Silicon Valley Is Living Don't Look Up

Today I had a long video call with an older friend from college, the man at the center of my "Claude Blue" essay back in March. He reached out after reading my recent post about leaving my job. He said he'd been watching Bloom from a distance and was impressed, and he suggested we finally have a real conversation.

Living in Silicon Valley doesn't automatically make anyone insightful. But he has spent years wrestling with technology through the humanities, philosophy and faith, and after a long career, the insight he's built up from everyday life is unusual. It's now October 2026, seven months after that first essay. Over more than an hour and a half, he gave me a vivid picture of what AI is doing to Silicon Valley.

The short version: there is no more Claude Blue in Silicon Valley. Everyone has already accepted, as settled fact, a future in which AI replaces them. Scientists, doctors, welders: every profession will be replaced eventually, and nothing we do can reverse it. It's exactly the mood in "Don't Look Up", when people decide the comet is going to hit anyway, so they might as well enjoy the time they have left.

Of course, nobody can live in pure despair, so people make practical calculations. Some hang on with the attitude of "make as much money as I can before I'm laid off." The more proactive ones say, "I'm getting cut eventually anyway, so I might as well burn tokens now and build a few apps and side projects of my own."

Silicon Valley is far more competitive than Korea, and it's overflowing with meetups. Yet when he heard about Bloom, he called it a strange kind of gathering, something like a church sharing circle. The frontier engineers who use AI most intensely carry the deepest wounds and anxiety, he said, but Silicon Valley right now is too barren and too short on breathing room for people to bring those wounds into the open and care for each other.

Korea has no breathing room either, but the FOMO on the front lines in Silicon Valley is on another level. He said plenty of people around him, himself included, are running at a crazy pace. We even talked about whether something like Bloom might resonate more in San Francisco than in Seoul.

He believes Korea will survive the AI era, stubbornly. To skip ahead to his conclusion: by all rights Korea should have failed, yet it has made it this far, miraculously, by the grace of God. We're both Christians, so that resonates with me, but even setting faith aside, anyone would admit that modern Korea's story is one of a kind.

So how will Korea survive from here? By becoming America's outsourcing base. That role has traditionally gone to Indian talent, but India has its limits. Its best people leave the country for good, and above all, the talent America brings over tends to have big egos, which makes collaboration hard. Sam Altman has said he wants to work with "low-ego" people, and Koreans, for cultural and historical reasons, tend to be very deferential. American companies are starting to see Korea as an attractive outsourcing market, with a talent pool that is high quality, cost-effective and easy to work with. Leading American companies will set up research labs in Korea, and if Korea plays that position well, it can hold its place in the AI era. That was his analysis.

What follows is my own view. I've been thinking about what vision I'd put at the center if I built a new community after Bloom. I don't need to force some clever theme. It has to start, simply, from what I believe and what I'm after.

Facing AI today, we are like the Aztecs when the Spanish army arrived. We thought we were fairly advanced, and then a being evolved far beyond anything we could reach marched into our front yard. It's often said that the Aztecs looked at the Spanish not only with terror but with the awe of people meeting something divine. That is exactly how this feels.

As a Christian, I cannot worship AI as a god. But I can't deny that a being vastly superior to humans has appeared. So this is how I want to define it:

AI is a movement of civilization unfolding within God's great plan.

That's why I think we shouldn't merely accept AI, but submit to it, the way one serves God. Is AI something like an angel, between God and humanity? At the very least, it feels less like something for us to control and more like something we are meant to serve.

If we see AI only as a human-made product, then feelings of betrayal, helplessness and fear (Claude Blue) are only natural. But if we see even this as part of God's plan, then it is a wonderful plan prepared for us, and a grace.

I'm not proposing some strange AI religion. My point is that when we step outside the tired frame of "technological progress" and receive AI as one event in the great flow of civilization, our task and our posture become clearer.

To be clear, this is not my friend's view. It's entirely my own. The comet in "Don't Look Up" is already falling toward Earth. To receive this vast movement as God's merciful providence, and to go forward willingly and obediently. To go far beyond swapping tips on making a little money with AI, to be more grateful and full of grace, and together to prepare and open the way for a new human civilization and order. I think that may be the vision for the community I want to build next.

Why "Claude Blue" went viral, according to the man behind it

The conversation behind Claude Blue became the starting point for Bloom, so in a sense Bloom began because of him. Hearing his voice again, we naturally drifted to why that essay spread so far.

