My American Friend Who Built a Sequoia-Backed Unicorn

In 2015, I helped out with a Startup Grind event in the basement of MARU180 in Yeoksam, a startup hub in Seoul's Gangnam district. Startup Grind is a global startup event series that Google sponsors in each country. An older friend from college worked at a startup in that building, and while helping them out I naturally ended up on the organizing team. I wanted to do something more before I left for my military service.

I was really into video production back then, so my friends and I filmed the Startup Grind event. That was the first time my eyes opened to the tech startup scene. As a marketing student, I'd assumed I'd build a career at an ad agency, but my interests went in a direction I never expected.

Speakers included people from VCNC (the startup behind the couples app Between) and Altos Ventures. Since it was a global event, it ran fully in English, and beer and barbecue were served American style. There were a lot of foreigners, too.

I started a video production club called Kiwi Productions and helped run the event through it.

That's where I first met Teddy Cross. I was in my mid-twenties, so Teddy must have been in his late teens. Meeting a kid from Silicon Valley in Seoul was a novelty in itself, so we became friends quickly. Even after I went into the army, I'd get in touch with him whenever I was on leave, and we went to concerts together, so the friendship kept going.

After my discharge, I had an awkward nine months before I could go back to school. While looking for an internship, I happened to find a marketing intern position at Buzzvil, an ad tech startup. I liked marketing, and I'd gotten interested in startups after making friends with developers in the army, so applying felt natural. And there was Teddy, working as a software engineer. Another strange coincidence.

Between finishing that internship, going back to school and building a career, Teddy and I somehow ended up knowing each other for more than ten years. Along the way I knew he had founded Playco and was growing it with Justin Waldron, a co-founder of Zynga. But only after doing this interview did I really feel how impressive Teddy is. We'd known each other so long that we talked more about personal things than work, and that's exactly why I heard the deeper story so late.

Playco was founded in 2020 and raised a Series A of about $100 million (140 billion won) co-led by Josh Buckley and Sequoia, with 230 billion won raised in total. It's a unicorn valued at more than 1.4 trillion won (over $1 billion). Teddy mostly goes back and forth between the US and Japan, the two biggest markets for instant games, but because he lived in Korea for a long time, he comes to Seoul often too. You'll see why below.

I casually suggested we do a Claude Bloom event together, and we did a pre-interview with a reporter from the Dong-A Ilbo, a major Korean daily. That was the first time I properly heard the arc of his whole life, not just his achievements. He was homeschooled from childhood, learned Korean after watching 2NE1 (a K-pop girl group), and came to Korea when Buzzvil recruited him. Along the way he started a company, and it became a unicorn.

Two things stood out to me. One was how he sees AI. As a true engineer, he saw AI as a technology when it arrived, not as a business trend. As soon as ChatGPT came out, he sensed it would change the world, and he started using it right away, even when it wasn't very good yet. That finally explained why he'd been stressing AI so much since two years ago.

The other was writing. He tries to hand every task to AI, but he said he never hands over writing. Writing is the only way to express our thinking, and if we hand even that to computers, thinking itself will wither. What stayed with me was his point that humanity gains nothing when what makes us human withers away.

I've written as a hobby since I was a kid, and after AI came along it bothered me that writing seemed to be losing its value. But once I started thinking of writing as a mirror of my own thinking and identity, rather than an output, my attitude changed. That goes for how I feel about this blog I keep writing, too.

That was a long intro. Here's my conversation with Teddy, in as much detail as I can. (For reference, here's the article by the Dong-A Ilbo reporter who did the interview with me: The secret of the young American who built a Silicon Valley unicorn with no diploma or degree, in Korean.)

1. A person who never went to school

Teddy never went to school in his life. Not kindergarten, not elementary school, not high school. Homeschooling laws differ by state in the US, and Tennessee, where he grew up, allows it.

His hometown is a small rural town in Tennessee. He said there wasn't much to do. He didn't have many friends. His mom had been a teacher, so she taught him herself, and his dad worked in software. His dad was a founder who started a company from home and later grew it to about 50 people. Sitting next to his dad while he coded, Teddy naturally got comfortable with computers.

His tools for teaching himself were Wikipedia and books. "Imagine how much more I could have learned if ChatGPT had been around," he said. He learned only what interested him and skipped what didn't. He added with a laugh that this is a bit of a problem too. His favorite subjects were science and engineering. He also named IT, coding, books and swimming.

