Hello, this is PD Gu Beom-jun from Sebasi 45. Everyone, in just a year and a half, the world's capital has flowed into AI by a staggering 3 quadrillion won. Today, we have invited Professor Choi Jae-boong from Sungkyunkwan University to provide the clearest analysis of this era of upheaval and present the survival strategies we need. Hello, Sebasi fans. I am Professor Choi Jae-boong. I am here to talk about AI again. I will do my best to explain clearly how real changes are unfolding in the AI ​​era. You recently published a new book titled " Choi Jae-boong's Global AI Trends," didn't you? Yes, that is correct. I wrote a book called "AI Sapiens" last June. I usually write a book about once every two years, but shortly after I finished it, massive changes began again. As of June 27th of last year, I added up the capital gathered in the top 10 AI companies, and it was 23 quadrillion won. I checked again a few days ago, and it is 37 quadrillion won. Wow. As of June 27th of last year, the total market capitalization of all Korean companies on the KOSPI and KOSDAQ was 2,500 trillion won. Now, recently... It has surpassed 4 quadrillion won. People are talking about how we hit 42 trillion won today, and most of them are AI-related companies. Exactly. That is correct. It is to the point where we can say the entire world is currently swept up in an AI frenzy. When this capital pools together, especially in the digital age, it easily and quickly attracts talent. People invest without asking about economic concepts, such as revenue or profit. That is why investment is currently taking place in everything related to AI. The expectations placed on South Korea, in particular, are for physical AI. In fact, for physical AI to develop, one needs to be proficient in LLMs, have a strong semiconductor industry, and have a solid manufacturing sector. South Korea and China are the only countries in the world with all three of these strengths. However, from the US perspective, they cannot partner with China, right? Jensen Huang came, made a pact, ate chicken, and the media described that meeting as an "8300 trillion won meeting." South Korea's science and technology R&D budget for this year is 34 trillion won, so if Jensen Huang sells a 1% stake, it amounts to 70 trillion won. But why, of all places, did he choose to partner with our Samsung Electronics and Hyundai Motor Company? Did you form a partnership with Kkanbu while eating chicken and beer? My biggest interest right now is humanoids. You believe that, aside from China, this is the only partner capable of achieving the best results in physical AI. So, is the reason you view it that way because of the data? What is clear is that semiconductors are extremely important. Yes, it is obvious that TSMC is the best and most excellent partner. However, from the perspective of the ordering company, they were raising prices by 50% and demanding, " Develop this quickly, release it." But the company could say, " Hey, wait two years. I have to supply Apple, and there are too many other places to supply, so I ca n't just give in easily." This can be seen as the harmful effects of a monopolistic company. Then, you need to cultivate partners. It is a form of checks and balances. That is correct. That is why Samsung Electronics signed a contract with Tesla, but from NVIDIA's perspective, Samsung Electronics possesses memory that TSMC does not have. They possess memory that TSMC lacks. Since they cannot operate without memory, next is automobiles. Moving forward, car factories will need to incorporate a lot of humanoid robots. Tesla, Figu, and Boston Dynamics are all working on this, but Figu and Tesla are likely to be quite picky. We have many investments, but Boston Dynamics is affiliated with Hyundai Motor, so they are relatively more lenient. NVIDIA is currently focusing heavily on Omniverse, the software for humanoid development, and Cosmos, a training platform. And they are the perfect, flexible partner to collaborate with on these projects. I mean, Tesla's market capitalization has exceeded 2 quadrillion won; even if you sell just 1% of their shares, that amounts to 20 trillion won. That's why they are building data centers and such. On the other hand, Hyundai Motor Group, including both Hyundai and Kia, has barely surpassed 100 trillion won recently. However, their technological capabilities are excellent, right? Their technology is n't bad, and their financial performance is also strong. Last year, Hyundai and Kia sold over 8 million vehicles, whereas Tesla only managed to sell 1.6 million. So, if they join hands and NVIDIA leads the way, I believe it is entirely reasonable to think that they could be excellent partners when entering the humanoid manufacturing or mass production system. Then, how can we view the future of Korea's technology and economy ? I believe that, relatively speaking, we can be seen as being significantly superior. We are always worried about China, and that is correct. Despite their economic difficulties, China is investing