지금도 늦었다. 60만 명 해고가 말해주는 AI 시대의 잔혹한 현실 | 최재붕 교수 [세바시45 ep.94]

Hello, I'm PD Koo Beom-jun of Sebashi45. In just a year and a half, the world's capital has poured over 300 trillion won into AI. Today, we've invited Professor Choi Jae-boong of Sungkyunkwan University to most clearly analyze this era of upheaval and suggest the survival strategies we need. Hello Sebashi fans, this is Professor Choi Jae-boong. I'm back again to talk about AI. I'll do my best to explain how real change is happening in the AI ​​era. Choi Jae-boong's Global AI Trend. You published a new book, right? Yes, that's right. I wrote a book called AI Sapiens in June of last year. I usually write a book once every two years, but not long after I wrote it, another tremendous change began. As of June 27th of last year, I added up the capital accumulated by the top 10 AI companies, and it was 230 trillion won. I did it a few days ago, and it's 370 trillion won. Wow, as of June 27th of last year, the combined market capitalization of all companies in Korea, on the KOSPI and KOSDAQ, was 250 trillion won. Recently, it has surpassed 4 trillion. We're talking about hitting 4,200 today, but most of them are AI-related companies. So, yes, that's right. So, you could say the entire world is currently swept up in an AI craze. When this capital gathers, especially in the digital age, it's easy to quickly absorb talent. They invest without asking about economic concepts like sales or profits. So, investment is happening in everything related to AI. The expectation for South Korea, in particular, is physical AI. In fact, for physical AI to develop, you need to be a good LLM candidate, have good semiconductors, and have a strong manufacturing industry. The only countries in the world that have these three strong pillars are South Korea and China. But from the U.S.'s perspective, they can't partner with China. Jensen Huang came, made a deal, had chicken, and the media described the meeting that day as an 8,300 trillion won meeting. South Korea's science and technology R&D budget this year is 34 trillion won. If Jensen Huang sells his 1% stake, it will be 70 trillion won. But why did he choose Samsung Electronics? Did you bond over chicken and beer with Hyundai Motors ? What you're most interested in right now is humanoids. You think that, aside from China, there's no other partner that can achieve the best results in physical AI. So is the reason you think that way because of the data? Let's be clear. First of all, semiconductors are very important. Yes, it's clear that TSMC is the best partner. But from the perspective of the company that places the order, they keep raising the price by 50%. They said, " Hurry up and develop this. Release this." But they said, " Hey, wait two years. I have to give it to Apple, and there are too many places to give it to. I ca n't be so easygoing. This can be seen as the evil of a monopolistic company. So, they need to cultivate partners. It's a kind of check and balance. That's right. So Samsung Electronics signed a contract with Tesla, but from NVIDIA's perspective, Samsung Electronics also has memory that TSMC doesn't. They have memory that NVIDIA doesn't have. They can't do without memory. Next, automobile factories will need a lot of humanoids. Tesla is doing it, and they're doing figure robots, and Boston Dynamics is doing it too. Figure robots and Tesla will be a bit picky. We're investing a lot, but Boston Dynamics is Hyundai Motors, so it's a bit. It's a bit easy. Right now, NVIDIA is really focusing on Omniverse, a software for humanoid development, and Cosmos, a training platform. They're the perfect partner to work with. Tesla's market cap is over 200 trillion won, right? Even if they sell 1% of their shares, it's 20 trillion won. So they're building data centers and stuff, and they're building data centers, but Hyundai and Kia combined have barely surpassed 100 trillion won these days. So, their technology is good. Right? Their technology is not bad, and their performance is good. Last year, Hyundai and Kia sold over 8 million units. Tesla only sold 1.6 million units. So, if they join hands and NVIDIA leads, I think they can be good partners when they start manufacturing or mass producing humanoids. So, how do you see the future of our country's technology and economy ? I think we can see that they are relatively superior. We are always worried about China. That's right. Despite China's economic difficulties, China is investing enormously in the AI ​​field. They are also training a lot of human resources. So, the shock of DeepSec, Alibaba developing a new semiconductor for AI, that kind of shock. The country that is most wary and afraid of that is the United States. So, the US-China hegemony war can be said to be an AI war. From the US's perspective, they have different amounts of money in LLM and capital, and they have different amounts of software power, but their manufacturing industry is a mess, right ? Look at the shipyard. Also, cars are losing their competitiveness now, right? But when you look at the world, Europe has already gone crazy. Right? Japan and Taiwan are partners, but Taiwan is actually listed in the UN as a vassal state of China. It's a bit risky to form a partnership with them, so from their