Chinat · August 25, 2026

Manage the Forest, Not the Trees

A hackathon is not a business. It is a season. I spent last weekend in New York with over a hundred people who run these events for a living, and the sharpest framing I heard there was this: a single hackathon is a tree. It needs sunlight and rainwater, and both are seasonal. Some years the sponsors show up, some years they don't. You can grow a beautiful tree and still watch it die in a bad season.

The organization that has survived longest in this space doesn't really run hackathons at all. It manages the forest. Its problems are ecosystem problems: setting norms, keeping organizers from burning out, making sure next year's events get planted. The individual trees come and go. The forest persists.

I have been running trees for three years. Some of them were great trees. Our biggest single contract was real money, and the events themselves worked: hundreds of builders, real projects, sponsors who came back saying it was the best-run event they had touched. But every one of them was a harvest, not a farm. The moment it ended, revenue went to zero and the team started from scratch on the next one. Delivery and sales at the same time, forever. On the call I said it plainly: the hackathon work is too seasonal to be the business.

So the question we have been sitting with is what our forest actually is. And I think the answer has been in front of us for months.

The strongest thing we have built this year is not an event. It's a pipeline. Founders Bridge started as a program at Stanford: take student builders, run them through a structured incubation sprint, and end at a demo day where the work meets real investors. One of the teams from that first cohort is now in a final-round interview with a top seed fund. Last night we launched the Hong Kong chapter, with a Stanford professor who flew out for it, partners in the room, and students running the show. This week I pitched the same playbook to a student incubator at Johns Hopkins, where the pain is exactly the pain the forest layer solves: leadership graduates, continuity dies, and every new team reinvents the wheel.

That's the reframe. Student incubators and accelerators are trees too. Each campus program is seasonal, dependent on whoever happens to be running it this year, rich in energy and poor in infrastructure. Nobody is managing that forest. There is no league for student demo days, no shared playbook for cohort programs, no software layer that makes the pipeline from hackathon to cohort to demo day repeatable across campuses.

We already have the pieces. We have run the hackathons that feed the top of the funnel. We have built the demo day software and watched what engagement data from those events actually shows. We have run the cohort programs at two campuses on two continents. What we have not done, until now, is admit that the product is the connective layer, not any single event in it.

This means letting something die, and I want to be honest about that. MentorMates as a hackathon business is over. We said it out loud this weekend and it stung less than I expected, because the thing replacing it is not a retreat, it's a promotion. Running one great event proves you can grow a tree. Making the pipeline work at Stanford, then Hong Kong, then Hopkins, with the same playbook and the same software, proves you can manage a forest. Only one of those compounds.

The near-term test is unglamorous. A hackathon in September feeds a filtered hack day in October, which feeds a five-week cohort, which ends at a demo day that investors already told us they will attend. If that chain runs on a second campus without me personally holding every link, the thesis is real. If it doesn't, I'll have learned that I was just describing my own labor in ecosystem language, which is a failure mode worth naming too.

There's a personal version of this too. For three years my identity has been the person who plants spectacular trees, each one a sprint, each one ending in a clean zero. The people I respect most in this space stopped optimizing for the next tree a long time ago. They pick a forest and tend it for a decade. For the first time, my plan is not about the next event on the calendar. It's about whether the pipeline runs without me.

Check-in date: by the Hopkins demo day in December, the second-campus chain either ran or it didn't. I'll write about it either way.

← Back to all posts