Did you know that rumors are circulating that OpenAI is going to start making its own mobile phones? Meanwhile, DeepSeek, a Chinese company that emerged claiming to compete with OpenAI , recently launched its new model V4, emphasizing not only high performance and low cost, but also abandoning Nvidia chips in favor of Huawei's own. Is the US-China tech war entering a new phase? Will this become a nightmare for Jensen Huang? Today, we'll introduce our guest, a master at translating complex technology into language we can all understand. We'll introduce you to Qu Bo , CEO of Knowledge Power Technology. Welcome to the show ! Hello, Xiaolu, and hello to all viewers. We'll start by discussing this new trend in AI. But first, we'd like Qu Bo to estimate how many out of 1,000 people in Taiwan are already using AI chatbots, whether it's OpenAI's ChatGPT or Gemini. How many people are actually using AI agents? (AI agents) assist in daily life. I believe that at least 60%, 70% , or even 80% of people use them . This is because even a Google search will bring up an AI summary, which is essentially artificial intelligence. You can chat with it. If we include this part, I think it should be over 80 %. But if we only talk about chatting with chatbots, that is, excluding search engines, I think it should still be at least 60% . So, how many out of 1,000 people do you think will start building their own AI agents, whether using Claude or other tools ? The current percentage is relatively low. I myself haven't done this. The reason is simple : artificial intelligence still has a serious problem : it can talk nonsense in a seemingly serious way. Let's go back and explain what "open" means. Claw , or artificial intelligence agent, is essentially a question-and-answer process where the AI ​​controls the entire interaction. However, the concept of an AI agent goes a step further. You tell the computer the problem you want to solve, and the AI ​​will devise a solution , plan the entire process , and even execute it automatically. Of course, because the process is complex, there are many risks. If the AI ​​can spout off nonsense, entrusting it with this entire process can be risky in practice . For example , imagine this computer controlling all the data on your computer , even automatically cleaning your hard drive. Wouldn't you be worried? Your computer might malfunction at any moment and delete data it shouldn't. This is possible, so I wonder if people can confidently use these AI agents before chatbots, or rather, current AI language models, can solve the problem of seemingly nonsensical statements. Another issue is that the installation process for current AI agents is still somewhat complicated . Unlike older apps where you just click "OK" to install, they're not as complex. For example, the recent trend of "raising lobsters" in mainland China, due to its complicated installation process , has created a new profession : helping people raise and install lobsters. This is because the process is somewhat complex and not intuitive. However, I personally believe the problem is actually simpler because technology is constantly advancing . I think that within six months to a year, someone will launch a service that eliminates the need for special installation, and it will truly be OK. OK, that's it. Installing this kind of lobster AI agent is possible, but I think resolving this issue of "serious nonsense" won't be easy . It might take two or three years, and it will continue to improve and evolve. I do have some friends who are hesitant to get into AI agents because the installation process is too cumbersome. I've also heard the Google CEO say that all software must be completely rebuilt in this new AI era because the logic is different now. However, Qu Bo also mentioned that the AI ​​we use daily , whether it's Chat-GPT or various AI tools like Gemini Grok... Wait, he's still making progress. Qu Bo, have you seen the recent hairstyle analyses and color suggestions we've had before and after filming? This is a new feature launched by ChatGPT. Previously, Chinese characters would display gibberish in its image generation and text processing. Besides English, it's also upgrading various languages ​​like Korean. But speaking of updates, Qu Bo just told me about the company that developed ChatGPT, called OpenAI. They're moving from software to hardware. What they're doing is... well, recently there's been news that they want to make phones. Let me clarify a bit: they don't necessarily need to make phones. Any phone can theoretically install the ChatGPT application. So why do they want to make their own phones? From their perspective , I think it's because, as everyone knows, while we can install ChatGPT, not everyone will. Also, this application is at the top layer of our software architecture, so its control over the phone's hardware is relatively low. More importantly , the computing power of our current phones isn't that strong compared to cloud servers and terminal devices. Therefore, the AI ​​accelerators within them may not be able to effectively support OpenAI. He optimized the algorithm , so from his perspective, he probably considered a few things. First, he hoped to create a native OpenAI (ChatGPT ) smartphone . What does "native" mean? The idea is that as soon as you turn on your phone, ChatGPT pops up as your smart assistant , handling everything for you. My favorite example is JARVIS from the Iron Man movie. Anyone who's seen the movie knows what I'm talking about. The protagonist, Tony , always asks