enormously in the AI ​​sector. They are also training a massive workforce. Therefore, the place that is most wary of and fearful of a "deep-seeking shock" —or the shock of Alibaba developing a new AI semiconductor—is none other than the United States. So, the hegemonic war between the U.S. and China can be described as an AI war. From the U.S. perspective, they possess vast amounts of money and software power on a different scale, including LLMs and capital, yet their manufacturing sector is a mess, isn't it ? Look at the shipyards. Also, the automotive industry is losing competitiveness right now, is n't it? However, looking at the global landscape, Europe is already in a bad state, isn't it? Japan and Taiwan are the only remaining partners, but from their perspective, Taiwan is actually listed as a vassal state of China in the UN. The risks of forming a partnership with them are quite high, so from their standpoint, the most approachable partner might actually be South Korea. In fact, just looking at this humanoid robot alone, the reason cooperation with China is difficult is that if a war breaks out, it could become an enemy soldier. What about autonomous vehicles? That’s true. Yes, given that suspicion arises merely from things like the existence of backdoors, the U.S. cannot easily import humanoids or autonomous vehicles developed in China. Therefore, they form an alliance with a trustworthy nation, and Korea can be the best partner. It seems they are pursuing a strategy under the slogan "Sovereign AI" due to these circumstances and environments. What exactly are the details? If you could explain this in more detail, Sovereign AI involves various concepts. For example, some people ask if we can catch up to the US or China by independently developing an LLM. We can't do it. But if we just give up and sit back, waiting for them to make it for us, what happens later is this: we want to develop an AI doctor. We have a massive amount of medical data, and South Korea is truly the best in the world when it comes to that. Why? Because there isn't another country where people get a health checkup every two years, right? That's correct. On top of that, we have the highest rate of CT and MRI scans in the world. Ah, I see. That makes sense. So, using this, creating an AI doctor becomes a really easy task, and accuracy is bound to increase. However, if we don't pursue an LLM, people often say things like, "We can just use what the U.S. has done later." But would the U.S. really give it to us then? The moment the data is transferred, it is over. Therefore, we need to closely observe what the U.S. is doing, establish cooperation to adopt their LLM, and build a system capable of performing tasks even without providing our own data. I believe that is the core of what I call Sovereign AI. To achieve this, what is needed? What would be required? That is when data centers are needed. Regarding data centers, they agreed to supply 260,000 GPUs this time. Currently, the world's top is Grok 4. Elon Musk's Grok 4, and the next one is GPT-5, created by Sam Altman and OpenAI. These two are the top, and in the case of Grok 4, they assembled 200,000 GPUs to build XAI's data center. And regarding the data, they took it from what XAI built after acquiring the Twitter platform and are continuously upgrading it. They are constantly upgrading it. Ultimately, the performance of this LLM forms the foundation for creating physical AI, doctor AI, and war AI, and they have reached the top with just 200,000 units. Each unit costs 50 million won. However, when introducing the new Blackwell product last year, they said, "We are introducing it, but all for this year have already sold out, so don't ask us. It's a hassle." That is how popular Jensen Huang is. Even though they are selling it for money, giving 260,000 units makes it a very generous deal. Of those, 50,000 are going to Samsung Electronics. Samsung Electronics intends to use that to train data across the entire semiconductor field, covering manufacturing, sales, and related sectors. This essentially leads to the emergence of vertical AI. In Korea, the government is currently planning to build a data center with a capacity of 100,000 machines for sovereign AI; with just that, we can train multiple models. The reason developing LLM is difficult is n't the programming itself, but the fact that performance can only be evaluated after training. The cost of using GPUs for training is simply 8 billion won per session. Since training requires billions of won, the government should cover this. If we train using domestic medical data and materials created by teachers, we can create private tutors, AI doctors, and various public services. Then, by applying this vertically to autonomous vehicles and robot development, researchers can conduct various studies using government-supported GPU costs. This will play a significant role in boosting Korea's AI capabilities. We shouldn't be unable to use it for too long, but we need to reach that point. Just the fact that