He said he wanted to be an inventor from a young age, while other kids dreamed of being doctors or firefighters. Building things with LEGO, making games with code, and now building a company: looking back, it was all an extension of "building something."

With no school, he looked after his three younger siblings and did the laundry by himself, starting at age 10. At 15, his first job was as a lifeguard at the local pool. He'd been on a swim team for eight years, so it was a natural pick. With twenty people sharing one pool, he watched over all of it alone. In the US this is a normal teenage life. At 16 you drive, you work, and you get around on your own.

2. Viral at 14, Silicon Valley at 15

He first got into coding through FIRST LEGO League. In 2007 he gathered his friends and formed a team himself. It's a competition where you build LEGO robots and program them to move. He led the team for three years, and in the last year they took second place at the state championship. That's when he really fell into coding.

2010 was when HTML5 was just appearing. At 14, Teddy built a tower defense web game with HTML5 Canvas. He'd loved Flash tower defense games since he was little, and the genre was second nature to him, so choosing tower defense for his contest entry made sense. The occasion was the 10K Apart contest, hosted by Microsoft.

The challenge was to build a complete web app using only HTML5, in under 10 kilobytes of code. Ten kilobytes is tiny, and HTML5 itself was brand new back then. The contest entries can still be seen in the archive.

He didn't win. But he'd set it up so players could post their score to Twitter when a game ended, and the game went viral as soon as he posted it. He also posted it to Hacker News, run by YC, and it hit number one.

Although I didn't win the contest, going viral on Twitter, reaching #1 on Hacker News, and then launching my career from it was a far better result.

People in Silicon Valley got to know him. The original game still runs in the archive, and an upgraded version is playable today at canvas-td.teddy.io.

Then an email came from a web game company asking him to work with them. One of its co-founders was Justin Waldron, who had co-founded Zynga. Zynga made FarmVille and was the biggest game company on Facebook in the 2010s. After the interview, Teddy and his mom visited the company in person to check it out, and he started as an intern.

So at 15, Teddy moved to Silicon Valley on his own. He rented a room in a big house shared with three Stanford PhD students. When I asked if he wasn't scared to live with strangers, he said, "They were Stanford students, so I figured they wouldn't be weird." His room cost $800 a month. It was less a bedroom than a space about the size of a small office. Today you'd have to pay three or four times that in the same area. He found it through a friend of a coworker, at a time when you could ask an email list "Does anyone have a room?" and get nothing. Then as now, people from all over the world were flooding into Silicon Valley.

It was his first time living away from his parents, and he said it was a little scary. His parents encouraged him, saying it would be a big opportunity. Looking back, he said that early independence became the biggest source of his confidence.

3. Instant games, and Playco

Playco launched in 2020 with four co-founders: Justin Waldron, Michael Carter, Takeshi Otsuka and Teddy. Michael Carter is the developer who invented the WebSockets protocol. The real-time communication in Slack, Discord and every messaging app today is possible because of WebSockets. Takeshi Otsuka is a game producer who pioneered Japan's mobile social game industry and led DeNA's Mobage platform. The company formed when people with a similar vision merged their teams. "Similar vision, similar work. We realized if we work together we can build something much bigger."

The mission is clear: no download, no install, the game starts right away from a single link. Social games are the core. He gave Anipang (a Korean mobile puzzle game people played with their KakaoTalk friends) as a good example. You play with friends, share scores, and see who's playing. Now games start inside the LINE messenger without any install. He said a game in Japan that combines pet raising with puzzles is doing well. Like many game studios, they deliberately keep their game brands separate from the company brand.

Solo games are not the future. People want to play social games.

His point was that the long run of solo mobile games like Candy Crush came from the limits of technology and platforms. It wasn't what people wanted.

Timing came up too. When they tried the same idea in 2014 and 2015, it was too early. Phones weren't powerful enough, and hybrid web technology wasn't ready. By 2019 and 2020, phones had caught up, and the timing was right. "Too early is the same as wrong."

When COVID-19 hit in 2020, the timing clicked. Demand for social games exploded, and Playco games went global, like Heads Up! running in Zoom calls and EverWing running in Facebook Messenger. Playing with friends from a single link with no install fit the no-contact era perfectly. That's also when the $100 million Series A, co-led by Josh Buckley and Sequoia, came in. Josh Buckley founded the game company Mino Games with YC backing, sold it, and then became a solo VC. He was an early investor in Figma, Rippling and others. Being a former game founder himself, he was a good fit for Playco.

