Reinventing Politics With AI

The third guest on the Bloom podcast was Choi Dongin of The Hackathon, the team behind "Black and White Developers," which went viral recently (the name riffs on "Culinary Class Wars", the Netflix cooking show known in Korean as "Black and White Chef"). Almost nobody around me can go deep on both technology and politics, so I found the whole conversation fascinating and a lot of fun. We jumped across society, culture, history and more, which I loved (haha). I'm even more curious how it lands for the people watching.

"I'm building what comes after democracy"

When I asked Dongin to introduce themself, the answer was one sentence.

I'm Choi Dongin, and I'm working on what comes next for humanity's operating system after democracy.

When I asked why they picked politics as their issue, they said it started with a thought: politics is such an important part of society, but is there the will and the drive to change it as the times change? From there they asked themself how they could change it, in what area, in which direction and with whom, and that one sentence came out of it.

Every book at home was a book on ruling

The interest goes back a long way. Their father studied law and their mother painted in the traditional East Asian style. The home itself had a strong traditional, East Asian feel, so they grew up memorizing the Thousand Character Classic and reading the Four Books and Five Classics of Confucianism. Along the way they noticed something.

I read the Four Books and Five Classics, and they're all books on how to rule.

They're about how to govern well, whether it's your body, your family or your country. For high school they went to Hanil High School in Gongju, South Chungcheong Province, an extremely controlled environment where they had no cell phone until they were twenty and had to use pay phones. They spent three years reading during self-study hours. As it happened, their homeroom teachers in the second and third years were big fans of Marx, so they read "Capital" and social philosophy during that time, and since they could read German, they read some of it in the original.

Perfect test scores, and a D

Germany came in between. Their father was posted there as a legal cooperation officer, so they lived in Munich and Berlin for a few years in elementary school. When they got there, kids their age were debating Nietzsche, Heidegger and Wittgenstein. The way we read classical Korean poetry, they were reading modern philosophy.

The shock came with the report card. They got every test question right and still got a D.

Getting the test questions right isn't enough. You have to contribute to this class.

The standard was that if you didn't contribute to class with questions or comments, a perfect score meant nothing. From then on they started thinking about what a good question is. A good question, in the end, is about finding the problem. They said the framework was noticing a problem somewhere, sharpening it through your own point of view, and going as far as proposing a way to solve it.

The other thing they saw in Germany was how dense political activity was. There are multiple parties that form coalitions with each other, and there are youth parliaments, so from a young age kids participate and say which party they like. When I brought up the glass dome on the Reichstag in Berlin, they said there's a reason things like that get built.

Where are the philosophers now

They wanted to get into Seoul National University's political science department and didn't make it. The interview topic was what kind of trend the sharing economy, like Uber and Airbnb, represents. They answered that it was a communist movement. The old picture was the working class rising up in revolution to bring down the capitalists. Now cars and homes, which used to be consumer goods, were being turned into capital, opening a path for workers to become capitalists too. Wasn't the trend shifting in that direction? The professor didn't like it and asked whether they had actually read the books. Dongin laughed and said they probably weren't trying to give the answer the professors wanted either. They admitted to a bit of a contrarian streak.

So they ended up in business school, and it hit them hard that they were in the wrong place. They wanted to do philosophy and were interested in politics, and they had no interest in this field. Then, talking with a friend from a philosophy reading group they'd been in for a long time, they landed on something.

Nobody has produced a real philosopher in modern times.

Digging into why, they concluded that people like Elon Musk and Steve Jobs are the philosophers of this era, and that's why they started founding companies. Over two or three years they tried all sorts of things. They built and sold a health supplement brand, and they picked up a pile of ideas because they wanted to enlighten the world. Then importing game IP from China and Japan and selling merchandise went better than expected. About a year in, the question came back.

Why did I start this? Can I do this for ten years?

At the end of 2023 they handed off that business. They wanted to try something completely new, and they felt they finally had the room to do what they really wanted.

