Bloom Member Talks: San Francisco Is 30,000 Fake Elon Musks

This time we ran a Bloom meetup with no corporate partner, and only our own community members as speakers. The truth is, the Bloom stage is open to anyone, and we want to bring up as many people as we can going forward.

Kyutae built Mirr AI, a social media automation tool that got a lot of attention on Threads, and has come to our events several times. Later, over a coffee chat, they walked me through the social media ecosystem in depth. Then, wanting to see a bigger market for themselves, they left with their co-founder for 50 days in San Francisco, New York and LA. They shared, very candidly, what shocked them in a good way on that trip and what turned out to be surprisingly disappointing. I loved how they described San Francisco: "a city that feels like 30,000 fake Elon Musks gathered in one place." But they also said some of them really are next-Elon-Musk material, so you can't write it off.

Celina has walked the full elite track: MIT, then Wall Street, then Berkeley Law, then a New York law firm, then an AI startup. Yet she says she never stopped wrestling with her career at every step. With AI now digging deep even into legal, long seen as the most conservative field, she talked a lot about the mindset and attitude it takes to move from Blue to Bloom. She coaches clients on their careers in Korea, the US and 20 cities around the world, and surprisingly, she says everyone carries similar worries. So her advice really resonated with me: don't overthink it, just start by doing one thing.

Bloom keeps running AI community meetups one to three times a week, with about 100 people each time. Some people tell us it must be exhausting, but it's become a routine for us, which makes it more fun, and I get a lot of energy from it too. Thank you again to both speakers.

That day, it was two of our own members

To start, I asked people to introduce themselves. One studied international trade and leads an overseas sales team, using AI to build everything from work automation to websites and apps, and came to stretch their thinking because what they build probably feels like a tiny piece of a tiny piece. Someone building a service with AI on their own. Someone who'd just been put on an audit automation task force at an accounting firm. A recruiter at a blockchain company. What a researcher at an international organization said stayed with me. They think the professions of strategist and lawyer are both in decline, and came because they wanted to know whether AI could actually save them.

We didn't bring in a company that day. The week before we'd done an event with Snowflake, and the next week we were doing one with other companies, but this one was purely two community members I picked myself. The key was that their backgrounds were opposites. One had been building a service in Korea and had just come back from the US. The other had spent her whole career in the US, through law firms and Wall Street, and was in Korea for a while. New York and San Francisco are different, and going from Korea to the US is different from coming from the US to Korea. I thought that setup would be fun.

A 40-million-won toy project

First up was Kyutae Park, CEO of Mirr AI.

It's a social media marketing solution for small businesses. Put simply, it's an AI agent that runs your Instagram for you. It helps with the full cycle, from planning to production to account management, and over 90 percent of its customers are Korean. Two people build it, and monthly revenue peaked at 40 to 50 million won (roughly $30,000 to $37,000). That's a great number when you think about a typical salary. But on LinkedIn they call it a toy project, so I asked what's behind that. They said they still think of it as a toy project. They started as an indie hacker, built something they needed, used it, and customers came in and paid. In a way it's a pretty ordinary story. But in the process of growing it into a business, they realized they hadn't been looking at the 3Cs: what workflow customers really use it in, who customers see as the competition, and why we're the ones who can do it well. Until they have answers to those three, they said, they'll call it a toy project no matter how much revenue it makes. I was struck that with revenue coming in well, they don't count it as a win and are going back to check the fundamentals.

30,000 Elon Musks

I started by asking why they went. The biggest reason was that global software has never come out of Korea. There are a few well-known solutions in their domain in Korea, but go abroad and nobody's heard of them. Nobody knows Naver or Kakao (Korea's biggest internet companies) either. They figured at this rate it would end up a domestic-only tool, and at the very least they wanted to see with their own eyes and decide whether to stay domestic. They kept saying "we should go, we should go," until their co-founder hit them with "how long are you going to put this off?" They decided over a meal and bought tickets that day. They said they'd been a Silicon Valley hater. They assumed there'd be nothing special there. They went, and there was something special.