Someone asked what the hardest moment in a startup is. He laughed for a while and said that's the hardest question, because there are too many challenges to pick just one. Then he put it this way:

Running a business means a team overcoming endless challenges together.

He said it's short and simple, but it's the most honest answer from his own experience.

4. AI is not a bubble

When ChatGPT first came out, Teddy saw it right away. "As soon as I saw ChatGPT, I thought this is gonna change the world." But the market wasn't ready. That was still true a year ago, and now everyone is excited. It also shows that timing is everything.

At his company, they kept trying AI even back when it wasn't good. The art team went first. First backgrounds, then characters, then animation. Each month, the range of what they could use it for grew a little. After two years, AI was handling almost all of the art. The coding team was the same. "Six months ago everyone was still writing code. Now nobody writes code at all." ChatGPT came out three years ago, and three years later, coding has stopped. Nobody knows what three years from now looks like.

Because they kept practicing even when AI was bad, they could jump on right away when it got good. They were a technology-first company in mindset, so they took AI in naturally as just another new technology.

Our use of AI is very similar to our use of web technology. Bleeding edge, but very normal for us.

On the bubble debate, he said this. Some dominoes may fall. Some infrastructure investment may have overheated. But that doesn't change the fact that the technology itself will change the world. Compared with electricity, the internet or the iPhone, AI is spreading at a completely different speed. Electricity took decades to reach every home, and the iPhone took three years just to arrive in Korea. AI is software, so it reaches everyone at once. There's no language barrier either.

How will this change the world. Not WILL this change the world.

I brought up something a teammate once said: within five years we'll get either utopia or Skynet. Teddy's answer was "Nobody knows." But the fact that this is even the question shows it's already a given in Silicon Valley.

He said something interesting about Korea. By ChatGPT usage, Korea is in the top three to five countries in the world. Koreans catch trends fast, he said. But he added that the important part is that this is a change, and calling it a trend misses that.

This isn't a trend. It's a change.

5. Writing is the one thing he won't give to AI

Teddy said he isn't a good writer. He was homeschooled, so he never formally learned writing. He uses AI for internal documents and customer-facing posts. But anything that carries his own thoughts and feelings, he writes himself. He doesn't let AI write it. Messages to employees and outside communication alike, he sends raw, by dictation. He barely even checks grammar.

Here's why.

Writing is just the expression of our thinking and thoughts. The lasting thing humans can do is writing.

When AI writes, you get safe, correct writing. That's the problem. AI writing is safe. It has no unusual phrasing and no risk. When Teddy writes himself, the grammar might be wrong and some phrasing might be odd. That's him. The moment AI becomes a filter, the person disappears. "If I put an AI in front of me, that's a filter. That's not me."

He said you can tell right away when AI wrote something. "It doesn't feel authentic, it doesn't feel sincere." If someone sends him a message written with AI, he can always tell.

There was deeper context. Growing up, his family had no TV. They watched a movie together once a week, but there was no TV in the house. These days when he goes out, the babies are all staring at screens. He sees blind faith in AI as similar to depending on screens. Losing critical thinking is the biggest danger, he said.

If you leave your thinking to AI, it'll feel easy and nice at first, he said, but you'll end up not even knowing what kind of thinker you are.

If we stop thinking, we're in real trouble.

He said they don't just trust AI output either. "We don't trust AI." They always review it, and if they think it's wrong, they throw it out.

6. Why Seoul

He's lived in many cities: Tennessee, Silicon Valley, Shanghai, Tokyo and Seoul. He started by saying that the right city is different for each person. With that said, Seoul is the one that fits him.

His ties to Korea go back a long way. He first settled in Seoul while working at Buzzvil, the ad tech company. He lived there about four years, then moved his base to Tokyo. When he started Playco in 2020, he came back to Seoul. Now he spends about 70 percent of the year in Seoul and 20 percent in Tokyo. The rest is in other places, including Silicon Valley.

When he first came, it felt like a city of the future. New infrastructure, great subways, delivery for everything, energy around the clock. But what struck him most was how safe it is. That didn't seem important when he first arrived, but living there changed his mind. "Americans can't understand how people in Korea leave their bags at a Starbucks and walk away." Seoul is a place where that's just normal.

He also said Korean people suit him. He compared it with Tokyo. He works in Japan often, but culturally it doesn't fit him. Koreans have energy and are kind, he said.