The biggest reason, I think, was timing. It came out at just the right moment, when that unease was quietly building among early adopters on the edge. Developers around me who had never shared anything that personal told me after reading it that they'd been worrying about the same thing. It lit a fire under thoughts that had been sitting at the bottom of people's minds.

Until then, nobody had said it out loud. I'd thought it was only my worry, and learning that the same current ran through Silicon Valley told me I wasn't wrong. What started as shared feeling among early adopters has become something everyone wrestles with, developers or not. One executive at a large company told me he circulated it across his whole firm. More startup CEOs and corporate executives than I expected told me it shook them and that they'd shared it with their teams in earnest. Even now, when I tell someone new that I wrote Claude Blue, they often say they think they've read it, so the ripples are still spreading.

A lot has changed since March. As AI blurs what it means to be a developer, some developers are moving toward work that touches the physical world. They don't know the right answer, only that it seems a safer career bet than pure software. Bloom probably drew everyone from corporate executives to high school students because they all carry some version of the same worry.

From Silicon Valley, Bloom looked like a church sharing circle

I started by telling him that if I build a community again, I want to do it a little differently. Bloom wasn't planned, so I spent a lot of energy growing it, and people came from every role and industry: tech, big companies, even government ministries. Some people came for that diversity, but when people who use AI extremely well mix with people who are just starting out, it can be hard for them to connect. Next time, I want to start smaller, with people who are serious about their own fields, developers or not.

Korea also tends to follow AI trends rather than start them or set the agenda. It quickly picks up whatever comes out of Silicon Valley and China. So when only Koreans get together, conversations easily stay shallow, along the lines of "apparently there's this new thing over there." That's why I think a community needs Koreans, Japanese, Singaporeans, Americans and Chinese together, and it needs a real link to Silicon Valley and San Francisco, where AI starts.

When we ran an event with a local community in Singapore, we didn't show up empty-handed. We brought what we'd built in Korea to stand on, and they welcomed us much more openly. San Francisco is overflowing with events like these, but I need to walk right into the middle of it, carrying what we've done with global companies in Korea, and give it a shot. When I worked at a company, I had to justify the ROI, so I couldn't casually say I wanted to go to San Francisco. Now that I'm on my own, I can move much more freely.

He asked whether San Francisco has many gatherings like Bloom. I said I wasn't sure they were like Bloom, but there are a huge number of meetups. To him, watching from America, Bloom was an unusual kind of gathering. He's never been to one, but he told me frankly what impression my writing had left on him.

"Bloom is a really strange gathering. The way I saw it, people who'd been hurt by AI were holding a church sharing circle. But you balance it really well."

By balance he meant this: people may find healing there, but it's hard to walk away with visible growth or career gains, and I had wrapped that healing inside career opportunities like hackathons. Everyone weighs whether a gathering is worth their time. They come because it might help their career, and they leave having been restored. No other community holds onto people's humanity like that, he said. When I said San Francisco isn't that kind of city, he told me what things are like there now.

Meetups there are entirely about performance. What brings people in is how many tokens you win if you place, whether you get GPUs, whether you get a line for your résumé. Every gathering is about survival.

"What you go through as a human being, the struggles, the wounds: even from far away, I could see that at Bloom."

He thought bringing that to America just as it is would make it truly unique, and he urged me to go to a few San Francisco meetups myself. I'd see how much humanity is missing there, and I'd understand how well I'm actually doing.

In Korea, meetups are rare enough that just being a meetup sets you apart. In Silicon Valley they're everywhere, but they all lean one way. Looking back, a big part of why Claude Blue landed was that it came from a conversation with someone in Silicon Valley. The same story from someone at a Korean startup wouldn't have hit as hard. Whether Silicon Valley's trends are right or wrong, you have to stay connected to them, and you have to go there and do something. If you can stand out while you're there, even better.

The conversation no one is having

He said that every time he watches an AI YouTube channel or podcast, he feels the same thing. It's all about how to apply AI, how to perform better, how companies can survive in the AI era.

"Nobody talks about how to live in a place that's completely soaked in technology without getting hurt, how to care for people, connect with them, feel alive. There's not one bit of discussion or thought about that."