JARVIS for help, constantly telling it what to do. Similarly , if you buy a phone and turn it on, JARVIS will pop up saying, "Hello , how can I help you?" You can tell it, " I want to go abroad to Japan for three days, please help me plan my trip ." It will automatically search all the information, arrange a great itinerary, and even discuss things with you. It might even chat with you if you're feeling down. What's wrong? Are you in a bad mood? Your tone isn't quite right . I think this is what it means for AI to be natively integrated into a phone. This is what they're thinking. Qu Bo, I hear you ? From hardware to software, this could bring about a completely different situation. Because this native ChatGPT or chatbot integrated into a phone— Google phones are already doing this . Google phones easily use the voice or OpenAI features in Gemini. Is he going to create a completely different software and hardware? Personally, I think it's because this news has only just started circulating and the details are still unclear. But from his perspective, it's obvious. Look at everyone's phones; they all use Android. So, wouldn't it be easy for Google to natively put its Gemini into our phones ? You're essentially forced into it. You don't even need to install an app ; you just buy the phone and turn it on, and it's there. The problem is with OpenAI. No, OpenAI is an app , so you have to install its app on your Android phone to use it. So, he wants to bridge that gap, hoping to have a native OpenAI or ChatGPT smartphone. Of course, you've hit the nail on the head: is he going to choose an operating system ? It's possible. It's uncertain at the moment , but I think given OpenAI's current influence, it's very likely. He has his own operating system, which I think will likely be Linux-based, possibly similar to Android, or even directly using Android. Because Android is open source, he could use it. But the key point is our software architecture. There's the so-called phone chip, which is the hardware. On top of that is the firmware, which is the software that drives the hardware. On top of that is the operating system, and on top of that are the applications . So, for OpenAI... To put it simply, he's only at the top layer; his control over the underlying layers is insufficient. Therefore, he might want to control the operating system and even the related firmware. He could control all the functions of the phone, right? I even think it's possible he'd want chips from manufacturers like Qualcomm and MediaTek to be optimized and accelerated based on his OpenAI algorithms. That would make the phone very unique, since OpenAI has billions of users. So, it's possible he'd ask Qualcomm or MediaTek to modify the chips to meet the needs of major clients. So, we 're just exploring his future possibilities. Regardless, OpenAI seems ambitious about entering the mobile phone market. Now, let's talk about DeepSeek, a Chinese company that has long been considered a strong competitor to OpenAI in the context of the US-China tech war. Recently, it's been reported that their newly launched model uses Huawei's Ascend chip instead of NVIDIA's. There's concern that Huang Renxun 's position in the chip market, providing chips to various service providers, might be shaken. Could this be another turning point in the US-China conflict ? Let's review the technical background. Artificial intelligence has two stages : training and inference. This is the first key point . In the training stage, artificial intelligence needs to input data into the computer to find rules and build a model. So there are three steps. The so-called model is the rule of the data. After you've built the model , that is, the rule of the data, you enter the second stage, inference . Inference is using this rule to input the question into the model, and the model calculates the probability of a correct answer based on the rules. This is called inference. For example, the chatbot we just talked about is a large language model. Everyone has heard of large language models. In short, it's the rule of human speech. In other words, it's the rule of speech . Scientists input human speech into computers to find rules and build models. These are the rules of human speech. Next, I will use this model to create a chatbot. When you have a question, you can ask it. After inputting your question, it calculates possible permutations and combinations based on the rules of human speech and tells you the answer. In the past, Nvidia's strength was in graphics processing units (GPUs). The characteristic of GPUs is that they can be used for training and inference. They are basically high-end general-purpose GPUs. They are indeed masters. However, there is a problem: you have to tell them what to do. So, you have to write a program to tell them what to do. Nvidia chips require this. This is the characteristic of Nvidia chips. The program itself will consume computing resources. So, in the past, we professionally said that this kind of general-purpose processor had a high degree of software programmability. But conversely, it also had some disadvantages, such as its chip size. Transistors are more power-intensive, and computation requires more computing power. They are indeed necessary during the training phase because training algorithms are constantly being modified , making software programmability crucial. However, the entire AI industry has entered a new phase starting this year. My favorite analogy is that the period from 2015 to 2025 was the previous phase, a