What made it special was the sheer density of talent in the offline community. In Korea, fewer than 10 companies seriously do what they do. There, there are 2,000. So at any meeting you go to, there's always at least one person who's already dealt with the problem you're dealing with now. In Korea, just finding that person and getting a meeting takes a long lead time. There, you can stop anyone on the street and they're a founder. One story: on a train, the person in the seat in front asked for the Wi-Fi password. They chatted a bit, and it turned out to be the founder of a hot company worth about 1 trillion won (roughly $700 million). The thing people call serendipity really exists, they said. I've heard this story a lot too, and I think you can't really get it until you experience it.

There was plenty of disappointment too. They felt there were a lot of frauds. When I asked what percentage, they said 95 percent looked like frauds. Most people meet investors with every metric inflated, and they'd heard secondhand about a founder who put on a show in front of investors, saying they'd kill themselves if they didn't get funded. Investors find that entertaining, so the money goes there. Lots of money flows in, and frauds follow the money. That was their conclusion, and I thought it was pretty logical. The global market we talk about isn't Silicon Valley. There's clearly a bubble unique to that place, mostly a side effect of all the capital pouring in.

So I asked how they'd define Silicon Valley, and this was the answer.

"Put 30,000 people who want to act like Elon Musk in one place, and that's San Francisco."

I said, isn't that a good thing? They said there's a real Elon Musk among them. One percent are real and 99 percent are fake, and when people like that gather, you get an interesting ecosystem. Your mindset shifts just by going. At any meetup you'll meet at least one person better than you. In Korea you can go your whole life without meeting someone like that if you don't want to. There, if you go out, you will. They also described it as a city where software is mainstream, with software ads on every bus. That stuck with me.

Founder blues, and escaping in a day

People who know our community's story might remember this: while they were in Silicon Valley, they posted that they were feeling a bit of the blues there. I asked what those blues were. X and LinkedIn are like Instagram stories, they said. Everyone you meet is already playing at being Elon Musk, says how much they raised, and seems to be making an impact on the world, and your whole feed is nothing but that. It's the same structure as Instagram stories where everyone looks happy, so the founder blues hit hard. They escaped within a day of getting to New York. Statistically, it makes no sense that every founder you meet is doing great. I thought living there long term could be really hard on your mental health.

The comparison of the three cities was fun too. If San Francisco is vision, vision, vision, founders in New York run much more practical businesses. They took the same business to both and got opposite advice. In Silicon Valley the tone was "your dream is too small, what's the point of being No. 1 in AI content marketing?" In New York it was "if your customers are in Korea, wouldn't you be better off raising money in Korea?" Silicon Valley, with a tech scene built by immigrants, seemed more welcoming to foreigners, they said. But if they were building a global product, they'd actually meet fewer founders in Silicon Valley, because it drains you and you'd get swept up in the overvaluation. Overhype was one of the keywords of the night.

I also asked about the decision to put their co-founder, who'd been in Korea, on a plane to the US partway through. There were two reasons. The Korean business is a market they know, so they can play it remotely, and if the Korean business can't run remotely, the US business was never going to work in the first place. And the co-founder is more aggressive than they are. In Silicon Valley, 30 people line up in front of one high-profile person and wait 30 minutes for a one-minute conversation, so an aggressive person who can make an impact in that one minute has a clear edge. Surprisingly, the most important networking skill turned out to be listening. When five non-native speakers mix in one conversation, listening comprehension gets brutally hard. You can always piece together something to say, but you need to understand at least 70 percent to fill in the rest.

I also got tips for anyone buying a ticket next week. San Francisco has 10 to 15 meetups a day on Luma, so sign up early before you get bounced off the waitlist. Go to vertical meetups that overlap with your interests. At a builder meetup without that overlap, everyone just tells each other "you're the best, no, you're the best," and afterward you can't remember who built what. Koreans tend to lean a lot on EO (a Korean startup media and community brand), and they did their first-ever English pitch there. But what stayed with them more was the Koreans who were doing well. Seeing someone with the exact same background get that far motivated them much more than Americans doing well. Korean founders also had by far the highest response rate to cold emails. And it was there that they first felt BTS really is world-class. There was a BTS listening party on San Francisco's Luma.