Jensen Huang promised to provide it can be seen as a significant help in allowing our country to quickly surge ahead as the leader. I am curious about the direction in which AI technology is currently evolving; could you please explain that? Well, initially, the one that garnered the most attention was LLM. It is a jack-of-all-trades, all-around assistant—a versatile assistant that answers questions well. So, if you ask something, it organizes the world's knowledge and provides a report just like a secretary. This was something that hadn't been possible for 70 years, but the Big Bang happened in 2017 when Google published a paper called *Transformers*, and exactly 35 months have passed since then. The reason it is scary is that Generative AI mimics the human brain. That is correct. So, people engaged in a competition to increase brain capacity and train it to a large amount. So, where has it gone now? It has reached 1 trillion parameters. Yes, the parameters. GPT-5 was announced on August 8th. Wow, they have increased the brain capacity to almost 1 trillion. Since GPT-4 had 600 billion, this one knows almost everything. You might think AGI was complete, or perhaps it could have become Artificial Super Intelligence (ASI). But it did n't work. It can be considered that it has been largely verified that this method doesn't work yet. The research teams said they hit a wall, so they shifted direction toward Vertical AI. They thought, "Wow, with just this level of intelligence, there seem to be many things you can do." And that is when Physical AI, in particular, began to gain prominence. After all, people benefit from purchasing goods, or robots emerge to replace their labor— these provide tremendous, tangible economic effects. Next was LMM, right? So, they created programs by learning multimodal sentences, and they could even draw well. Look at it these days; Google's Nano Banana is so massive that people say it is software that will put Photoshop to waste. In fields like design, video, advertising, and marketing, the emergence of LMM has made disruptive innovation a reality. Look at Physical AI as well. In the past, we used to call them autonomous vehicles and robots, so why the sudden shift to Physical AI? LLM is involved. It does the coding itself. It learns from videos and codes how to drive on its own. Tesla trained its system with 4 billion kilometers of driving footage over 10 years and developed that LLM within its system. Full Self-Driving (FSD) 12 was released in January 2024. Think about it. You commute to work and say, "I just came here reading a book today." You have never once driven from parking to parking. But you are experiencing it right now, aren't you? You are experiencing it. China introduced 400 driverless taxis in Wuhan, right? They recently released data, and taxi drivers' incomes reportedly dropped by 20%. Once people started riding them, they realized it was pretty good; the fares are one-third of what they are now. So, they reportedly increased the fleet to 1,000 vehicles. What is the most important data? The accident rate came out to be one-fourteenth that of human drivers. In the US as well, Google Waymo has finally received approval in New York. It means the technology has been recognized to the extent that it received approval in a complex city like New York—one of the most complex and difficult cities on Earth to drive in. The number of people who rode Waymo last year alone The number has already exceeded 10 million. If you ask what role LLM has played— previously, engineers worked hard and tirelessly coding based on sensors— LLM is now replacing that. What is the concept behind Tesla's Optimus? You bring this robot into your home, and it records your entire house. Then, you start recording, saying, "This is how I brew coffee at home." You just need to show the coffee machine and the equipment with a bit of precision. After recording everything, you send it to headquarters. Then, it learns. Since it can understand the movements of human joints, it generates thousands or tens of thousands of similar cases. And through that, all the coding is done using LLM. Then, in the morning, you download it and say, " I will brew coffee now." Metaverse technology and LLM are now combining. NVIDIA has started providing that software for free. However, to use it, training is required. This also becomes tied to NVIDIA GPUs. So, you have to use our GPUs to use this for free, right? Because that training must have specialized learning algorithms. So, NVIDIA is the school for robots— that training. Providing the platform training platform for free is correct. However, if you want to train here, you have to be a robot using our GPUs. That is correct. That is Jensen Huang's dream. He is someone who really dreams of such things; he is an amazing visionary. But if you look at it from another perspective, isn't that just an attempt to create a complete monopoly ? Oh, that makes sense. They looked at the case where NVIDIA monopolized the GPU market and realized it wasn't