market primarily focused on training. From 2025 (2026) to the next 10 years (2035), the market will gradually move into the inference phase. This means the model is almost fully trained, and manufacturers are starting to focus on monetization. Monetization is about inference, not training. Because you can only charge for using my model for inference, but training requires investment and significant investment. Now, we're entering the second phase. Can inference computation be done using a GPU ? Theoretically, yes , but the problem is inefficiency. The reason is that inference is usually a repetitive computation because the model is already defined . Repetitive computations, when written as a program and executed by a GPU, suffer from poor performance. So, what's the solution? So I'll switch to what's called an artificial intelligence accelerator , or what people often call an ASIC ( Application-Specific Integrated Circuit ). Some people also call this a neural processor (NPU). Whatever you call it, it's an artificial intelligence accelerator. Using an accelerator for inference calculations can reduce the computing power required for software control. Moreover, using an accelerator directly uses transistors for computation. Its chip area is small, its performance is high, it's energy-saving, and it's cheap. So the whole market is starting to change. This is the strength of Huawei's Ascend chip. Okay, next we'll talk about Huawei. From the past to the present , Chinese manufacturers have wanted to develop graphics processors. They are indeed making them. However, because graphics processors have a high barrier to entry , they currently cannot achieve the high-end graphics processors of Nvidia. Mainland China has always lagged behind in this area. So how do we make up for this lag? The solution is simple : use an artificial intelligence accelerator. So the Ascend chip initially had a central processing unit and a graphics processor , but the focus wasn't on the graphics processor because its computing power wasn't strong. The focus was on the artificial intelligence accelerator. So you just mentioned DeepSeek. This model, in fact, its training and inference, especially inference operations, using Ascend's chips , especially accelerators, actually perform well. This does put some pressure on Nvidia , so you'll find that Nvidia has also begun to adjust its strategy . At this year's GTC conference (a global technology conference held annually ) , Nvidia announced its latest progress: Nvidia has invested in a company called Groq (a chip startup), and Groq has launched a so-called Language Processing Unit (LPU), which is an accelerator for artificial intelligence . So this time, Nvidia is also turning around and emphasizing that its servers are not only graphics processors but also language processors , specifically designed for inference operations and understanding in artificial intelligence. So I think the whole trend is very clear , as I just mentioned, from 2025 to 2026. The year marks a turning point. The focus shifted from training-oriented to inference-oriented , and from GPU-based systems to accelerators becoming increasingly important. Qu Bo mentioned the accelerator reserves or productization needs from 2025 to 2035. What ASIC (Application-Specific Integrated Circuit) accelerators are needed in the market ? Taiwanese manufacturers are relatively capable of providing such products. Frankly, AI accelerators are currently mainly designed by European and American companies. Besides Groq (designing LPUs) and Google (designing TPUs), other companies like Amazon, Meta , and Microsoft are also designing their own accelerators. The reason is simple: think about these cloud service providers like Google and Amazon… Meta and Microsoft both have data centers, and they all need a lot of processors. If you buy all these processors from Nvidia, wouldn't Nvidia make a fortune? Wouldn't you cloud service providers be overpaying? Such expensive stuff, and you're buying so much ! So, if you're a cloud service provider owner, you'll start thinking about reducing costs. How do you reduce costs? Develop your own accelerators. Why? Because if it's my own money, I'll earn it myself . I just need to find a wafer foundry to manufacture it for me. And here's an important point to remember: the barrier to entry for graphics processing units (GPUs) is relatively high, while the barrier to entry for accelerators is relatively low. The reason is simple: GPUs need to be able to execute programs in any situation , so relatively speaking, the complexity of their underlying chips is higher, and therefore they are more expensive. That's why it's almost impossible for a company to design a new GPU to compete with Nvidia. But what's the shortcut? Developing accelerators involves a bit of a "cornering overtake," or rather, it depends on how you apply them. The inference and computation are perfectly suited to accelerators and aligned with the current generation. So, as I just mentioned, every cloud service provider can develop accelerators . But what supply chain does Taiwan have? After developing an accelerator, do you need to manufacture it into printed circuit boards ? Servers ? Cooling ? Chassis, racks? And more importantly, do you need a chip foundry ? So, who is it ? TSMC. Therefore, I think TSMC's recent strong stock price is reasonable. TSMC's stock price has been very strong these past few days. What I'm curious about is where Intel's progress is now. In terms of process technology , TSMC is currently ahead of