A city where token maxing is a session topic

I asked what's most hyped in San Francisco right now. Sessions on how to use Claude Code or Codex well happen almost every week, they said, and what stood out was token maxing. How do you evaluate who's good at using AI in a company? What do you do when token bills get too high? Those get their own sessions. If people are discussing token maxing, it means a lot of companies are already pushing AI hard. Here, many companies are still debating whether to use it at all. As people keep delegating to agents, there really are cases where it ends up costing more than salaries, and some even say it's better to hire more people instead. I also liked their account of a room where a CPO, a CTO and a CEO had completely different views. The tech side said, as long as you hit the goal, isn't that fine, and if you don't do this you'll fall behind. The CEO side said costs have to be cut.

As a meetup organizer, I picked up one format I'd love to have in Korea. A company that sells insurance for startups runs a café, and the café is more famous than the actual business. It's a café just for startups, so anyone you strike up a conversation with is a builder. Insurance is a product that needs an opener to sell anyway, and people who fit the ICP exactly show up on their own, which even gives you a reason to start talking. They called it genius, and I agreed. When I looked it up while writing this, the company was Corgi. It's a Y Combinator company that sells insurance just for tech startups, including AI liability coverage, and its valuation went from $630 million in January this year to $4 billion in July. The café was a happy accident: to rent the fifth-floor office, they had to lease the ground-floor retail space too. The CEO compared the café to a watering hole on the African savanna. That's where the lions wait. VCs actually go there to find people. It's not a profitable café, though, and the CEO admits they're losing money on it, so it's closer to spending marketing money in the form of a café. Still, if they'd spent the same money on banner ads, no watering hole would have formed. One more thing: the café's address is on Claude Lane. For a community that started with Claude, all we could do was laugh.

Admitting there's no moat

The last question came from an audience member preparing to start a company. With execution getting so fast in the AI era, their biggest worry is a moat. What did they come to think after Silicon Valley? It's a topic they've been chewing on for hours at a time lately, they said. Every software-only company worries about it, and nobody can say with confidence they won't be replaced when Claude ships its next version. Even for a company as small as theirs, five or six copycat products have appeared in Korea alone. We're getting to the point where you'll ask AI to copy something and it'll whip it up with only the colors changed.

So they've decided to move beyond a software-only business. Growing revenue several times over from here seems doable, they said, but if someone pours in crazy money to compete, they can't be sure they'd win. So they're trying to shift toward a kind of business that needs CapEx, something AI agents alone can't easily copy. They thought about moats a lot in Silicon Valley, and their conclusion was an admission: nobody really has one. These days they have one thought: you have to build a business that's tangled up with people.

From Wall Street to career coach

Second was Celina.

She was born in LA, went to elementary school in Korea, went back to California for middle and high school, and then studied management science at MIT. She loved math and science, so it seemed like she'd go into engineering, but a business class she happened to take as a freshman was fun because it applied to real life, so she ended up doing business grounded in engineering. She'd planned to go to law school, but a string of coincidences turned an internship into a full-time job, and she started her career on Wall Street, which wasn't in the plan. I asked how close it was to "The Wolf of Wall Street." Parts of it are similar, she said. It's a place full of people who value money above everything, so money is the goal and money is the value system, and even if you're not really like that, the environment can change you.

Less than two years in, she went to law school. People around her told her to stay because it was a good career, but her thinking at the time was clear. If I work really hard here for five or ten years, I could become my boss. Is that the future I want to picture? Her answer was no. Why she picked a West Coast law school was fun too. After getting in, she toured schools. It was spring, but on the East Coast it was snowing and gloomy, while Berkeley was bright and sunny and people were kind. If school was going to be hard anyway, she figured she might as well do it somewhere with good weather, and it really did turn out to be a place where people helped each other. She's lived in New York for a long time now. California's never-changing weather actually got boring, she said. You have to be a little cold to appreciate the warm days.