because they made good semiconductors. They understood that the hardware control software that sets up AI programs to run based on those semiconductors was what was profitable. So, that's how this is going as well. What exactly do you think they want to sell? NVIDIA is working hard on manufacturing the chipsets that go into robots. Tesla is designing them, and a company called Figu is also involved. This is because the core idea is to enable them to learn quickly and mimic human tasks. Chipsets are used here, and massive data centers are deployed in servers that require enormous training capacity. However, the thinking is: "If you use our software, you must buy our hardware. So, since I have the money, I should give it a try." Ultimately, while various operators will emerge, it plays a role much like Windows did in the early days of PCs. That is correct. The dream Jensen Huang has— monopolizing operating systems while selling products—is this: Since model houses exist anyway, he films people going to those model houses to cook, clean, brew coffee, and wash dishes, and then digitally models that footage. That is the software called Omnibus. Then, you can use any robot, whether it is Tesla's Optimus or Boston Dynamics' All-New Atlas. However, for the robot to observe the footage and learn to mimic the actions exactly, if you use Cosmos, we will handle all the hardware setup for you, so Tesla You can buy Optimus or Boston Dynamics' All-New Atlas and train them using NVIDIA's Cosmos. Then, when you move on, you will only have to pay the usage fee for Cosmos. You can consider this the dream that NVIDIA and Elon Musk's Tesla are pursuing. In the case of Optimus III, they have reportedly already placed an order for 18,000 units with 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 is not far off when robots will become as numerous as cars in our daily lives and play a role just as important as automobiles. Our engineers said back in 2023 or 2024 that it wouldn't work; they said it would take more than 10 years. We couldn't mass produce it because there would be no revenue. Humanoid robots need to sell to generate revenue, and that revenue needs to come in for investment to occur again, thereby establishing a continuous virtuous cycle, right? But what happens when 35 quadrillion won is gathered? It means there is no need for revenue; massive investment has taken place. Especially in the US-China hegemony rivalry, China spent about 100 trillion won to build 450,000 parts factories. Since the US can use parts that aren't related to telecommunications, they are thrilled that cheap parts are flooding in. As a result, over 100 humanoid robot companies have been created—there are companies in China and the US. So, have they sold? Mass production has n't started yet; not a single one has been sold. Building a factory costs another 100 trillion won; it will likely cost around 500 trillion won. There will be no revenue until 2028. But they aren't worried. If they were worried about this, everything would go bankrupt. That is exactly what has been repeated until now. Why hasn't humanoid robots advanced? Well, it costs 500 trillion won; how can you trust and proceed without revenue? That's why they did n't do it. But now that 500 trillion won has appeared, the possibility of it becoming a reality has increased significantly. In our last SEBASI 45... When you appeared, you said that if you want to be rich, become a doctor, and if you want to change the world, become a programmer. But you can't become a programmer in such a short time, right? Oh, right. It is true that a programmer started it. If you look at the moves of Mark Zuckerberg, who is already becoming a veteran of the AI ​​industry, he has now scouted a friend named DateKera, who founded a startup called Vercept, as an engineer. He spent a huge amount of money on it. What did I say? I told him, "If you transfer to our company this week, I'll give you 160 billion won," and what did he reply? "No thanks." Is that it ? So, 330 billion won—yes, I'll go with it. That is the gist of it. As this spread on social media, it caused quite a stir. When I was talking about it on *Sebasi 45* last year while writing my book, I said, "They're offering 10 billion won each. We're in big trouble. We're in big trouble." But it turned into 330 billion won. Not only that, they just casually threw around 100 billion won. Then I did an interview. A reporter asked, "Is it really necessary to spend 20 trillion won? What on earth is the reason for spending so much money to secure talent?" And I replied, "Even so, it doesn't even amount to 1 percent of our company's equity." However, if we can't secure talent now, there is no future. Winners take all. They are all programmers. Among programmers, what kind of people are they? People who think deeply. While Vibe Coding and various other tools handle the coding, the idea of creating these programs and services to ultimately realize them—that is, resolving one's business model through coding. AI will certainly help