From Claude Blue to Bloom

AI now produces more code than any team can realistically review. At one of the largest companies in the world, that gap started showing up in the infrastructure itself.

The people who feel it first are not the ones falling behind. They are the ones standing furthest forward.

Claude Blue named the feeling. Bloom is what we built in response.

This piece is based on conversations between Bloom founder Wonjun Han and two people working in Silicon Valley on the same day: a senior AI engineer at Meta, and a startup founder.

The call that started it

"Our infrastructure is falling apart right now."

That was the first thing Wonjun heard on the video call. On the other end was a former colleague, a senior AI engineer at Meta. His explanation was simple. AI was generating code faster than any human could read it, and people across the company had started shipping work they had not really reviewed. Infrastructure that was supposed to be boring and stable had stopped being either.

The conversation was never meant to go anywhere. Wonjun wrote it down as a personal note. It traveled much further than expected, and the response to it turned into the first gathering of the Claude community in Seoul.

What Claude Blue actually means

Claude Blue is not a product. It is a name for a specific kind of professional grief that shows up among people working at the front of AI.

The important detail is who feels it. It is not the people who ignored the technology. The engineer in that call holds a PhD in physics and engineering and has worked at major companies and at Meta. He put it this way.

"When I write, I had my own structure, my own way of visualizing things. That was my edge. Now I ask the model to organize it and a clean draft comes back."

What changed was not the fact of being replaced. It was the speed. In earlier waves, the feeling of "I have maybe a few years left" arrived over months. With a new model release, it arrived in a day.

A day inside an AI-native team

The phrase "AI-native" gets used loosely. His routine makes it concrete. He uses Claude Code at work and Codex at home. When a task arrives, the first move is not to open the codebase. It is to decide what the agent should own and what he should own, then set the boundary explicitly before anything runs.

The bottleneck moved. Writing code is no longer the slow part. Deciding what is worth building, and being accountable for what ships, is.

From Blue to Bloom

The word Blue was accurate for the feeling, but it was not a place to stop. If the anxiety is real, the useful question is what people do with it.

In Seoul, the answer turned out to be unglamorous: get in a room together. The first meetup drew thousands of applications, which was not something anyone on the team predicted. What people wanted was not another framework. They wanted to hear how someone one step ahead was actually working, and to say out loud that the ground was moving.

Claude Blue became Bloom. Same conditions, different response.

Why this keeps mattering outside Silicon Valley

There is a comfortable reading of all this: the disruption is happening over there, and the rest of us have time. The conversations suggest the opposite. The reason most teams do not yet feel a bottleneck is not that they are using AI well. It is that they are not using it enough to hit the wall.

Korea sits roughly one to two months behind the US on adoption. In most industries that gap is noise. In this one, two months is a release cycle, and the compounding is unforgiving.

That is the whole thesis behind Bloom. Not optimism, and not doom. Just the belief that the fastest way to close a two month gap is to put the people who are already living in the future in front of the people who are not, in the same room, regularly.

Join Bloom

Bloom builds offline rooms where people and technology meet. We run them in Seoul, and now beyond it.

• 💬 Discord community, where events are announced first
• ▶️ YouTube, full talks from past events
• 📸 Instagram, photos from the floor
• 🎟️ Upcoming events