She shared one memorable scene from law. During a private equity deal she worked for a week without sleep, and the bill for that week was a million dollars. As she went to get the partner's signature on it, she thought:

"We're an industry that sells trust."

Lots of firms can do the work. What you're paying for is a signature that says, on the firm's reputation and the partner's name, that everything has been legally vetted. That trust and reputation are why a lawyer's time is priced that high. A friend at a law firm once told me they track their time in three- and five-minute increments, so this clicked for me right away.

AI has come to the most conservative industry

With services like Harvey doing well and companies like Legora working with big Korean law firms these days, I asked about the mood in the industry. Of the industries she's been in, law was the most conservative, she said, but in the face of AI, even the legal world seems to think it has to follow the trend. People really are using it a lot. But AI is taking over much of what junior lawyers used to do. Junior work always had a lot of repetitive patterns. There's also a kind of person law firms are actively hiring now: someone who helps lawyers use AI efficiently. Few people have both AI experience and legal experience, so they're hard to find. Hearing this, I thought the profile of the ideal hire is changing. Just the day before, I'd heard BCG is hiring a huge number of developers, since AX projects have to go past consulting all the way to execution. The kind of talent companies want keeps evolving, and if you can't keep up, even a long résumé might not save you.

A number came up too. Nobody openly says it's because of AI, but data showing that first-year, entry-level lawyer hiring declined for the first time since 2014 has the industry buzzing. Juniors shouldn't fixate only on big firms and should look at a range of opportunities, she said. Since the rules are being rewritten right now, a lot of opportunities that never existed before will open up too. Seniors have their own problem. There are cases where a junior does something with AI, a senior passes it along without reviewing it properly, and it causes trouble. In the US, it's all public which lawyer at which firm handled what, so you have to be even more careful.

What doesn't get posted on LinkedIn

An audience member asked a good question. Sometimes the work you feel destined for doesn't pan out. What mindset should you bring to exploring your career, and what do you do when finding the right work feels endless? Celina's answer stayed with me more than anything else that night. From ten years of working as a career coach: there's a US statistic, a bit old, showing that more than 80 percent of people hate their jobs. The reason is that you can't really try a career before you're in it. She started banking and law without really knowing them, and what law school teaches is different from the actual industry. When people around you say something is great, that's how it looks from outside. How you actually spend your days inside is different from what you imagined.

So you start wondering if you chose wrong, and life doesn't go the way you want. But she's coached a huge number of people, and not one of them had never hit a low point.

"Not one person I've coached has never hit a low point. They just don't post it on LinkedIn."

In coaching, she knows when someone is going through a really hard time, but their social media looks glamorous and great. Most people don't have that information, so they feel like they're the only one struggling while everyone else is doing great. That's an illusion. So be a little kind to yourself while you work your way through this. Let go of the idea that it has to turn out a certain way, throw yourself into the things that are fun, and pick what fits. If it works out, great. If not, oh well. In her group coaching she assigns a rejection challenge as homework, and whoever collects the most rejections gets a prize. In the AI era you'll end up trying lots of things, and people with a solid inner core last longest, and if you try a lot, one of them will work. It's building your rejection muscle.

Finally, I asked what mindset helps you ride the wave in times this turbulent instead of getting swept away. She told a story from when she applied to law school. She happened to apply in a year when the economy suddenly turned and news was going around that law schools would cut their classes sharply the following year. She's an anxious person by nature, so seeing that news sent her hunting for even more anxiety-inducing news. She'd go on forums and soak up useless information, "that school isn't admitting anyone, this school isn't either," and get even more anxious. Looking back years later, she has one thought: I should have spent that time writing better essays and studying harder for the test. The AI era is similar, she said. News has to be sensational to get clicks, and because the era is changing, there may actually be more opportunities. When most people are anxious, they think a lot and don't act. From her years of coaching, that's when people are most depressed and most lost. So now that a great tool called AI is here, act more than you think, build something, anything, and learn it with a light heart. It connected perfectly with Kyutae buying a plane ticket over a meal in the first session.

