Looking back, written on the day we flew to New York. Part 1 of 5 of the Launchpad series.
Launchpad, part 1: Seven days before
A call from René, twenty minutes to decide, Tenerife cancelled and one week of building before the flight to New York.
The flights to Tenerife were booked in my name. So was the rental car. I cancelled both, one day before I would have left.
It started about a week before Launchpad, with a call from René. Launchpad is a startup competition run by Stan: two weeks at a camp in the Pennsylvania woods, around 30 teams picked from about 700 applications, challenges, eliminations, and only five teams make it to demo day. The winning team gets USD 100,000. René wanted to know if I was coming.
I needed ten, maybe twenty minutes. I didn’t think much in that time. It was gut over head, and my gut said yes right away. I applied for my ESTA that same night.
What it cost: a week away with friends that was already paid for, a few hundred euros for sure, and two university courses I dropped for this semester. All that while I had a few good offers from other companies on the table.
I’m not new at Kobe, by the way. I’ve been building it with René on the side since March, and in spring the two of us drove a van up to the Arctic Circle and coded on the way. Since September 28 I’m full-time. The call turned “on the side” into a full yes.
Quick version of what Kobe does: you say who you want to meet, and Kobe finds the best path to a warm intro through your own network. We’re a team of three. Nils does sales, pitching and social media, René is the CTO, and I’m the founding engineer. For the next two weeks that mostly means one thing: build.
The hardest moment wasn’t the call. It came right after. One night around September 27 we cut the scope radically and set the product up again. On one side was what we wanted to build and had already promised. On the other, what actually worked. That pulled me down a lot, and I was close to quitting. We stuck with it. We didn’t throw the big plan away, we just pushed it back.
After that things moved fast. In the night to October 1, I moved the UI of our Mac client from SwiftUI to Electron, so we’d stay on the same TypeScript stack as everything else. According to Git that took from 1:10 to 1:44 am. Thursday and Friday were the new frontend: merging our old backend with Nils’s much nicer prototype frontend, and keeping only what really works.
What I cared about most were loading states that show something is happening. So “waiting for your mail”, “connected”, “starting import” instead of an empty screen that makes you wonder whether the app crashed. Plus errors shown right where they happen, not somewhere in a log. And a network graph that shows more when you zoom in.
On Friday night René and I coded until 2 am. This morning I packed at sunrise after exactly one hour of sleep. Now nine hours of flying to New York, my first time there.
Tenerife would have been a beach. Pennsylvania will be forest, cold nights and days that start with a session at 9 am and end somewhere in the night. Whether the trade was worth it, I’ll know in two weeks.