He started learning Korean because of K-pop. It was the time "Gangnam Style" was blowing up and a movie starring Jun Ji-hyun was getting famous. He'd never learned a foreign language in school anyway. Looking around for a language to learn, he chose Korean. He learned Hangul at a language school, so by his first trip to Korea he could already read it.

7. How to read trends

I asked his secret to catching trends. The first thing, he said, is staying connected to Silicon Valley. Many of his friends are still there, and whenever they meet, that's what they talk about, so he always sees what's at the very front firsthand. Engineers there have to learn new technology every day to survive, so following change quickly has become second nature to them. Avoiding it isn't an option.

Someone asked what to do if you don't have Silicon Valley connections. He said that even before he had them, he just consumed content hard: blogs, YouTube, Hacker News, Twitter. He followed people who actually work in the field instead of AI influencers. Hands-on developers who became known by building things themselves, like Steve Yegge or Peter Steinberger, the creator of OpenClaw. He tried the projects those people made and built side projects with the same technology himself. These days he listens to a lot of podcasts. He also makes good use of the X algorithm. Once you start watching AI content, you only see AI.

Learn first and learn fastest, and you get an advantage.

He doesn't follow particular podcast shows. He picks by episode, using three criteria: the guest is interesting, it was a hot topic on X, or a friend sent it. He said AI episodes stay relevant for months. News goes stale by the next day, but AI podcasts are philosophical and predictive, so there's still plenty to think about long after.

He named the episodes he keeps going back to: Dario Amodei's conversation with Dwarkesh Patel, Andrej Karpathy on No Priors, Steve Yegge on Latent Space, Richard Sutton with Dwarkesh, and Sam Altman with Satya Nadella. Each of them builds, designs or theorizes about AI. He listens to them speak for themselves.

Beyond content, he makes time for offline too. Chance conversations at coffee shops, small gatherings cooking dinner together in a friend's kitchen. He said he gets a lot of insight from those.

8. The experience of making comes first

The Dong-A Ilbo reporter who joined us brought up their son. He loves games, but they couldn't tell where the will to make games, rather than just play them, comes from. They added that Korean parents tend to dislike games.

Teddy's answer was LEGO. He loved LEGO as a kid. In the US there are a lot of toys kids build themselves, and LEGO is the classic example. Once making things becomes fun in itself, coding comes in through the same door. Practicing building what he wanted with his own hands from a young age carried all the way to today, he said.

Even more important, he said, is the chance to try many different things. He swam for eight years, and if he'd never tried it he wouldn't have known how much he loved it. Everyone ends up with things they love, but to find out what they are, you have to try first.

The opportunity to try many different things is important. Because if you don't try, you don't know if you like it or not.

Jobs came up too. The reporter asked about young Koreans who can't find jobs and keep stacking up degrees. Teddy said up front that he doesn't know the Korean job market well. But one thing he's sure of: a skilled developer can go after remote positions at US companies, and US companies still need people. And in any field, experience matters more than school.

In the early 2010s, companies that cared about pedigree and companies that didn't existed side by side, and Peter Thiel's Thiel Fellowship (a program that gives $100,000 over two years to people who drop out of or skip college) was seen as a bold move back then. Today, degrees and pedigree barely matter.

He was blunt about MBAs as well. The mood now, he said, is: you could have spent that time working at a company or starting a startup, so why did you do an MBA? He gave the University of Waterloo's co-op program as an example. You have to complete three or four internships to graduate. Because of that, students build real experience before they finish, and they go straight to Google, Amazon or Facebook right after graduation.

If you can show me something cool you built, that means more to me than anything.

In the end, he said, Silicon Valley wants attitude, not credentials: people who are curious and explore and figure things out on their own.

9. The most active, most honest AI community

As our conversation wrapped up, Teddy said that Patrick had built the Claude Bloom community, and that not many people can pull off something like that on their own. The reporter sitting with us nodded. I was grateful.

For context, Claude Bloom aims to be the most active and most honest AI community. We work with Anthropic/Claude, Microsoft, Oracle, MyRealTrip (a Korean travel platform) and others, and we've been growing fast, bringing in tech influencers like JoCoding and Builder Josh (Korean tech YouTubers), an xAI engineer, and speakers from abroad. Our Discord passed 1,600 members in five weeks, and in July we have community events coming up in Singapore and Japan. We're working toward a global AI community grounded in the humanities. We run a YouTube channel too.

Originally written in Korean on Brunch, May 30, 2026.