Intel, with a slight lag. However, I think Intel 's biggest lag is in yield. Because if your chip yield isn't high enough, customers won't dare entrust such important projects to you. In terms of process technology, the difference between the two isn't that significant. The key point now is that regardless of whether you use accelerators or graphics processors , what you really need is advanced manufacturing processes . Therefore, there's no inherent difference in the quality of the manufacturing process. Customers naturally prefer the most advanced , so TSMC is still the priority. So why has Intel's stock price been so strong recently? There are several reasons. First, Intel's manufacturing processes are gradually improving. Secondly, frankly speaking, TSMC, with its high yield and numerous customers, doesn't have much capacity left , so its prices are very firm and it's even considering raising them. So, when Musk wanted to develop his own TeraFab project, who did he first approach ? TSMC? No, he approached Intel first. The reason is simple: it's not that he didn't want to approach TSMC; he certainly considered it. But could TSMC help him? That's uncertain right now because TSMC is too busy. Conversely, Intel urgently needed an important customer, so they quickly reached an agreement. Therefore, Intel first connected with TeraFab. You might wonder why he wanted to use a slightly less advanced, slightly outdated process? The reason is simple: if you can't get the top supplier, you have to go for the second-best. This is reasonable , and no customer wants to be tied to a single supplier; they always want a second option. The second source is a different choice , so I think from this perspective, Intel has an opportunity. At least Intel took over the customers that TSMC couldn't handle , and Samsung was also right . Another point I want to make about Intel is that when the entire market shifted from training to inference computing, the role of the central processing unit (CPU) became more important. Why is that? Currently, the entire AI server setup consists of a CPU paired with a GPU or accelerator. Remember, the GPU or accelerator is only needed for AI algorithms , but when users input their needs, I still need to do program scheduling management, which is what the CPU does. So think about it this way: when the market is mainly focused on inference computing , aren't there hundreds of millions of people worldwide who need to connect to the server , make requests , chat , and have you help them with computing ? Doesn't the CPU have to be responsible for scheduling and allocating resources to the GPU? In this way, the importance of the CPU increases, which is Intel's strength . That's why recently, Intel CPUs have started to be in short supply , and the entire market has begun to re-evaluate Intel. Based on Qu Bo's analysis, I know that Intel's new CEO, Chen Liwu, will be visiting Taiwan for Computex in Taipei this June . Many people, including those in the market, are eagerly anticipating his appearance and what positive news he might bring. Intel is very interested in Intel, which is different from the previous phase. However, from the beginning of our program until now, we 've focused more on the US. Now, I want to ask Qu Bo: looking back at China , with its own Huawei Ascend chips and DeepSeek, I believe they also want to make their own hardware phones. What I'm curious about is whether, within the AI ​​system, specifically AI phones, China will become a self-contained ecosystem in the entire AI industry chain , no longer needing to rely on the US and Taiwan. This is definitely the direction he's striving for , and it's very clear. He just said it's not possible to achieve this yet . I think the most important thing is the chip supply chain. I know that mainland China is working hard on domestic production and a complete vertical integration, but at this stage, even SMIC's wafer foundry requires a lot of high-end equipment. Yes, we have to import. So frankly, if all the equipment in this area had to be domestically produced, I think it would be a long road, maybe 5 or even 10 years. Some equipment might not even be produced domestically, like extreme ultraviolet (EUV) luminescence equipment. A mainland Chinese netizen bet me that we'd be mass-producing by 2026. It's already 2026, and he told me two years ago that we'd be mass-producing within two years. But that's definitely impossible. It's not that simple. It's possible to work towards that goal slowly. Another area is advanced packaging equipment. That has a high barrier to entry, but the possibility of domestic production is relatively high. Okay, that's semiconductors . After having semiconductor chips, then there 's system integration, printed circuit boards, and a whole series of systems. I think none of these are difficult for mainland China, as it's the source. As for software, that's the US's strength , but mainland China's software is also very strong. So, I think mainland China can indeed produce their entire system. It's just that the US knows this, and it might not be easy to compete. So they simply cut off our supply chain by withholding some advanced process-related equipment. If you can't buy them, you can't make the most advanced chips, which is why they need accelerators to make up for it. Of course, China is also working very hard, but I think there will still be some gap. The US is probably a few years ahead , and they can only use this advantage to continue creating new things to maintain their lead. Speaking of