perspective, the partner they can most easily join hands with is South Korea. In fact, just look at this humanoid alone. In fact, just look at this humanoid alone. Why is it difficult to cooperate with China? If there's a war, it could become an enemy soldier. What about self-driving cars? That's right. Yes, there's a backdoor. Just thinking about things like that makes people suspicious, but from the US's perspective, it's not easy to import humanoids or self-driving cars developed in China. So, if we form an alliance with a country that we can trust, then Korea can be the best partner. It seems like they're using a slogan like "Sovereign AI" as a strategy because of this situation and environment. What's the content? If you could explain this in more detail, there are several concepts behind sovereign AI. For example, some people are wondering if we can independently create an LLM to catch up with the US or China. We can't do it. But if we just give up and don't develop and wait for them to make it, what happens later? We want to develop an AI doctor. There's a ton of medical data. South Korea is truly the best in the world. That's because there's no country that gets a health checkup every two years. That's right. On top of that, we have the highest rate of CT and MRI scans in the world. Oh, I see. That's right. So, making an AI doctor with this is really easy, and the accuracy is bound to increase. But if we don't do an LLM, people say things like, "We can use what the US has later. Then, will the US give it to us?" The moment the data is transferred, it's over. So, we need to closely watch what the US is doing, form partnerships, bring in LLMs, and build LLMs that can be used without having to give us our data. I think that's the core of the sovereign AI that I'm talking about. What do we need for that? What do we need? Then you need a data center. This time, we decided to give away 260,000 GPUs. The world's top one right now is Grok 4. Elon Musk's Grok 4, and then there's GPT-5, made by Sam Altman and Open AI. These two are the top. In the case of Grok 4, these two are the top. In the case of Grok 4, they gathered 200,000 GPUs and made XAI's data center. And the data is taken from what they made with X's Twitter platform, and they are continuously upgrading it. In the end, the performance of this LLM is the basis for making physical AI, doctor AI, war AI, physical AI, doctor AI, war AI, and they hit the top with 200,000 units. One unit costs 50 million won. But last year, when Blackwell introduced new products, they introduced them, but this year's units were all sold out, so don't ask. It's annoying. Jensen Huang is that popular. They are selling them for money, but if they give away 260,000 units, it's a huge deal. Among them, 50,000 units are now sold. Samsung Electronics is entering the market, and Samsung Electronics plans to train all semiconductor data related to manufacturing and sales with it. This is a vertical AI. In Korea, the government will establish a data center for sovereign AI with a capacity of 100,000 units. With that alone, multiple AI systems can be trained. The reason why developing an LLM is difficult is because the problem isn't writing the program. After writing it, you have to train it for performance evaluation. The GPU cost for training is 8 billion won per session. It costs 8 billion won, or even billions of won, to train it once. The government should do that. It will train our country's medical data and study materials created by teachers. Then, we can create private tutors, AI doctors, and various public services. And now, we can apply this vertically to autonomous vehicles and robot development. Then, researchers developing AI can conduct various research with the GPU costs supported by the government. This will play a significant role in enhancing Korea's AI capabilities. We can't afford to be unusable for too long. The fact that Jensen Huang said he would provide this level of information can be seen as a huge help in Korea's rapid advancement to the forefront. I'm curious about the direction in which AI technology is currently flowing. Could you explain it to me? Well, the LLM was the most popular at first. This is a versatile secretary who responds well to questions. So, if you ask, it organizes all the world's knowledge and gives you a report like a secretary. This was not possible for 70 years, but in 2017, Google published a paper called Transformer, causing a big bang. And exactly 35 months have passed since then. The reason it's possible is because generative AI mimics the human brain. That's right. So people competed to increase brain capacity and increase the amount of learning. So, how far has it gone? It's reached 1 trillion parameters. GPT-5 was announced on August 8th. Hey, it's almost 1 trillion brain capacity. GPT-4 had 600 billion. And this one can do almost everything. AGI, as we know it, might have been completed, or it might have become Artificial Super Intelligence (ASI), so it did n't work. It's almost been proven that this method doesn't work yet. We hit a wall. The research team said that, and so we turned to vertical AI. It seemed like there were a lot of things that could be done with just this much brain power. And now, what's starting to get the