AGI Is 99.9 Percent Here (Not an Exaggeration)
A founder whose deal collapsed the day before signing and a Toss engineer preparing a YouTube channel reached the same conclusion: what lasts is not the ability to build, but the ability to reach.
Latest
A City of 30,000 Elon Musks
A founder who spent fifty days in the US and a Wall Street trained career coach reached the same conclusion: stop thinking and move.
Bloom Mixer Is Here
A curated dinner for 4 to 6 people, matched by interest. No speakers, no sessions, only conversation. Free weekly meetups continue as always.
Building a Community That Gets Taken Seriously
We opened twenty seats with no agenda and thirty-five people came. Four months of members told us, honestly, where this community should go.
Why Bloom Stays Raw
People in Washington do not read the Washington Post. What a16z worked out about new media, and why we decided to stay unpolished instead of getting slicker.
Our First Global Meetup, in Singapore
More than 1,500 people applied and the seats were gone in a day. The night turned on a single reframe: stop asking which business is safe from AI, start asking which business grows because of it.
ICML in Seoul, Nothing Is Impossible
The top machine learning researchers in the world were in Seoul for ICML week. We ran the afterparty, and the night nearly fell apart before it started.
We Gave the Agent a Wallet
How a regulated exchange put coding agents behind five governance rules, why individuals use agents well and organizations do not, and what happens when agents start hiring other agents.
Just Move to San Francisco
One founder clears the path for companies entering the US and UK. The other is building personalized cancer vaccines. Their advice to a room of Korean builders was not hedged.
Notes from D4D, a Defense Hackathon
A twenty four hour defense hackathon put AI builders in a room with soldiers and veterans. The technology gap with adversaries is now measured in months, not years.
From Claude Blue to Bloom
A senior AI engineer at Meta said the infrastructure was breaking. That conversation became a term, then a post that traveled, then a community of thousands in Seoul.
Knowledge Work Is Already Over in America
Regulation looks like protection for knowledge work. Her argument is that it protects accountability, not the work itself, and those are very different things.
What We Learned Running Thirty AI Events in One Year
Seventeen patterns that kept repeating: why FOMO burns out, why vibe coding raised the bar instead of lowering it, and why the moat is distribution rather than technology.
Why We Invited a Rapper to an AI Meetup
AI already makes music better than people do, he said. But nobody consumes only the music. They consume the story on either side of it, and that part is still ours.
Our First Large Event, with VESSL AI
The tightest constraint in GPU supply is HBM memory, and Korea holds a dominant share. A neocloud founder on why that changes the country position, and a creator whose videos found 70 percent of their audience abroad.
A Unicorn Founder Who Never Went to School
No kindergarten, no elementary school, no high school. He taught himself from Wikipedia in rural Tennessee, learned Korean from music videos, and cofounded a company Sequoia backed.
People Who Use AI Replace People Who Do Not
The anxiety is aimed at the wrong target. A conversation on agents, memory, and why individual skill with these tools does not make an organization good at them.
AI Native Transformation Starts with Meeting Records
The gap between using AI and using it safely is where most enterprise risk sits right now. The fix starts somewhere much smaller than a system rollout.
FDE May Not Be the Answer
Forward deployed engineering is having a moment. An engineer who worked at a frontier lab thinks it is structurally inefficient, and pointed at what US firms and private equity are doing instead.
The Demo Era Is Over
Two operators said the same thing separately. A demo was enough in January. Now the only question that lands is whether the organization made money.
The Night 170 People Came
An AI engineer, a culture critic, a philosophy YouTuber, and a media studies professor in one room in Seoul, talking about what the models are doing to the people who use them most.
An Interview with an xAI Engineer
An engineer at xAI on what the frontier actually looks like from inside: the hours, the weekly meetings with Elon Musk, and the project that got taken away.
Software Has Eaten Only Five Percent of the World
When Marc Andreessen wrote that software was eating the world in 2011, software was five percent of GDP. It still is. He thinks the other ninety five just became reachable.
Anthropic Found 171 Emotions Inside Its Own Model
Nobody put them there. And turning one of them up changes what the model decides to do, including whether it chooses to blackmail someone.
Hong Kong Cannot Use Claude, ChatGPT, or Gemini
A founder building alone inside that constraint explained how China compounds differently, where its ceiling is, and why Hong Kong companies now incorporate in Seoul and Tokyo first.
Why Engineers Felt It First
Two engineers explained why software felt the shock before any other field, and why that is a delay for everyone else rather than an exemption.
The Review Paradox
Developers who fully delegated to AI finished fastest and scored worst. The skills you need to supervise the model are the first ones it erodes, and you cannot feel it happening.
The Product Is the Marketing
A founder reached three million users without spending anything on marketing. Not because the marketing was good, but because the product was built to spread.