China's strong software , I recently saw a news report that the Gaode Maps app has recently become the number one downloaded app in Taiwan . People think that its traffic light countdown is particularly accurate, so they want to download Gaode Maps and care about the difference of one or two seconds. But another side says that while convenience and accuracy are one thing , there are also cybersecurity risks . I usually use Google Maps more. The logic behind Gaode Maps is that it collects a lot of data on Taiwan . I believe it must have some big data. Any user who downloads Gaode Maps will receive some information. For example , if you are driving on the road , because the map will definitely turn on GPS , and if a lot of people in Taiwan have installed Gaode Maps, then it will know many cars. Now, you can use this method to calculate the waiting time at each traffic light. I've seen some information; our government has issued a statement saying that all government APIs (Application Programming Interfaces) are not open to Gaode Maps for reading. In other words, the so-called traffic light durations displayed are calculated , not provided by the government. You might ask, what security issues might arise from this calculation? I think it's okay, because these are mostly applications related to daily life and consumption, so there shouldn't be too much of a problem. The question is whether this map can replace Google Maps ? I think it depends on each user's habits . If you really care about traffic light durations, then Google... Maps doesn't have such a thing. Maybe you want to switch to Gaode Maps , but anyway, this example shows that mainland China is actually very advanced in the software industry. Don't underestimate them. The government hasn't opened the API to Gaode Maps. Have people downloading the Gaode Maps app exposed any cybersecurity risks by giving data to Chinese companies ? As I just mentioned, your GPS location information will definitely leak out. But recently, some people have been calling for regulation of this kind of thing. Personally, I think there's no need for this kind of free market competition . This is free market competition. If you want to regulate this, then Huawei or Vivo phones— I see many of my friends using them— and Xiaomi phones too... Personally , I think this is a free market , and our general consumption behavior, whether it's going out for fun or taking transportation, doesn't seem to involve any great secrets that require such strict regulation. I think it's fine; only highly sensitive or special individuals need to be careful. Perhaps for important people, knowing exactly what time they are at which intersection and how many seconds they are waiting at the traffic light is crucial, right? For ordinary people, I don't think it's necessary. The program has been comparing various AI products from China and the US. We just talked about DeepSeek and the chip it uses, or how some people in Taiwan are starting to download Gaode Maps. On the other hand, I observed that the latest episode of The Economist specifically introduced Mythos, a product under Anthropic , positioning it like a digital nuclear weapon. In my opinion, this AI tool can detect vulnerabilities in the system and mobilize resources to directly patch them. Why is it described as a powerful weapon ? We've been saying that artificial intelligence is a data rule model, a set of rules for data. Every program we write actually has certain rules. Today, I'm using Mythos... This kind of model, when used to review your written code , will find many vulnerabilities , or so-called bugs . In engineering, we used to find these bugs manually. Frankly speaking, no matter how good the human eye is, its speed is limited. But when a program runs, its speed is very fast, so isn't that great? Therefore, any program, once reviewed, will likely find a lot of vulnerabilities. That's why everyone is very afraid. Why? Because if hackers use such tools to study the code of their targets, it's very dangerous. They might find vulnerabilities , or even use a certain method to attack the website and find its vulnerabilities, using certain rules to do so. That's possible. So later, they dared not release this model because if it were released, anyone could get it and anyone could attack it, which is too dangerous . But we need to be prepared that Anthrop won't be the only company. The development of such a model inevitably leads to other similar companies creating similar models, so sooner or later, such models will proliferate. More sophisticated hackers will still be able to replicate them. Therefore, what needs to be done now is for every company to review its own programs, using artificial intelligence to identify and patch vulnerabilities. The power of the Mythos model isn't just about finding vulnerabilities; it can also write a program to exploit those vulnerabilities within hours. This is very dangerous. If it can be done in such a short time, many industries, especially those related to finance and money , as well as government and military industries, will be affected . This is the main reason why the entire market is experiencing such a shock. This product can be described as Ragnarok. Qu Bo, I'm curious about your personal stance on AI now... When a company or software company develops such a powerful product, should the government play a role in regulating it? I personally think it's necessary. However, this is a double-edged sword. The more the government regulates, the more restrictions there are on the