spotlight is physical AI in particular. People, I guess, can buy things and benefit from them, or robots that can replace my labor, and this can have a huge economic effect that can be felt directly. Next was LMM, right? So, it learned multimodal sentences and created a program that can draw well. These days, look at Google Nano Banana. This is amazing, to the point that it's almost like software that will ruin Photoshop. In the design, video, advertising, and marketing fields, the emergence of LMM has made it possible to say that disruptive innovation has become a reality. Also, look at physical AI. In the past, physical AI was known for self-driving cars and robots, so why is it suddenly called physical AI? It's an LLM. It does the coding itself. It learns from videos and codes itself to learn how to drive. Tesla trained its cars on 4 billion kilometers of driving videos for 10 years and developed that LLM in its system. The full self-driving FSD 12 was released in January 2024. Think about it. I was just reading a book on my way to work today. I was just reading a book on my way to work today. I've never parked a car before. I'm not experiencing it right now. I'm experiencing it. In China, they introduced 400 self-driving taxis in Wuhan. They recently released data. Taxi drivers' income decreased by 20%. When people started riding, they said, "This is okay. The fare is one-third." Then, they increased the number to 1,000. What's the most important data? The accident rate was one-fourteenth that of human drivers. In the US, Google's Waymo has finally received approval in New York. Oh, even in a complex city like New York, it's now technically recognized enough to receive approval in the most complex and difficult-to-drive city on Earth. Last year, Waymo alone Over 10 million people have ridden it. What role has LLM played? Until now, engineers had to work hard and tirelessly code based on sensors. LLM is replacing that. What is the concept behind Tesla's Optimus? Bring it to your house. It records everything in your house. Then, it starts recording. You just have to show the coffee machine and this in a slightly restrained manner. After recording all of that, it sends it to headquarters. Then, it learns and can understand the movements of human joints. Using that, it creates tens of thousands of similar cases. And through that, you do all the coding with LLM. Then, in the morning, you download it and I'll start making coffee now. Metaverse technology and LLM are now combined. NVIDIA has started providing that software for free. However, to do that, you need to learn it, which is also tied to NVIDIA's GPU. So, you have to use our GPU to use it for free. That's because that learning will have a special learning algorithm. So NVIDIA is the robot's. The school provides the training platform learning platform for free. That's right. But if you want to train here, you have to use our GPU-powered robots. That's Jensen Huang's dream. He's a visionary who dreams of such things. But if you look at it from another perspective, isn't that creating a complete monopoly ? Oh, that could be true. If you look closely at the case where NVIDIA monopolized the GPU market, it wasn't because they were good at making semiconductors. They realized that the hardware control software that sets up AI programs to run based on semiconductors is what makes money. So that's how it's going. What did they originally want to sell? NVIDIA is diligently developing the chipset that goes into robots. Tesla is designing it, and a company called Figure is also working on it. The key is to enable rapid learning and mimicking human tasks. That's why chipsets are used here, and massive data centers are also built on servers that require massive learning capacity. But if you use their software, you have to buy our hardware. Then, it becomes, " I'll do it because I have the money." Ultimately, various businesses will emerge, but it's like Windows in the early days of PCs. That's right. This is the dream of Jensen Huang, who monopolizes the operating system and sells products. Since there are model homes anyway, people go to those model homes and cook, clean, make coffee, and do dishes. They record all of this on video, and then digitally model it. That's the software called Omniverse. Then, whether it's Tesla's Optimus or Boston Dynamics' all-new Atlas, you can use any robot you want. But if you use Cosmos, which allows the robot to see and imitate what you're doing and learn to do it, you can configure all the hardware settings. So, buy a Tesla Optimus or Boston Dynamics' all-new Atlas and train it with NVIDIA's Cosmos. Then, when you go later, you just have to pay the Cosmos usage fee. You can think of this as the dream of NVIDIA, Elon Musk, and Tesla. In the case of the Optimus Three, they've already placed an order for 8,000 units to a Chinese robot actuator parts company for mass production. Elon Musk's dream is to open a mass production factory next year. So, the day when robots become as common as our daily cars and play as important a role as cars is not far off. Our engineers said that it wouldn't be possible even in 2023 or 2024. They said it would take more than 10 years. Mass production is impossible because there are no sales. Well, humanoid