Roundtables: what you know but haven't done

As always, the last hour was roundtables. There were two questions. What do you know you should do but haven't done yet? And if you had one extra hour, what would you do?

A few scenes from the eight tables stayed with me. Someone meditates to get away from the dopamine of vibe coding. Someone had already bought a ticket to San Francisco in October because they want to ride a Waymo. Lots of people gave basic answers like exercise and studying English, since in the end it's a battle of stamina. One table summed it up well: there's plenty of desire to try new things, but the hardest part is turning what you know into action. At another table, opposing views clashed. One side said in a world this good for building, if you don't build fast, someone else will build it first. The other said that exact environment keeps you on the run, so you need rest but can't take it. The most real moment of the night was someone describing how their company gives them a daily token allowance, and since it renews on weekends too, they end up working weekends.

A table where people said they couldn't move because they hadn't found a direction came to a funny conclusion: if you can't get yourself moving, spend money. Like signing up for personal training first because you won't go to the gym otherwise. If you pay and force yourself to a seminar, you'll get something out of it. A table where everyone was already acting talked about how to spend time. One person sets aside time every week to meet people in their industry. Another reads at a café the morning before a work-from-home day. Another turns off all devices on weekend mornings to journal or recharge at a sauna. I said the last one seems the hardest. I know very few people who actually do it. Someone had arrived from Canada the day before and found us on Luma. They'd been a CTO at a startup in Canada, wound it down when the investment fell through right before the final round, and now work on an open-source AI harness. Asked what they know but haven't done, they said "success," and everyone laughed. They also said that outside of work they try to barely use AI in daily life, because it wears them out. That stuck with me.

The last table's presentation was the perfect close. The presenter said honestly that what they know but haven't done is quit their job, and that since their side project has started paying off a bit, they'll probably quit at the end of the month. Then they summed it up. AI has reopened the market. Anyone can start developing, start a company or make content now, so the real wall isn't technology anymore. Developers get stuck on sales and marketing, and students find it harder to study the fundamentals because AI writes all their code. So the competitive edge in the AI era isn't the ability to start. It's the ability to find a problem and solve it all the way through, and what matters is a mind that doesn't break. That table proposed carving out a 40-minute golden hour every week in the evening to build up the experience of finishing one thing. The presenter spends their golden hour boxing, and everyone laughed when they said they'd gotten a girlfriend out of getting hit at the boxing gym. They ended with: take care of your health, because your neck might go before you quit or exit.

In the end, both sessions were saying the same thing

The two speakers that day came from opposite backgrounds. One had been building a service in Korea and went to the US with no plan. The other walked the elite track in the US and is in Korea for a while. One wrestles with moats in software. The other wrestles with people's careers. But they reached the same conclusion: don't just think, move. Kyutae bought a ticket over a meal to see with their own eyes whether to stay a domestic tool. Celina said when you're anxious, don't consume more information, just build something.

Honestly, I don't think we gave people any especially new information that day. A fair amount of what Kyutae said is already out there on YouTube and LinkedIn. But hearing it in person from someone who actually lived it is different from watching it on a screen. I think that difference is why people give up their evening and come all the way to Gangnam. One thing Kyutae said has stayed with me: people who've done it and people who haven't are completely different, and you can't make someone who hasn't done it understand just by talking. So if there's one thing nagging at you as you read this, I hope you just try that one thing. It's nothing much, but almost nobody does it, so if you do, it becomes something.

Last, we recently started something called Bloom Mixer. At our retrospective gatherings, the thing we heard most was that people want to network with the others who come. So we take applications and put groups together based on interests and location. Think of it as a curated roundtable. It could be dinner, breakfast or even a ping-pong meetup. We plan to keep experimenting with the format.

Originally written in Korean on Brunch, August 14, 2026.