a lot with the coding part in the future, right? But the capability to think from a young age, "I will write these programs, create these services, and lead innovation like this"—that is not easy for AI to replace. So, if you look at the examples of people who have achieved such success, from a young age... I had a strong interest in business, and one of the core elements of that business was software development. With the help of AI, a world has begun to open up where one can realize their business ideas more quickly without significant investment. The competition is incredibly fierce. However, if the competition is fair, a market is opening up that Koreans can realistically compete in. I believe there is hope for the AI ​​native generation if they study hard, but what should those currently working do now? This is one of the toughest sectors right now. Companies like Google, Microsoft, and Amazon have laid off a massive number of employees; they even announced plans to lay off 600,000 by 2033. Furthermore, the number of robots deployed in Amazon's fulfillment centers has finally surpassed one million. Even with the current humanoid robots being somewhat sluggish and slow, the numbers are already at that level, so it is clear that things will only get worse in the future. Moreover, just think about it: aren't those the companies that American college students want to work for the most? That's right. But regarding the layoffs due to AI job overlap, most of those are related to developers, advertising, and marketing— all of which are generated automatically. So, have the number of employees decreased? They say it's practically the same. When I looked at what has increased at Microsoft, their AI-related departments have expanded dramatically. Since layoffs have already been happening in the U.S. since 2023, it is bound to become a social issue. I mean, imagine if Samsung Electronics, SK Hynix, Naver, or Kakao in Korea laid off 10,000 people each. It would be absolute chaos, wouldn't it? And imagine someone asking, "Hey, you went to a good company, why did you get fired?" and the answer being, "They said my job overlapped with AI." Just think about it. If that conversation went back and forth, the students would go crazy. That's right. That's why universities are in an uproar. Already? Already ? It’s already a huge sensation. In the case of Palantir, they put forward the argument that college is useless, conducted a four-month internship, and eventually hired twenty-two people this time. They said those kids would do much better in the future. That also explains the future trend. I mean, tuition is $80,000 a year, which is 120 million won. If you include living expenses, that’s 200 million won. You graduated from MIT or Harvard, but you have 800 million won in debt. It’s hard to get out of that, right? Yes. So, don't do that. Come here. I’ll make something for you. I’ll create a Palantir Mafia for you. We talk a lot about the PayPal Mafia, you know. I think they did the four-month internship with 300 people. Ah, so that must have become a network too. Twenty-two of them were hired. Ah, so twenty-two were selected from the interns. That’s right. Then those twenty-two kids are being groomed into the Palantir Mafia. Since they are 18 or 19, in about five years, when they turn 24, they’ll be going somewhere for 20 trillion won and stuff—they are already the ones Peter Thiel has raised. One of the Palantir Mafia members is Palmer Luckey. You know those VR glasses? Palmer Luckey is the very first person to create them. He is a genius, a genius, and very young. That is why he sold his company to Facebook. Facebook acquired the company on the condition that he would work there, and he became incredibly wealthy. While he was working hard at the company, he showed a slightly right-wing tendency. In Silicon Valley in the US, being friendly with the right-wing Trump is unacceptable, so he was kicked out. Peter Thiel sought out this man, who was depressed after being kicked out. Peter Thiel is originally from a defense company—which is very close to Trump right now, isn't it? That is correct. So Thiel told him, "Don't do that, let's focus on national defense," and the company he founded was Anduril. Based on the technology he possesses in the metaverse, he created self-destructing drones. And as these achieved tremendous success in the Russia-Ukraine war, this company grew into a 40 trillion won enterprise. Right? That is what increased the success rate of drone attacks against Ukraine. So, it operates in conjunction with the system called Gotham, which Palantir created. Anduril has also reached out to our country. Where is Joseon? Where is the American guy? So, they are building unmanned combat vessels with HD Hyundai. Ah, surface vessels, and then partnering with Korean Air for unmanned fighter jets. These aren't suicide drones, but combat aircraft. The current dilemma for the U.S. is that they want F-35s; Europe is demanding them too. They fear a Russian invasion. They produce only 12 a