robots have to sell to generate sales, and those sales have to generate more investment, so a virtuous cycle will be established, right? But what happens when 3.5 trillion are gathered? Even without sales, they'll invest. A huge investment has occurred. A huge investment has occurred. Especially in the US-China hegemony competition, China spent about 100 trillion to build 450,000 parts factories. The US can use anything not related to communications, right? So, they're excited because cheap parts are flooding in. So, more than 100 humanoid robot companies were created, some in China and some in the US. But have they sold? They have n't been mass-produced yet. Not a single one has been sold. It would cost another 100 trillion to build a factory. It would cost about 500 trillion. There will be no sales until 2028. But don't worry. Don't worry. If you worry about this, everything will fail. That's what's been happening until now. Why hasn't humanoids developed? Hey, it costs 500 trillion. How can you believe in it even without sales? So, we didn't do it, but now we have 500 trillion. So, the possibility of it becoming a reality has increased significantly. In the past, we When you appeared on Sebashi 45, you said, "If you want to get rich, become a doctor. If you want to change the world, become a programmer." But you can't become a programmer in that short period of time, right? Oh, that's right. It's true that programmers started it. If you look at the actions of Mark Zuckerberg, who is already becoming a senior figure in the AI ​​industry, he is now scouting a friend who founded a startup called Dateke and is now an engineer. What did he say by spending a huge amount of money ? I told him that if he moved to my company within this week, I'd give him 160 billion won. What did he say in reply? It's okay, that's it, right? Then I'll give you 330 billion. That's it. This has spread on social media and become a huge issue. Last year, when I spoke at Sebasi 45 and wrote a book, I said, " We're giving away 10 billion won each. We're in big trouble." But it ended up being 330 billion won. Not only that, they just threw away 100 billion won in a ridiculous way. And I did an interview with a reporter. I asked, "Isn't it really worth spending 20 trillion won? What's the point of spending so much money to secure talent?" He said, "Even so, it's not even 1% of our company's shares. But if we don't secure talent now, there's no future." Winners take all. They're all former programmers. People who think a lot about what kind of programmers they are. Vibe Coding and various other things do the coding, but they create these programs and these services, and ultimately, they solve their business model through coding. AI will help a lot with this coding part in the future, right? But the ability to think, 'I'm going to lead innovation by writing programs and creating these services' from a young age is not easy for AI to replace. So, let's look at the examples of people who have achieved such success. If you look at it, he was very interested in business from a young age, and one of the cores of that business was software development. If you get help from AI, then a world is opening up where you can realize your business ideas faster without a lot of investment. The competition is incredibly fierce. The competition is fierce, but if it is fair competition, a market that Koreans can enter is opening up. I think there is hope for the AI ​​native generation if they study hard in the future, but what should people who are working now do? This is one of the areas that is really difficult now. Google, Microsoft, Amazon, these guys have laid off a ton of people. They say they will lay off 600,000 people by 2033. And the number of robots in Amazon's fulfillment centers has finally surpassed one million. Even with the humanoids that are a bit slow and clumsy now, it's clear that it will get even worse in the future. Besides, think about it. Aren't those companies the places that American college students want to work the most? That 's right, but most of the layoffs due to AI job duplication are done automatically by developers, advertising, and marketing, so did the number of employees decrease? They say it's almost the same. If you look at Microsoft, for example, what has increased, the number of AI-related departments has increased significantly. In the US, layoffs will already occur starting in 2023, so it's bound to become a social problem. No, imagine if Samsung Electronics, SK Hynix, Naver, and Kakao each laid off 10,000 people in Korea. There would be chaos, right? And if you ask, "Hey, you worked at a good company, why were you fired?" and they say, " Oh, my job overlaps with AI." Think about it. If this conversation goes back and forth, kids will be in chaos. That's right. That's why universities are in chaos. Has it already been? It's already a mess. In the case of Palantir, they came up with the theory that college is useless, had four-month interns, and then finally hired 22 people. They said they would do much better in the future. That also explained the future trend. But no, the tuition is 80,000 dollars a year, 120 million won. Then, including living expenses, they graduated from MIT and Harvard and have 800 million won