What they saw helping their father's campaign

Around then their father ran for the National Assembly and asked for help, and that's when they saw raw politics for the first time. Politics stopped being an abstract idea. How the game actually works, how voters and politicians communicate, whether politicians are bad people: all of it hit them at once. They campaigned in a Spider-Man suit, calling themself "your friendly neighborhood Spider-Man."

The conclusion they drew fit in one line: the ways of communicating are really bad. Mass texts, banners, broadcasts and debates are all one-way, so you have no choice but to say the same thing to everyone. You have to say the same thing to a teenage boy and to a grandmother in her nineties, so it resonates with no one. Talk to the grandmother about policy for teenagers and it doesn't land. Talk to the teenager about policy for grandmothers and it's the same.

The only message that works on everyone is "the other side is bad"

So I asked what's left. The answer was short.

If you ask what resonates with everyone, it's just that the other side are bad people.

All you can say is that the other person has moral problems, because that works on teenagers and on people in their nineties. Setting up a clear enemy and talking as if removing that enemy will solve every social problem is easier to take responsibility for and easier to execute. "Support me and I'll bring that person down" also makes the payoff crystal clear.

The problem goes one layer deeper. They started with the hypothesis that changing the media and the means of communication would make politics better. After about a year of trial and error, they found the underlying structure was tangled up much more tightly than that. The media and the speakers weren't separate, and they were playing this game on purpose. It took time to realize that.

You can rack up hundreds of millions of views and in the end nothing changes. It actually just gets worse.

Because the message stayed the same. "That person is no good, I'm better, help me and I'll bring them down." No matter how much you changed the format, that skeleton didn't change.

"I like underdogs"

The politech startup they founded at the end of 2024 was called Chamchi Sangsa. "Chamchi" means tuna in Korean, and plenty of people remember it as a company that sells tuna or mistake it for one. But here chamchi stands for "chamdoen jeongchi" (true politics) and "chamyeo jeongchi" (participatory politics). They even added a tip: when they meet a left-wing politician they say participatory politics, and when they meet a right-wing politician they say true politics.

Right after the martial law declaration, they wanted to see for themselves what kinds of people were showing up, so they went to both left-wing and right-wing rallies and sold T-shirts and hats. They said they saw there that there are structures capable of mobilizing hundreds of thousands of people, with give-and-take relationships inside them. The conversation went on to the idea that politics comes down to groups fighting other groups over their interests, a contest over whose side has more people and who persuaded more. And that the office politics people talk about inside companies has the same structure.

During the presidential race, they helped with publicity for the campaigns of Lee Jun-seok, Kim Moon-soo and Kwon Young-guk. They made things like a campus cafeteria tour of universities, "Junstock," a one-to-many Q&A format, and "Minsaeng Four Cuts," a twist on the Life Four Cuts photo booths popular in Korea ("minsaeng" means people's livelihoods). They said they borrowed a lot from the playbook of celebrity fandoms. I asked whether they felt any disconnect helping candidates whose views didn't match theirs. The answer was that party isn't their criterion.

Watch on YouTube

Because I think what always changes the world is the underdogs rising up, and the world changes as they gain power.

It could be a Democratic Party candidate running in Daegu (a conservative stronghold), or it could be someone doing the hard work in a third party from the start. Regardless of party, they thought about how someone in an underdog position could get better known. They said they especially like people who haven't just repeated what others say but have lived those words through their own lives. When I asked who their favorite underdog is, they said they'd listened to a speech by former president Roh Moo-hyun just the day before. The reason was that he proved himself through a track record of ten, twenty years.

If AI could be Plato's philosopher king

When I asked where they want to apply technology, the answer was that they want to make people's thinking a bit more complex. Instead of "this person is good so I'll vote for them, that person is bad so I won't," they want people to weigh what a policy means for them given their gender, age, region, job and family. It's impossible for one person to cross-check all of that, and that's where technology comes in.