year; people ask when that will happen. Even if 200 were produced, it wouldn't be enough. Furthermore, the cost is prohibitive, and they cannot manufacture them. So, the question is, what should the U.S. do? Small drones can be produced in abundance in places like Anduril, but since they cannot manufacture those either, the idea is to partner with Korea to build them. How should our country's education be structured? People in our country's education system also ask me that a lot. "How do you detect homework that was done by AI ?" It seems our country has a stronger tendency to crack down on and regulate things. Kids might think, "I no longer need to do that kind of homework," right? The first thing that has emerged is that a jack-of-all-trades, all- purpose assistant can teach me any knowledge I want, anytime. Then there is no reason for me to waste time memorizing it. Instead, I have to generate new ideas from that. For example, that could be a business model. So, what is most needed in various ways? The power to explain my thoughts, the power to build the structure of my thinking— these are what are needed, right? Right. So, how should classes change? If you tell students to express their thoughts on video and use AI freely, their AI skills will improve, and their capabilities will grow. The second is the power of presentation. This is very analog. Improving communication skills by giving presentations and debating with others—these two are being said to be the most necessary competencies in the AI ​​era. Honestly, as the recent poll showed, 90 percent of people are afraid of the changes brought by the AI ​​era. The number one concern is, " Won't I be pushed out of my job?" Number 2 is changing this fast, so is it really okay for me to just stay still? Do I not have to study? This is all fear. It's all fear, isn't it? But what about the children? If you're born into an AI world and think you can just live that way, the generational gap will become enormous. But how are adults acting right now? They say, "I want to protect the world I created until I die." But where do they invest? Let's do this in the world that will be newly created. This is the duality I speak of regarding the duality I hold in my heart. It is not too late. I mentioned earlier that it has been 35 months, didn't I? That is correct; it has also been 35 months since the money was saved. It was exactly the same when the internet first came out 30 years ago. People had so many thoughts about whether it would work or not. 10% believed it would work, threw themselves into it, and worked hard, while the remaining 90% thought it wouldn't work and were swept away. That is correct regarding revolutions. So now, I don't think there is a chance that either Representative Gu or I will be scouted for 100 billion won. It is unfortunate. Then what should we do? Let's all level up together. What Korea is truly good at is that if we study with curiosity, we excel at everything. So, if you try things out yourself and level up, your worldview definitely expands and changes little by little. In terms of positioning, we are not bad. Although people say we are rigid, other countries don't have the manufacturing, service, and platform industries like Naver and Kakao that we have built through hard study over the years, right? So, we have a talent for studying together as a whole to rise. Therefore, if we level up together like that, we can earn respect from the younger generation as well. It exists, and based on that respect, we can take another leap. I think that is what we need to do most right now. So, let's study AI presentation for 30 minutes a day. Just close your eyes tightly for exactly three years. The AI ​​Nano Banana has been released. You try writing a manuscript or a book, or making a video. As you do that, the possibilities and the speed of development are so fast that I believe if you do it for 30 minutes a day, you will definitely become a truly different person in three years. So, Europe actually released a self-reflection report in response to the AI ​​era. They kept regulating digital, but after 20 years, the GDP of the US and Europe were the same, yet in 2002, there was a 30% difference. And you know that story from recently about an EU employee who criticized Trump getting their email sent to them? It was basically, "If you hate the US, we'll cut you off." So, we realized we didn't have sovereignty over email, we didn't have the sovereignty to do so. But looking back, we didn't have a data platform; there was nothing to learn. We imposed harsh environmental regulations, so we don't even have electricity. Hey, regulation isn't the only solution. That's right. But what has happened now is that regulatory experts remain, and innovators... Everyone has left. These two AI geniuses who won the Nobel Prize last year both have PhDs from the UK, right? That's right. When asked why they left such a good place like the UK, they said they left because the fields they wanted to research were subject to regulation in