in debt. It's hard to get out of this. Right? So, don't do that and come here. Then, I'll create something. I'll create a Palantir mafia. We often talk about the PayPal mafia. There were 300 people who did the four-month internship. Oh, so there must have been a network there too. Twenty-two people were hired. Oh, twenty-two people were selected from the interns. That's right. So, those twenty-two people are now being raised as the Palantir mafia. They're 18 or 19 years old. So, in about five years, when they're around 24, they'll have about 20 trillion won and go somewhere. They've already been raised by Peter Thiel. One of the Palantir mafia is Palmer Luckey. He's the one who first created the glasses. He's a genius, a young genius. But he sold his company to Facebook. They acquired the company on the condition that he work there, and he became incredibly wealthy. He was working hard at the company, but he showed some right-wing tendencies. He was kicked out because he was close to Trump, a right-wing figure in Silicon Valley. Peter Thiel visits him, and he's depressed because he was kicked out. Peter Thiel is a defense company that's very close to Trump. So, let's not do that and focus on defense. So, he founded Anduril. Palantir's technology in the metaverse space was used to create a suicide drone. And that's how the company achieved tremendous results in the Russia-Ukrainian War, becoming a 40 trillion won company. That's right, it increased the success rate of drone attacks in Ukraine. So now, it's linked to a system called Gotham, which Palantir created. Anduril is reaching out to Korea. There's no Joseon, what about the Americans? So, HD is making modern unmanned combat vehicles. Ah, now the surface ships, and then we're joining hands with Korean Air for unmanned fighter jets. This isn't a suicide drone, but a combat drone. What's the US worried about now? It would be nice to have the F-35. Even Europe is asking for it. Russia is going to invade. They say they can make 12 a year, but when will they make 200? Even if they make 200, it won't be enough. And the price is not right, and they can't manufacture them. So what should the US do? Places like Anduril can produce small drones, but since they can't manufacture them, they're saying, "Let's make them in Korea or in Korea together." What should we do about education in our country? I get asked this question a lot about education in our country, too. How do you catch homework that was done with AI? I think our country has a stronger tendency to grasp and regulate things. You could think that kids do n't need to do that kind of homework anymore, right? The first thing that's revealed now is that Dr. Chuck Chuck, an all-purpose secretary, can tell me any knowledge at any time. Then there's no reason for me to spend time memorizing and memorizing it. Then I have to create new ideas from it. For example, it could be a business model. So, what's most needed in many ways? The ability to explain my thoughts, the ability to structure my thoughts, things like that. Right? Right. So how should classes change? Express your thoughts in a video. Then, use AI as you wish. This will improve your AI skills and other capabilities. The second is the ability to present. This is very analog. Presenting and discussing with others to improve communication skills. These two are the most necessary capabilities in the AI ​​era. To be honest, according to this recent opinion poll, 90% of people are afraid of the changes brought about by the AI ​​era. The first is that I'll be pushed out of my job, right? Number 2 is changing so quickly, can I really just sit still? Is it okay if I do n't study? It's all fear, it's all fear, right? But what about the kids? If you were born in an AI world, then you just have to live like that, and then the generation gap will be huge. But what are the adults doing now? I want to protect the world I created until I die. But where should I invest? Let's do it in a new world. This is the duality I'm talking about. It's not too late. I said it's been 35 months, right? That's right. It's been 35 months since we saved money. It was the same when the internet first came out 30 years ago. People thought a lot about whether it would work or not. 10% believed it would work, so they threw themselves into it and worked hard. The remaining 90% thought it would n't work and were swept away. That's right, in a revolution. So now, it's unlikely that Representative Koo or I will be scouted for hundreds of billions of won. It's a shame. So what should we do? Let's all level up together. What South Korea is really good at is that if you have curiosity and study, you can do it. So if I try it myself and level up, my worldview will definitely broaden and change little by little. If you look at our positioning, it's not bad. People say it's rigid, but no, no other country has platforms like Naver and Kakao, manufacturing and service industries that have been created through hard study, right? So, there's a talent for studying together and moving up. So if we level up together like that, the children will also. You can gain respect, and based on that respect, you can jump again. Isn't that what we need to do most right now? So, I thought, "Let's study AI PT for 30 minutes