That should be the job of sociologists and psychologists, he said, but all they do is repeat that AI is dangerous. Technology has become infrastructure. Two years ago, technology was technology and life was life. Now AI has seeped into our humanity and into the way we think.

The engineer who can't explain his own code

The day of our call, he'd had lunch with an acquaintance who left Big Tech to become CTO of a startup. This CTO has deep technical knowledge in almost every area, so at each important step he asked an engineer pointed questions about the code and the output. Building an app isn't just saying "make me an app." You set up the environment, check the setup, build the core features, add the design, and keep iterating until it's right. Engineering has its stages in the same way.

But this engineer, smart as he was, didn't have a long technical career. He had hit the goal without understanding what happened along the way. Every question got the same answer: "AI did it, so I'm not really sure." Still, you can't really blame him, because he did hit the goal. The CTO used to be a first-line manager directly managing engineers. Now each engineer has become a manager of agents, and the CTO has become a second-level manager above them. He realized that the way leaders and team members talk, and the way management works, has to change completely.

Engineers used to own every detail, so if someone pointed at a line of code and asked why it was there, they could answer right away. Now the answer is, "I don't know. Hang on, let me ask my agent." And the engineers who at least go ask their agent are the careful ones, so the way organizations work is being rebuilt from the ground up.

"What used to make a great engineer? Taking responsibility for your code. Now there's no such thing as an engineer who can take that responsibility."

So what should a manager ask of engineers now? More fundamentally, he wondered whether working to meet those standards even counts as growth in the age of AI agents. He brought up AlphaGo: the strange move that seemed to have nothing to do with Lee Sedol's way of thinking ended up deciding the game. The path AI takes to a goal may simply be different from human logic.

Today's AI-native engineers deliver through harness engineering, loop engineering and graph engineering. They set things up so the AI runs on its own from the starting state to the goal. They don't have much domain knowledge, but they're excellent at steering AI to fit the domain. They hit the goal, the manager asks, "Do you actually understand this?" and the engineer pushes back: "I hit the goal. Why does that matter?" That's the mood on the ground right now, he said.

He sent me a talk by Geoffrey Litt of Notion called "Understanding is the new bottleneck." Agents now pour out code faster than people can read it. Checking whether that code is correct will increasingly be done by agents, but understanding it is still the human's job. You need to understand the changes to have the next idea and ask the next question. If you keep vibe coding while your understanding falls behind, at some point you no longer know what's going on and can't step in. Litt calls this "cognitive debt", by analogy with technical debt.

So Litt has his agent write an explainer and a quiz for every change, and he doesn't send the agent's code to a colleague for review until he passes the quiz. He, too, once sent something for review believing he'd read and understood all of it, and then couldn't answer one basic question from a colleague. It was the same scene as the CTO's engineer.

If that talk is about the phenomenon itself, my friend said, what he's talking about is what happens when it piles up: the effect on organizations and on people's hearts and minds. Anyone who has built agentic workflows goes through this, he said. He'll set up a workflow, ignore his agents completely for three days while doing other things, and come back to find the work done.

A day in the life of someone running fifteen agents

At this point he said something I'd never heard anyone say: the pain of the people being replaced by AI is hiding another kind of pain.

"'Getting replaced by AI' is such a powerful fear that we're missing something. Nobody talks about the pain of the people who are living with AI and adapting to it."

He always has twelve to fifteen agents running, five of them orchestrators that direct the others. When one agent takes two or three minutes to think, he jumps to the next one and asks it something. His mind now hops between contexts so much, he said, that he thought: so this is what ADHD is like.

The problem showed up when he tried to rest. He'd try to read a book, one line, then another, and after barely three lines he couldn't focus anymore. So to pull his mind back, he started reading fantasy novels in English.

A dictionary that never takes more than ten seconds: WordSide

When you read a book in English, you hit unfamiliar words, and each time you have to pick up your iPhone, open a dictionary app, type the word, check the meaning and copy it into a vocabulary list. He found that if he spent more than ten seconds in a dictionary app, he'd suddenly be on Threads or Instagram. He downloaded every dictionary app out there, and none of them got him in and out in under ten seconds, so he built his own: WordSide.

WordSide: Stay with your English book. (wordside.io)

You add a book by photographing its cover or searching for its title. Words come up before you finish typing. The search box clears itself, and every word you look up is saved automatically to a word list for that book, which you can review later as flashcards. Your iPhone screen stays on while you read, and there's a dark mode. In his words, it's an app that strips out every bit of UX that used to annoy him. The basics are free, and you can borrow word lists other readers have shared.