Europe. And those subject to regulation are mainly lawyers and legislators, aren't they? These people have spent twenty years making a living and building the ecosystem. So what does that mean? Whether AI, humanoids, or physical AI emerge, it all became a situation where they were told, "Just wait, I will create perfect regulations for you." So what happens to all of that? So, you might think, "Oh, autonomous driving won't work here," and go to the U.S. Or you might think, "Humanoids won't work here either," and go to the U.S. My point is that we shouldn't have no regulations. However, there are regulations created out of fear. We need to be well-informed, and the mindset of the entire public needs to level up so that we can pursue innovation while adhering to certain standards. Therefore, there needs to be a counterweight; we need to balance things out. During a revolutionary period, we need to slightly lower regulations and significantly raise the bar for challenges regarding innovation. This requires effort within organizations, in the mindset of individuals, and across the nation and society. When we need technological advancement, we need a concerted effort to study. Yes, that is my view. It has been ten years since I first heard Professor Choi Jae-bung give a lecture at SEBASI. What he has been saying since then is that we can understand changes in the world through the movement of capital. No one had presented this so clearly and concretely to the public before. Looking back on the past ten years, I felt regret. Why didn't I follow that flow of capital that taught me that? Well, even I am like that—I did n't buy NVIDIA stock. Along with that, another thing is that he takes great pride in cultural competitiveness and offers a great deal of insight. As you mentioned, this book also discusses fandom. What is the story about regarding fandom in the age of AI? I understand that the AI ​​era is currently riddled with a massive bubble, and that it is bound to burst someday. But when it does, which companies will survive? I looked at the bursting of the dot-com bubble, and companies with superior technology—as long as it is overwhelmingly good—are the ones that survive. However, when I examined the criteria for judging whether that technology was good or bad, it was whether or not it receives customer choice. In the digital age, the consumer holds all the power. So, what is the ultimate goal of AI? All employees need to level up using AI, but that goal is customer choice. I refer to that as " subscriptions and likes." Please subscribe to and like Sebasi. It is the same message. Therefore, no matter how much AI improves, that choice ultimately determines success or failure. Whether it is humanoids, autonomous vehicles, or driverless cars, it ultimately comes down to the same thing: customers must choose us. That is correct. I have been consistently discussing this for the past 10 years, and it begins to emerge in industries with the strongest fandoms. A prime example is the power of K-pop—something that never dies. Cultural products are incredibly strong, aren't they? Look at what happened with Olive Young. The White House spokesperson came in during that brief window, said, " I bought Olive Young," and uploaded a selfie. And uploaded a selfie. In fact, Olive Young's sales increased by 28% last year, despite being an offline store. So I looked at the data and wondered, "What on earth is this?" It turned out that 10 million transactions were made by foreigners. Last year, there were 18 million foreign tourists visiting South Korea, and more than half of them—10 million—came and bought things without being asked. There are two founders of APR: Kim Byung-hoon and Lee Ju-kwang. Founder Lee Ju-kwang has since exited for about 400 billion won, and I had the opportunity to interview him. Their worldviews are very different. He was born in 1988. His worldview is fixated on social media marketing and Korea's K-fandom. He has the mindset that if you make cosmetics in Korea, you should naturally sell them globally. He believes that we can create decacorns worth trillions of dollars—companies worth $1 billion—anytimes we want. When people with that kind of worldview start building those things and taking on challenges, it becomes a tremendous opportunity. Ultimately, AI expands the areas where people can invest small amounts of money to try things out. So, where is the turning point? If you create a positive experience that makes customers think, "Wow, this is insane," and build a fandom, then cosmetics, food, and many other things can succeed. From that perspective, I speak of fandom as Korea's second chance. I say this all the time. Whether it's semiconductors, AI, or robots, the people who are good at studying always come first, right? Anyway, people who are good at studying can make a good living. But I have a talent for making tteokbokki. Oh, I'm good at drawing webtoons. I'm not interested in studying; I just love that in life. Actually, what Sebasi pursues is that people like that