a day. Just close your eyes tightly for 30 minutes a day for exactly 3 years. AI Nano Banana has come out. I'll write a manuscript book and make a video. If you do that, the possibilities and the speed of development are so fast, so I think if you do it for 30 minutes a day, you'll definitely become a different person in 3 years." So, Europe actually released a self-reflection report in response to the AI ​​era. They continued to regulate digital, right? But after 20 years, the GDP of the US and Europe was the same, but in 2002, there was a 30% difference. And now, there was something like that recently. An EU employee who criticized Trump lost his email. He said, " You hate the US, so I'll cut you off." Hey, our email does n't have sovereignty, so we don't have sovereignty. But when we looked back, we didn't have a data platform. There's nothing to learn. We imposed a lot of environmental regulations, so we don't have electricity. Hey, regulation isn't the only solution. That's right. But what's happening now is regulation. The experts remain, and all the innovators have left. This time, the two AI geniuses who won the Nobel Prize last year are both British PhDs. That's right. When I asked them why they left such a good country, the answer was that the fields they wanted to research were all subject to regulation in Europe. So, the people subject to regulation are mainly lawyers and legislators. These people have been making a living for twenty years and have created an ecosystem. So what are you saying? Whether AI, humanoids, or physical AI come out, just sit still and do nothing. I will create perfect regulations. So, what happens to all of that? Then, here, autonomous driving won't work here. So, going to the US, humanoids wo n't work here. So, going to the US, my point is, should there be no regulations? No. But, fear-driven regulations, and we know this well. If the mindset of the entire nation has leveled up, it would be good to maintain this level of awareness while pursuing innovation. So, there needs to be a balance. So, during a revolutionary period, regulations need to be slightly lowered and the challenge of innovation needs to be raised. This requires effort from individuals, organizations, and society. When we need to advance technology, we need to make a concerted effort to study. Yes, that's my opinion. It's been ten years since I first invited Professor Choi Jae-boong to Sebasi for a lecture. What he said then was that changes in the world can be understood through the movement of capital. No one has presented this so clearly and specifically to the public. Looking back on the past ten years, I regret not following the flow of capital that informed us of this. Even I, for one, did n't buy NVIDIA stock. Along with that, another thing is that we have a lot of pride in our cultural competitiveness. You shared some insights, and I see that this book also discusses fandom. What is this about fandom in the AI ​​era? I think the AI ​​era is a huge bubble right now, and it's going to burst someday. But when it does, which companies will survive? When I saw the dot-com bubble burst, I thought that companies with good technology had to be overwhelmingly good. But what was the standard for determining whether that technology was good or bad? In the digital age, consumers are the power. So what is the goal of AI? All employees have to level up with AI, and that goal is the customer's choice. I talk about subscriptions and likes. Please subscribe and like Sebasi. This is the same thing. So no matter how good AI becomes, in the end, that choice determines success or failure. Whether it's a humanoid, a self-driving car, or an unmanned car, it's all the same. Customers have to choose us. That's right. That's why I've been talking about this for 10 years, and it's starting to happen in industries with the strongest fandoms. A prime example is the power of K-pop, a cultural product that will never die. It's incredibly powerful, right? Look at Olive Young. The White House spokesperson came in during that brief gap and said, " I bought an Olive Young product," and then posted a selfie. And then, she posted a selfie. Olive Young's sales actually increased by 28% last year despite being offline. So, what the heck, when I looked at the data, I realized that 10 million of those purchases were made by foreigners. Last year, there were 18 million foreign tourists in Korea, and more than half of them, 10 million, came and bought things. They did n't even order anything. There are two founders of APR: Kim Byung-hoon and Lee Joo-kwang. So, Lee Joo-kwang, the founder, has exited, and I had the opportunity to interview him for about 400 billion won. Their worldviews are so different. They were born in 1988. Their worldviews are focused on social media marketing and Korean K-fandom. They think that if you make cosmetics in Korea, you should sell them globally. Why Korea? They think they can easily create decacorn companies worth 10 billion dollars. Friends with that worldview start creating products and taking on challenges. That creates a huge opportunity. Ultimately, AI can open up areas that people can explore with a small investment. If we expand it, then where will the competition be? If customers come here with a good