production, reliably, at scale.” Everyone and their mother is building agents. Just being in the space isn’t enough.

But that shift creates its own opportunity. Think about something that’s currently a service business — consulting, research, data entry, whatever. The question now is: how do you do that 10x better? How do you build the AI-native version of it? How do you scale what used to require a room full of people? The proof of concept is the easy part. Building the production environment — making it reliable, making it actually work for real users — that’s the hard part. But if you get that right, you can move a long way operationally.

The point is: pay attention to timing. What the opportunity looks like changes. In 2023 it was about showing something was possible. Now it’s about making it real. Both are windows. You just have to recognize which one you’re in.

It’s OK to not know what’s next

After TikTok, I went to Microsoft in Seattle. Good hikes, good pay. But features that could have shipped in two months took nine because of how many people were involved. I’m honestly not sure what I built there made much of a difference. I was paid well to do it.

Now I’m building CrowdListin as a startup, spending 40 hours a week on the product. We’re testing with three pilots. They’re happy with it. They’re surfacing real insights. But I also have offers on the table. I genuinely don’t know what I’ll be doing in three months.

And here’s what I’ve noticed: every previous stage of my life had a clear next step. In middle school, the goal was surviving the year. In high school, it was getting into college. In college, it was landing an internship. Getting a return offer by June. The conveyor belt was always there.

Post-grad, the conveyor belt stops. And I think that’s actually a good thing.

If I’d had a rigid plan, I wouldn’t have interned in Silicon Valley as a sophomore. I wouldn’t have had drop-out conversations. I wouldn’t have flown from Puerto Rico to San Jose on two weeks’ notice. None of the interesting things in my life were planned. They happened because I was open to them when they showed up.

I still don’t have a five-year plan. I might work on CrowdListin full-time. I might join an early-stage startup. I can postpone some of these decisions, which is something I only recently figured out. But as you collect more data points, the decision becomes clearer. Not knowing isn’t a problem. It’s just the state you’re in when you’re early in something fast-moving.

Close

I moved countries twice and gave up guaranteed paths twice. I still don’t know what I’m doing after graduation. The common thread across all of it is trusting my own read of the situation when the data was incomplete. That’s what I’d leave you with. You don’t need to have it figured out. You just need to be paying attention and willing to move.

Thank you.