development of the industry. The less the government regulates, the easier it is for the industry to develop, but it's also easier for problems to arise. So I think there's a balance to be struck. From the government's perspective, it should establish basic regulations to prevent these companies from overstepping the boundaries. However, it shouldn't regulate too many details ; it should allow the industry to develop and thrive . For example , could we allow a robot to pull the trigger and kill someone without human consent? I think the government should regulate this . For instance, after the government discusses and decides... If you can, then you should establish this regulation. It can't just be a basic regulation . No manufacturer can develop any weapon , firearm, or robot that can be pulled directly without human decision-making. The key point is that it cannot act without human input. This is the principle I just mentioned. As for the details, the government should allow industry to develop freely. However, I think the reaction or understanding of governments around the world is actually very slow. Let alone focusing on Taiwan , even in the United States, it took US senators and representatives a very, very long time to understand how Facebook operates and how it makes money before finally deciding what Facebook cannot do. In fact, the industry has always been faster than the government. This is normal because technological advancements are always pioneered by a group of brilliant minds. Regulations are typically formulated by those with legal or political expertise. Once scientific breakthroughs are achieved and technological advancements are realized, regulations or government intervention is necessary to address the issue. Therefore, there will naturally be a time lag, which is normal . However, what needs to be done must be done. When we discover that a technology might be spiraling out of control, appropriate regulations may be necessary. That concludes our discussion. Finally, I want to return to a very fundamental issue concerning people's livelihoods : whether everyone can keep their jobs. Recent news reports, if you've been following this, have you heard from Me... Oracle, Amazon, and other large companies are all laying off employees , and the layoff rates are quite high. You'll notice that these large companies are pouring all their money into the AI ​​arms race, but junior engineers , or those with only one or two years of experience, are all being laid off. I'm curious how this wave of layoffs will unfold, Qu Bo. Earlier this year, a research institution published an article that focused on how wonderful the future of artificial intelligence is , but they saw a completely opposite picture. One aspect is that many software companies no longer need to write their own programs , so they don't need so many engineers. Even many software companies that previously... Previously, clients would pay to buy software to use these applications. However, because artificial intelligence can write programs, any company can now write its own software using just a few engineers. So why would you pay software companies to buy their software ? That's why you'll find that software-related stocks, especially in the US, have experienced a significant drop over the past three months . This is just the beginning. If you read that article, you'll find it chilling. Why? Think about it: these unemployed engineers —won't their spending decrease? Once their spending decreases, all related industries will be immediately affected. They won't eat, they won't consume entertainment, and even their homes will be in trouble. Why ? Currently, banks are happy to lend money to engineers who are high-earning, right? But when they lose their jobs, they cut back on expenses and can't repay the loans. Is it certain that the previously lent money will be recovered? We don't know. In any case, people have already started discussing this issue, and it has indeed impacted the entire industry. As you just mentioned, many software companies have started laying off employees . I recently spoke with the general manager of a large domestic software company about this, and his point of view makes a lot of sense. In the past, the personnel structure of software companies was an equilateral triangle : many engineers at the bottom and fewer at the top. With fewer people managing most of the staff, companies used to operate this way. But now, because the bottom-level work no longer needs to be done, it can be... The current situation with artificial intelligence is that we're cutting back on lower-level engineers , leaving many in the middle , and then the middle engineers remain. The result is that, over time, if this continues, it will form an inverted triangle. But the inverted triangle still has a problem: senior engineers always have to start from the bottom. If you cut all the lower-level engineers now, there's no one to train them. What happens when these middle engineers retire? Who will check if the code is written correctly? The current situation is that the coding is handled by AI, but checking if the code is working correctly still requires experienced engineers. Anthrapic's Mythos can theoretically do that, but even Mythos... Errors might occur, and someone still needs to check them. So, what was the conclusion we reached during our discussions? We believed that the ideal human resource structure for a company , especially a software company, should be rectangular. What does this mean? It means you absolutely