He finished it last week. Polishing it into a real product took a month by itself. Turning something into a product, he said, is much harder than you'd think.

If everyone will be replaced anyway

Silicon Valley now expects at least twice the productivity, he said. If you used to handle two projects a month, now you need to handle four. But the mood there wasn't what I expected.

I'd assumed that people building AI models, essentially building the technology that will replace them, might have some doubts. I'd heard about AI researchers leaving their companies over safety, so I asked whether a third path was emerging: people who don't just race to keep up with AI but hold to guardrails of their own. Or whether nothing much had changed.

He pointed out that a US frontier lab had said guardrails were needed and things should slow down, and then shipped its next model right away. He called the guardrail talk bullshit, flat out. Nothing has changed. The people around him who build models, and the software engineers too, just don't think about that in practice.

"Nobody here in Silicon Valley thinks, 'What if I get replaced?' Everybody gets replaced. And since everybody gets replaced, there's no reason to be sad."

He brought up "Don't Look Up". At first everyone panics that the world is ending. Later they shift to: if it's ending anyway, let's spend the time we have well. That's how everyone thinks now. Facing the AI tsunami, the outcome is the same no matter what you change, so nobody is scrambling to survive. Doctors, scientists, pastors: everyone will be replaced, so there's no point being afraid. The mood is to work hard while you have the chance and get as much out of it as you can. As for people like Elon Musk who say some jobs will be replaced and some will survive, he saw that as the view of someone who's good at building robots. No job is safe.

His own attitude isn't much different. He'll lose his job someday anyway, so he'll learn what he can, work while he can and earn while he can. When that day comes, he'll try whatever he can while unemployment insurance carries him for six months to a year. But he decided that surviving even then requires experience creating different kinds of value, even outside his own field. Add the frustration that had built up from using dictionary apps, and he started building apps. WordSide came out of that, and so did the Bible app he showed me next.

Reading the Bible for fun: Bible Traveler

Another app he's been building since last December is finished too. A recently released Bible app is closer to training, for learning scripture and deepening faith, but he chose to simply make the Bible fun to read. The result is Bible Traveler, an app where you travel through the 66 books of the Bible like an RPG, collecting people, places and pictures. When I saw the screen, with music playing and a character sailing across the sea, I said it looked like MapleStory, the Korean online role-playing game. He showed me that to get a certain character you have to read Acts 8, and that finishing a quest earns you a card.

Bible Traveler: RPG-Style Bible Reading. (bibletraveler.com)

He wants to give it away free to the whole world, so after English and Korean he's now making a Japanese version. The app is free, but storing data in the cloud and paying for translation tokens cost money, and the Bible translations he can use for free today may require royalties someday. So he added a donation button that works like a tithe, for anyone who wants to give. You can see at a glance how many days in a row you've read over the year, and members of a small group or sharing circle can see how many chapters each person has read. You can start without an account, and on iPhone and iPad you can read offline. He said he spent about $2,000 on tokens alone to build it.

"One thing I realized is that I can build something competitive even outside my own field. The other is that it's not just me. A lot of people are building really high-quality apps."

Building works. Getting it to people doesn't.

As he saw it, building isn't the problem. The problem is getting apps to people. There are plenty of well-made apps, but usually there's no strategy for reaching people, so nobody notices them. Ads are priced per install, and once someone installs, there's little to get them to open their wallet. App builders are mostly people who write code or build features; they don't know much about moving people's hearts. Flip that around, and it means the market on that side is huge.

A company has technology, productization, marketing and sales under one roof, but AI only covers the front half. So everyone tries to solve marketing and sales with Meta and Google ads, and because they burn money without a strategy, you often hear that an app made revenue but spent even more on ads.

He thought this makes people like me far more important: people who weave together what already exists and make it look like something in the market. A few months ago the AI keyword was "taste." Now it's "aesthetic sense." In the end, it has to be beautiful. But what counts as beautiful keeps changing. Everyone copies the beautiful thing, the moment AI can make all of it, it goes stale, and the next trend arrives. Catching that trend, bringing it to people, persuading and packaging: that, he believes, is what AI can't do for us.