experience and go crazy, and if we create a fandom, cosmetics and food can succeed, and many other things can happen. From that perspective, I talk about Korea's second chance as a fandom. I talk about it all the time. No, whether it's semiconductors, AI, or robots, people who are good at studying are the best, right? People who study well can make a good living anyway, but I'm good at making tteokbokki. Oh, I'm good at drawing webtoons. I'm not interested in studying, but I like that in life too much. In fact, what Sebashi pursues is that all those people can do well, and the ecosystem can grow so that society can be healthy together. I've loved music since I was young. Think about it. TOP is BTS, right? Then dreams can change. Opportunities to challenge new businesses based on the things that truly made me happy begin to open up. So, diversity in our society is greatly increasing. That's right. People see the growth of pianist Lim Yoon- chan as very unique. I think it's a product of the digital age. Usually, a genius pianist doesn't emerge unless they receive lessons from a genius pianist from a very young age. But this genius, Lim Yoon-chan, has never left Korea, right? But he's like, "Wow, he was really inspired by Rachmaninoff's performances on YouTube, and he was able to learn his genius. He had another teacher." So people overlook this, but they say it's classical music. There was a lot of talk about how disappointed Im Yoon-chan was in the music industry. What was that? Where did you come from? Hey, I'm from over there, Juilliard. Where are you? He recalls how this Korean society, still dominated by legacy, made his childhood before winning the competition terribly miserable. In every field, there's a legacy, and it's so powerful, so why is that? Which high school did you go to? What university did you go to and what did you do? In this AI era, they say it's a meritocracy, right? A world based on meritocracy that opposes academic background is opening up. I think the content market has definitely opened up, and what ultimately dominates it is subscriptions and likes. Hey, this person has the skills, so they don't need anything else. In the future, this phenomenon will happen in almost every field, and the basis of that skills will be subscriptions and likes. So what do you need? Buying people's hearts Another way to say it is like subscribing and liking is empathy. Empathy Empathy is the humanities. That's why I tell my students this. Your most important asset right now is the ability to make that person empathize. Whether you're an engineer or a writer, no matter what role you play or what job you have, the size of the empathy you elicit will ultimately determine the size of the success you pursue. So what should you do? Looking into the hearts of others, whether it be consideration, a deep exploration of that person, or introspection about humanity, ultimately it is necessary, and the ability to communicate with many people is what I'm talking about. As Steve Jobs said, " Marrying technology, humanities, and humanity made customers' hearts sing." Even in the age of AI, the goal is still the power to make customers' hearts sing. In that context, I think Korean humanities will become a very attractive base in the future. I think that if those of you who study humanities don't just say, " I'll do my own humanities," but systematically uncover where this passion for Korean culture, which is happening in the US and Europe, comes from, it will be of great help to our children as they design their future careers. Technologies are advancing, and huge companies are investing heavily, but as you mentioned earlier, they're not making money. They're pouring everything into it. So, where are they going to make money? They said that these kinds of questions started to come up. That's right. But listening to what you just said, it seems that someone who creates empathy and is ultimately chosen is the one who ends up making money. That's right. Yes, Professor Choi Jae-boong is with us today. He gave us valuable insight, or rather information, that can increase expectations for future growth, from the speed of the AI ​​revolution to the power of the K-fandom economy and securing Korea's AI sovereignty. Ultimately, the most certain way to not fall behind in this revolutionary era is to do AI PT for 30 minutes a day. It's very realistic. Why can't we? Of course, of course, with just a little determination, you can do it. Just like the professor said about AI PT for 30 minutes a day, small actions will change our future. Thank you so much, Professor Choi Jae-boong, for joining us today. Everyone, Professor Choi Jae-boong's new book, Choi Jae-boong's Global AI Trends, is a must-read. Please read it as a must-read. Thank you. Thank you for the good words today. Professor, yes, thank you. This was Sebashi 45. See you next time. Thank you. Come get your skin care done. While you're here, let's get a health checkup. Then, when you go to the digital healthcare AI doctor in the future, please fill in one thing for me. I will take responsibility for your healthcare and take care of it. In two years, when it's time for a health checkup, come again. BTS will have a comeback concert. If we decorate it like the Olympics, like this.