Plato came up here. After watching Socrates get put to death by majority vote, Plato proposed rule by philosopher kings as the system to follow democracy. If you read how he describes the philosopher king, Dongin said, he's closer to a slave than to an elite.

He owns nothing, has no wife, and has broken free of his own interests.

And at the same time, he understands all the dynamics of the social structure and the system. No such person ever existed, so Plato's idea stayed an ideal, and Plato himself went around small states trying to educate kings and failed every time. What Dongin thought at that point was: couldn't AI do that?

I asked whether AI isn't also owned by someone in the end. If one company's model ends up running the whole National Assembly, wouldn't that risk concentrating power in a few hands? Their answer was to compare it with humans. Humans don't even explain why they made a decision, and it's hard for them to. A model, on the other hand, can be controlled, trained and improved, and its data and algorithms can be made public.

I mentioned that I like watching National Assembly hearing videos. I watch them because it's entertaining when lawmakers yell at each other for talking down to them. My question was whether that fun would survive if AIs argued with each other. We agreed that the fun comes from identifying with the people on screen, and you can only identify with your own species. So for a while, they said, what we'll see is politicians augmented by AI going up against politicians who aren't.

They laid out a practical picture too. They think the number of lawmakers should actually go up while staff gets cut, and they imagine AI taking some of the 300 seats, fact-checking and keeping people from crossing the line. Right now, though, politics uses AI for little more than chatbot search, and parties are even more conservative than traditional companies, so AX (AI transformation) isn't easy. So the conclusion came out like this.

Building an AI-native party is much better than doing AX on an existing party.

We also touched on why third parties struggle. They end up centered on a person, which makes it hard to build a vision, and when something goes wrong with that person, the whole thing collapses. The Democratic Party and the People Power Party keep their vision and direction even when their names change, and only the people get swapped out. So Dongin said three things are needed: a vision, a community of people who share it, and media to spread it outward. Black and White Developers was their attempt to build that media themselves.

The real reason behind Black and White Developers

So I asked what the one vision for that party would be. AI doing politics is a means, so there has to be a world you want to build with it. The answer was a country with more people living happily and absorbed in what they do, and they think Korea's situation here is especially serious. Right now the economy keeps running and work-life balance is more or less holding, so people are getting by, but once AI and robots start replacing labor, that balance will break at some point. Education needs to change starting now, yet it keeps repeating the same thing.

First get into a good college, first get good grades, first get into a big company. There's nothing but "first."

They called it a system that mass-produces people with no taste, no thoughts and nothing to get absorbed in. In the industrialization era it was the engine that let Korea grow fast, but now they think it's completely wrong.

That's where the intent behind Black and White Developers came into focus. Part of it was testing media reach and gaining a community and network. The bigger goal was to let people know that we're now in an era where you can build products with AI and make real money, and that you can find your own thing and try it. They estimate only about a million people, roughly 2% of the population, know this, and they want to push that number to 10 million and then 50 million.

That's also why their target audience is younger than Bloom's. They want teenagers to try things freely without risk and find out what they like. That doesn't mean everyone has to start a company. Whether you become a doctor or a lawyer, take an office job or work in a factory, before you make that choice you should at least explore what you like before going out into the world, and the ten years from middle school through college are a period with a safety net. Their reason for recommending everyone try starting a company once was the same. What idea you want to pursue, how you approach it and what strategy you use reveal everything about your personality. It's a compressed record of the path of your life.

To wrap up, I brought up someone I know. They work at xAI, and all their colleagues are high school students who've already published several papers. There's even a community on X where they find people to co-author papers with. The idea that degrees don't matter is old news there, and the sense of ranking people by age collapsed a long time ago. We closed the conversation on the thought that if Korea had more young people like that, our society and politics might move up a level too.

Originally written in Korean on Brunch, September 22, 2026.