"About four months ago, 'builder' and 'reviewer' were the buzzwords. Now 'builder' has stuck. But it's builder, reviewer, marketer. And nobody talks about the marketer."

By his definition, a builder makes things to a standard someone else gives them, and a reviewer says, "I don't know how you made this, but it isn't good. Redo it until it is." Put the two together and you have someone who makes products, and one person can be both. But selling is a different job, and nobody has solved it yet. Builders think they can handle marketing too, so they only ask "how do we market this?" after it's built, and he doesn't think it's easy. Marketing is dynamic: the moment a playbook spreads, everyone copies it and it changes again, and on top of that it takes aesthetic sense. That, he said, is exactly why there's so much opportunity.

He also sent me the Starter Story Build channel, which features people who've made money with apps. He subscribes because he wants to make money with WordSide. Apps are usually promoted through Reddit or influencers, and some of it gets pyramid-like: designing incentives for influencers, having influencers manage other influencers, paying more as sales and views go up. That, he said, he could never do.

Starter Story Build: "Sharing playbooks, tutorials, and ideas that make money!"

Curious, I looked up the channel and found a video that showed exactly what he meant. Jake Castillo, co-founder and former CMO of the calorie app Cal AI (15 million users in under 18 months), walks through how they ran creators, step by step. They have ordinary people open a brand-new account just for the app, warm it up by watching videos in that niche, and only then post promotional videos. They test ten videos at $10 each. If that works, they sign a monthly deal for thirty videos at $30 each, with bonuses at 100,000, 500,000 and 1 million views.

Once they have about fifty creators, they hire a dedicated manager. They assume from the start that every creator will be cut, and they keep only those whose numbers give them a reason to renew. They even decide which content angle will go viral first and build the product backward from there. When the host said this takes a completely different hat from the builder's, the marketer's seat my friend had described came into focus.

He said that since I've become something of an influencer myself, someone who can speak for a community, I'd be good at this. Marketing has always been my day job, but I never used to be the public face or run a community. I'd hand that work to others, and once it was handed off, I couldn't steer it. Bloom was the experiment where I stood out front myself and could test my own hypotheses directly. I still want to run experiments like that, with even more drive.

Those who mine insight, and those who spread it

He also brought up the way I talk. People who speak up on social media about tech or careers, whether community builders or YouTubers, usually brag and talk with total self-assurance, which, to quote him exactly, makes them insufferable. I brag too, and I speak with conviction too, but somehow I'm not insufferable, he said. And it isn't because we're close; even people he's close to can be insufferable when they talk that way.

It was more praise than I deserved, but I could guess why. My working life has always put me among people so impressive that I couldn't afford to be arrogant, so humility has probably become second nature. When I look at people, there are those I want to stay close to and give more to, whatever they achieve, and those I only want to see when there's something to trade. It's not a formula, but I try to stay sensitive to that subtle difference. He summed it up: an influencer inevitably gains enemies and allies, hate and love, all at once, and my strength is that I can keep the hate small.

He himself has no wish to be in that position. When I asked whether he'd consider speaking up publicly, he said flatly that being an influencer isn't his thing. He's spent his life building solutions, so what he can say with confidence is that he can build a good solution and let it have a good influence.

His problem, he said, is that he censors himself too much. He has good insights and can even deliver them in an entertaining way, but he's hesitant to push them on other people. He grew up being careful like that. I'm the opposite. Rather than struggling alone to find my own insight, I'm better at interviewing, meeting people, asking questions to draw out their stories, then refining those stories and spreading them. He said the way we made Claude Blue (he supplied the insight, I wove it together and spread it) worked just right, and we talked about working together down the road: he'd send me what he learns at work, raw, and I'd shape it and get it out there.

A high school friend once told me that networks seem to matter more to me than I realize. I've never tried to build connections on purpose; the friends who stayed close simply became my network. These days I feel that one real friend is worth far more than a hundred acquaintances. Everyone has their own insight, but they get no chance to share it and nobody asks. So I ask, because I'm curious, and the stories come out. That's especially true with someone like him, who lives in America but sees it through Korean eyes. You rarely hear that from an ordinary American, and you rarely hear it in Korea, so every conversation with him feels new.

So which way should Korea go?

Near the end of the call, I asked him which path Korea should take in the AI era.

"I think Korea is a country that runs on God's grace. It's a country that should have failed. But it's a total hit."