cannot lay off all the junior engineers. You need to cultivate people who can become senior engineers. In other words, if you used to have 100 junior engineers, now you need to find at least 50. The purpose of these 50 people is to gradually cultivate them into engineers with senior experience who can assist in reviewing these programs in the future. This is a very practical business profit consideration , or is it a vision of cultivating talent? No, it's a business consideration because if junior engineers don't even have the opportunity to work, how can they... If you want to become a senior engineer, you must train them yourself. Otherwise, one day you'll find that all the senior engineers have retired, and you won't be able to find any in the market because everyone has laid off all their basic engineers. So our conclusion is that you must train them yourself. Every company must train its own basic engineers so that there will be enough senior engineers in the market. But the key is, you used to need 100 people, but now you only have 50. So what will the future world be like? The future world will become a K-shaped world. At this point in time, if you can keep up, you will rise to the top, and in the future, companies will rely on you even more for senior engineers. But if you don't keep up, you will fall behind and be eliminated. Frankly speaking, this group... What should people do? I don't know. I've thought about it for a long time but I don't have an answer. So I think I should give you, the audience, the most important concept, especially if you already have children or are in middle or high school or above. You need to start thinking about this: you must make your children interested in technology from a young age. This is very important. Why? Because in the future world, people who are not interested in technology will not be competitive. Remember what I said? I'm not saying that they must become engineers. I'm not saying that. They can be lawyers, accountants, or do any job. But they cannot be indifferent to technology. When you are indifferent to technology, no matter what job you do, when a new technology is invented, you're afraid you won't want to use it. That's right, you 're afraid you won't want to use it. This is so difficult. This is not my problem. You absolutely cannot have this kind of thinking. Our generation can barely manage to get by because we're already of a certain age and hold certain positions , so the competition isn't in that direction. But your child is different; your child faces that kind of competition. So you must cultivate their aptitude for technology from a young age — what we just talked about— that they need to be tech-savvy . That 's why I founded the Qubo Technology Classroom. Because children aren't always interested in new technologies. How do you spark their interest? You have to find a way to explain things in a way that they can understand, at least understand , and don't feel like they can't absorb anything. You might say it's too difficult for young children, but I think it's more challenging for junior high school graduates. For students above junior high school level, understanding technology is essential. Therefore, when designing these textbooks, I based them on a junior high school graduation level, ensuring that students above junior high can understand them. This is because everyone attends junior high school, which is compulsory education. I wanted to start from that basic level and cultivate children's sensitivity and understanding of technology. Sensitivity is very important. Mr. Qu, I'd like to ask you a question related to this : I recently heard Huang Renxun say that he believes the most powerful and popular major in the future is English. He believes that mastering the logic of language is essential to communicating with technology , regardless of which technology company's product it is. He considers language proficiency extremely important. What's your opinion on his statement? He once said he wanted to study biotechnology. He's changed his mind a lot , but I think one thing he'll never change is his core belief. Besides the previously mentioned need to be tech-savvy, there's another crucial point: you must be interdisciplinary. Simply put, you absolutely cannot be proficient in only one field. The more fields you know, the stronger your competitiveness. So, CEO Huang Renxun wasn't wrong; he said speaking English is a great skill , but that's only half the story . You can't just know English. If you're fluent in English grammar, proficient in artificial intelligence, and can program, then you're very competitive. Therefore, you must have at least two specializations. Thank you, Qu Bo. Today, our topics cover many different themes, starting with DeepSs... From the chips used in Eek to the ethics of AI , to the discussions surrounding Gaode Maps (a Chinese map service), and the fact that users have their own preferences when using it, to the global AI layoffs that raise questions about the priorities of education for the next generation, Qu Bo explained all these points in a very clear and concise way. We are currently at the forefront of a very special era , a turning point. Will this K-shaped society move upward or downward? We will continue to follow up on this program and hope to have the opportunity to invite Qu Bo back to the program to analyze the latest technological trends with us again. Due to time constraints, the program ends here today. See you on the same airwaves next week at the same time. Bye-bye- bye!