The Second Lap Needs a Team Radio: Starting Season 2 of Nitro Seoul 2026
Eleven months.
That is how long it has been since Opus 4.5 came out in late November last year. I remember that week as the moment vibe coding truly began. You described an app in words and it existed. What you fixed overnight was running by morning. Back then, the trait that mattered most was the ability to execute.
Season 1 of Nitro started on that feeling. We believed that if AI took the weight of implementation off our shoulders, one or two people could do the work that once took ten. We ran with four teams all spring, and that belief turned into numbers sooner than we expected. One team raised from Hashed Venture Fund III and now brings in hundreds of millions of won every month. Another grew revenue at a startling pace, was acquired, and has already exited. All of this happened within a few short months.
And before a year had passed, the air changed. Most services can be cloned within an hour. Screens, features, prompts, everything visible is becoming open source. Haven’t we all felt in our bones how easy building has become? If so, now that anyone can build, does simply building still matter? Where has the bottleneck moved?
Signals pointing the same way arrived from around the world. On a stage this spring, Karpathy confessed he had never felt so far behind as a programmer. He drew a line between vibe coding and agentic engineering. If vibe coding is about making prototypes fast, agentic engineering is about directing many fallible agents while keeping your precision and taste intact.
Beijing released an action plan to nurture one-person companies. At a demo day in Tokyo, an AI employee that single-handedly runs marketing for small businesses took the top prize. One American startup reportedly handed its fundraising to its own agent, which fielded more than 130 investors. Sequoia predicted that the next trillion-dollar company will come from services that run like software.
Lay these signals on top of one another and a paradox appears. Once anyone could run a company with a few agents, building well on your own stopped being enough to win. The ability to build has become a public good, and the contest has moved back to the organization.
Formula 1 went through this first. Even when the grid is full of teams running the same engine, the teams on the podium differ. The gap comes from pit stops of barely two seconds and from judgment on the pit wall, and what connects that judgment is the team radio. When a driver says in a few words that the front tires are sliding, an engineer lays that single remark over the data and changes the strategy two laps later.
Models are now an engine everyone buys. In the agent-native teams I have watched closely, the radio was alive in two directions. Inward, context flows freely between people and agents. Outward, a single line of complaint from a customer changes the product, the price, and the message that same night. The easier building becomes, the more look-alike options customers face, and the blurrier the reason to choose us. Teams with sharp go-to-market find that reason in the customer’s radio one lap ahead.
These days I hold on to the same question every day. What cannot be forked? Even ways of working move across in a handful of public prompts. Any work that comes with an answer key will eventually be copied.
So I turn the question around. Some things do not come along even if you take the whole repository. A sense of why a customer left last month and why they came back this month. A team’s agreement, after countless arguments, about what not to do. The people who pick up the phone at two in the morning. And the reason a founder bet their life on this particular problem. The organization is harder to copy than the product, and the founder’s heart is harder to copy than the organization.
Get Simon Seojoon Kim’s stories in your inbox
Join Medium for free to get updates from this writer.
When the Nitro team sat down together ahead of Season 2, Steven, who had just joined us, asked a question. Is one person using agents well the same thing as a team using agents well? Season 1 teams produced remarkable numbers through individual speed, but we rarely saw that speed accumulate into the team’s capability. Discoveries usually went cold inside one person’s laptop.
This question is not only for founders. It sits on the desk of every developer and PM who switches on an agent today. Ten people who each became ten times faster do not add up to a hundred times. More often, code and documents that nobody has read to the end pile up a hundredfold. When writing becomes free, a team’s speed is set by how fast it can read and decide.
Agents pour out code all night without tiring. But which of it to put your name on, and who answers for it when it is wrong, is still up to people. Developers become the ones who set the bar agents must clear, and PMs become the ones who decide what to give up. More valuable than writing the bar once is the judgment to rewrite it in the face of customer reaction.
So these are the teams we hope to meet in Season 2. Teams where agents have a place on the org chart, and where the line between people and agents is redrawn every week. Teams that turn one person’s discovery into the way the whole team works. Teams that are clear about who executes and who decides. Teams that relentlessly design the path from what they built quickly to customers and revenue. And teams that, in an age when everything visible gets copied, build up what is invisible every day. Even if you are alone, if your agents divide roles and context and turn together like gears, that is already a team.
For teams like these, Nitro decided not to follow the grammar of accelerators. The familiar formula is to teach a curriculum over a fixed period and graduate everyone at a finish line called demo day. But in an age when models change every quarter, a curriculum from three months ago is already an outdated map. At Nitro, there is a radio instead of lectures, and the next lap instead of graduation.
We call it Season 2, but it is not a batch selected all at once by cohort. We recruit on a rolling basis, and a team that is ready enters the track on the day it is ready. We ask for your understanding in advance, though. Selecting many teams is not the goal. For the radio to reach each team deeply, only a few teams can be on the track.
Because Nitro is a program for early-stage teams, investment runs through a process separate from Hashed’s main fund, and it moves very fast. But in connecting you to all of Hashed’s global network and resources, it is the same as the main fund. What we want to build together is a relationship where people know each other’s problems and respond. Records can be exported, but the trust built by making decisions together cannot be forked.
We do not want to be judges holding scorecards in the grandstand. We want to be engineers on the pit wall, watching the same screen and holding the radio.
A thousand builders came through Hashed events this year, but once the events ended, they scattered with no reason to meet again. That is why we are laying down a radio network now. It is a system in which one team’s radio reaches other teams, the developer community, and the Hashed team, and from there spreads to co-investors and to partners around the world, including our LPs. A single remark from a garage in Seoul can come back from a meeting room on the other side of the planet as an introduction to a first customer, or as the next round.
I imagine a Friday night. A founder in Seoul, wrestling alone with an agent, posts a single line about where they are stuck. A few minutes later, a developer from another team replies: we got stuck there too. The approach their agent tried follows, and the next day three teams try fixing it together. A place where a problem you were carrying alone becomes a problem to solve together. That is the Nitro I have in mind.
Accelerators draw a finish line. Nitro lays down a radio network. A year ago, the answer was execution. We do not fully know today’s answer either. But this much is clear: the answer lies somewhere on the side of what cannot be forked, and the radio that finds it cannot be switched on alone.
Today, a Track Day meetup introducing Nitro Season 2 takes place at the Hashed Lounge. You do not need a polished pitch. One problem you are stuck on right now is enough. Come if you hold the wheel, and come if you want to race alongside the one who does. Applications are rolling and open right now. https://nitro.hashed.com
The second lap has already begun. Now it is your turn to switch on the radio. We are listening on this frequency. All the way, on your side.