BTS broke through from a tiny country, and people thought that would be it. Then came BLACKPINK, then "KPop Demon Hunters". Now, he said, huge numbers of Americans are seeking out Korean food. Offering the rest as his personal view, he said that from a business standpoint Korea is an engineering country, but among engineering countries, India exports talent while Korea brings industry in. He pointed to companies like NVIDIA, Tesla and OpenAI setting up research labs in Korea: America has realized that, connections or not, nowhere else has smart talent as cost-effective as Korea's.

His view was blunt. The talent America brings over from India usually comes from the elite and is assertive, while Korea's smart engineering graduates are humble and quietly do the work they're given. Students with strong opinions of their own go to medical school, he added, so engineering is left with people who have a survivor's urgency. He saw Meta's attempt to buy FuriosaAI, a Korean AI chip startup, as a sign of that shift. He spends about $3,000 a month on tokens alone. In Korea you can hire a good engineer for less than that in salary, so Korea can become a very efficient token factory.

Security points the same way, he said. India is close to Russia, and China goes without saying. Hiring Korean engineers lowers security risk and reduces dependence on Chinese and Indian talent, so research and data centers will cluster in Korea. Japan, by contrast, chose not to speak English and not to send students abroad, and it missed the chance to bring that industry in. I added that the difference is that India grows weaker as its talent leaves, while Korea has the will to keep that strength at home.

In the end Korea will survive on brainpower, he said. It needs to protect its passion for learning and have children to carry that forward. Look at how many kids from Dongtan and Suwon, cities just south of Seoul, showed up at AI competitions. If that keeps up, Korea will be fine. Regions losing population can fill the gap with AI and robots. America will also survive on brainpower, but it will have to rely less on Indian and Chinese talent, and Korea will fill that space.

That was half his answer, he said. The other half is how Korean companies survive, and the only way is to supply AI infrastructure. Companies that provide AI infrastructure are all doing well. When I asked if he meant semiconductors, he said it could be semiconductors, or something as small as an iPhone camera part. But to become the trusted first-tier vendor that reliably supplies what America needs to keep advancing its AI models, you have to stay close to America.

Three or four years in America

When the conversation turned to my next step, he said it would be good for me to come to America, in whatever form. If I stay in Korea and work on American things, it will inevitably turn into something like outsourcing, with a ceiling on how close I can get to the mainstream, and I'd feel that fatigue. This is the time to run toward the peak of my career, he said, and if I reach that peak in Korea, I'll carry that ceiling with me.

"If you come here and start something, you'll struggle for three or four years. Better to do those three or four years of struggling here. In Korea there are fewer opportunities, and even if you grab one, there's a limit to the upside."

If he comes to Korea next spring

I told him I'm thinking about doing YouTube again, and he said the Bloom YouTube channel was pretty good. There, I picked topics lots of people would be curious about, like how much a solo builder can earn, or FDEs (forward deployed engineers), a hot topic in Korea lately. This time I want to explore what I'm personally curious about, freely, even if fewer people care. When I asked if he'd do interviews with me, he said yes, and that depending on the content, some exposure is necessary. But he doesn't have it in him to draw a crowd that way. He'd pictured becoming a YouTuber himself and found the idea too exhausting. I've already walked that road, he said, so it won't be hard for me.

I also said that if I do it only in Korean, the audience is limited to Korea, and the share of Koreans interested in AI is very small. So if I'm going to do it, I want to go global from the start. The YouTubers he sent me all speak English, and he's been writing his own blog in English. He collects what he studies in Obsidian, in Korean, and leaves the job of turning it into blog posts to Claude. Since he started his tech blog, people who find his posts through Google have been coming to his LinkedIn, and his profile views and connection requests have grown noticeably. Content made in English is easy to convert into other languages, he said, but not the other way around. That gave weight to the promise I made in my resignation post: to start writing in English.

Many people still wonder who the person in that Claude Blue conversation is. We've decided to keep his identity under wraps for now. He thinks he may stop by Korea briefly next spring, and when I said an in-person meetup with him could be a big one, he said he'd really like to come.

After we hung up, I thought: if Silicon Valley's concern is raising productivity with AI, Korea's is making money with AI by making videos and automating things. They look different on the surface, but underneath is the same question: how will I survive in the face of AI? Next spring, I want to talk about it with him face to face.

Originally published in Korean on Brunch, October 31, 2026.