can all succeed, and we can grow the ecosystem so that society moves forward together in a healthy way. I've loved music since I was young. Think about it. T.O.P is BTS, right? Then dreams can change. It means that opportunities to challenge yourself in new businesses based on what truly made you happy are starting to open up. So, the diversity in our society is increasing significantly. That's correct. People view the growth of pianist Lim Yun -chan as quite unique. I think he is a product of this digital age. It is generally accepted that a genius pianist does not emerge unless they receive tutelage from a genius pianist from a very young age. But this genius, Lim Yun-chan, has never left Korea. Yet, he says this: "Wow, listening to Rachmaninoff's performance on YouTube, I could really gain inspiration and learn that genius." It means he had one more teacher. So, people overlook this, but he says that in the classical music world, it is very... Im Yun-chan spoke a lot about his disappointment. He recalled how the Korean society, still dominated by legacy, made his childhood leading up to winning the competition absolutely terrible. He talked about how, in every field, there was a legacy that dominated so powerfully that you know, things like, " Which high school did you graduate from?" People talk about which university you went to and what you did, but as we move into this AI era, we talk about meritocracy, right? A world of meritocracy, opposing academic elitism, is opening up. I believe that has already clearly taken hold in the content market, and what dominates it is ultimately subscriptions and likes. It's like saying, " Hey, this person has the skills, so none of that matters." This phenomenon will occur in almost every field going forward, and the foundation and standard of those skills will be subscriptions and likes. Then, what will be needed? Winning people's hearts—another way of expressing subscriptions and likes—is empathy. Empathy is a humanities discipline. That is why I tell my students this: "Your most important asset right now is the ability to make others empathize with you. Whether you are an engineer or a writer, no matter what role you play or profession you pursue, the extent to which you elicit that empathy will ultimately determine the magnitude of the success you seek." So, what should you do? People often say that looking into the hearts of others—whether it be consideration, deep exploration of a person, or reflection on humanity—is ultimately necessary, and that this requires the capacity to communicate with many people. Steve Jobs once said that he " married the humanities and humanity to technology," and only then did the customer's heart begin to sing. Even in the AI ​​era, the ultimate goal is that power to make the customer's heart sing. In that context, I believe Korean humanities will become a very attractive foundation in the future. Instead of those in the humanities simply saying, " I will do my humanities however I want," if you could systematically shed light on where this passion for Korean culture emerging in the U.S. and Europe actually comes from, I think it would be of immense help to our children as they design their future careers based on that foundation. Technologies are advancing, and massive companies are investing heavily, but as you mentioned earlier, they aren't making money—right? They are pouring everything into it. So, where are they going to make money? I heard that questions like these have started to emerge. That is correct. But listening to you now, it seems that by creating empathy, one is ultimately chosen—and that person is the one who makes money. That is right. Yes, today we have invited Professor Choi Jae-bung, and he has shared valuable insights, or perhaps information, that can raise expectations for future growth, covering everything from the speed of the AI ​​revolution to the power of the K-fandom economy and securing Korea's AI sovereignty. Ultimately, the surest way to avoid falling behind in this era of revolution is a 30-minute daily AI presentation. It is very realistic, and why can't we do it? Of course, of course. It is something you can do with just a little determination. Just as the professor said, " 30 Minutes a Day AI PT," small actions will change our future. Thank you so much, Professor Choi Jae-bung, for joining us today. Everyone, Professor Choi Jae-bung's new book, *Choi Jae-bung's Global AI Trends*, is a must-read. Please read it as a must-read. Thank you. Thank you for your insightful words today, Professor. Yes, thank you. Well, that was Sebasi 45. See you next time, everyone. Thank you. Come get a facial treatment. While you're here, let's get a health checkup. Then, regarding the future of digital healthcare, the AI ​​doctor will fill this up for you when you leave. I will take full responsibility for your healthcare like this. Come again when it's time for a health checkup in two years. We are holding a BTS comeback concert. If we were to decorate everything like the Olympics...