Greatness is always achieved through perseverance! The latest interview with Huang Renxun from Stanford University: entrepreneurial journey, overcoming challenges, and the future of AI.

Author: Li Xiaoyin

Source: Hard AI

Key Points:

1. Faced with NVIDIA's 80% plunge to the "low point," returning to the "core" of the matter, stick to what I believe in, then nothing changes, and continue to move forward.

2. At the inception of NVIDIA, the first major decision was to determine that 3D graphics technology would be the first "killer app." The company's mission is to "build special computers to solve problems that ordinary computers cannot solve."

3. All of NVIDIA's work over the past thirty years has revolved around technology and the market, which is also NVIDIA's core: create technology before creating demand. In the next decade, NVIDIA's biggest challenges will come from technology and the market, as well as from industry, geopolitics, and society.

4. NVIDIA has made a unique contribution to driving the future of computing, which is one of the most important tools for humanity.

5. The latest breakthrough in AI is deep learning, another important breakthrough is the Reinforcement Learning with Human Feedback (RLHF) technology invented for language models, and NVIDIA has found a solution to implement this technology at the system level.

6. In the future, the way information is processed will undergo fundamental changes, generative AI will start from an information "seed," and the future of computing will rely heavily on generation rather than retrieval.

7. Regulation of AI in the future will come from two aspects: social regulation and product service regulation. The social issues brought about by the development of AI should be subdivided and regulated.

8. With the development of technology, information processing and software development will undergo fundamental changes, leading to significant adjustments in industry structure and organizational methods. Reshaping organizational structures aims to discard traditional hierarchical management and advocate for a more equal and open organizational culture.

9. A culture that does not believe in "information equals power" emphasizes the importance of trust and empowerment of employees. With only 30,000 employees, NVIDIA is the smallest large company in the world, but each employee is given significant power.

10. It is hoped that NVIDIA will be remembered by history as "changing everything" by persistently doing what they are good at and passionate about.

In the past year, riding the escalating wave of AI, holding the "hard currency" GPU, NVIDIA has successfully taken the throne of "AI chip leader," with the company's stock price soaring by 300%, reaching a market value of $2 trillion, ranking second in the United States.

On March 4th local time, NVIDIA CEO Huang Renxun, whose net worth has surged to the global TOP21, went to Stanford Graduate School of Business to participate in an interview at the "View From The Top" salon event.

In this latest public interview, Huang Renxun still appeared in his classic "black leather jacket." As a graduate in electrical engineering 34 years ago, Huang Renxun, as an alumnus of Stanford, shared how he founded NVIDIA, successfully obtained the first funding, established the company's organizational structure, and built the first "killer app." In addition, he also expressed his views on the artificial intelligence revolution and how to get through the "low point."

Striving to Create NVIDIA

Q: Jensen, it's a great honor to have you here. Thank you for coming.

A: I'm delighted to be here.

Q: Thank you. To commemorate your return to Stanford, let's start from the moment you first left. You chose to join the then very attractiveLSI Logiccompany, working with top figures in the tech industry, but decided to leave and start your own venture. What motivated you?

A: I was an engineer at LSI Logic, while Chris and Curtis were at Sun at the time. Working with some of the top figures in computer science at the time, including Andy Bechtolsheim, we were involved in building workstations and graphics workstations. One day, Chris and Curtis expressed their desire to leave Sun and start a business, inviting me to join them. Despite having a great job, they insisted on inviting me to start a company together. So, whenever they visited, we would meet at Dennis, by the way, that's my alma mater. I had worked as a dishwasher, my first job, and I did it very well. We started brainstorming, right in the midst of the microprocessor revolution in 1993 and 1992. The PC revolution was just beginning, Windows 95 had not been released, and Pentium was not announced yet. The importance of microprocessors was evident before the PC revolution was about to take off.

We thought, why not start a company? The goal was to solve problems that general-purpose computing couldn't solve. So, this became the mission of the company: to build special computers to solve problems that ordinary computers couldn't solve. To this day, we are still focused on this. If you look at the markets we have expanded into because of this, including computational drug design, weather simulation, material design, etc., we are very proud of it. There's also robotics technology, autonomous driving cars, what we call automated software in artificial intelligence, and more. We push technology to its limits, ultimately making computing costs almost zero, which led to a whole new way of software development, where computers write software themselves, which still amazes me to this day. That's our journey. Thank you all for being here.

Q: Today, these applications have caught the attention of all of us. But at that time, the CEO of LSI Logic convinced his biggest investor, Don Valentine, also the founder of Sequoia, to meet with you. I see many founders here eagerly waiting, but how did you convince Silicon Valley's most sought-after investor to invest in a team of first-time entrepreneurs building new products for a market that didn't exist yet?

A: I was at a loss on how to create a business plan, so I walked into the only bookstore that was still open. In the business book section, I found a book by a familiar author - Gordon Bell. The book was intimidatingly thick, titled "How to Write a Business Plan," straightforward and highly targeted. It seems like Gordon wrote only for a few people, and I unexpectedly became one of them. I should have realized it wasn't a wise move to buy it, after all, Gordon's outstanding wisdom meant he had a lot to impart.

Picking up this 450-page book, I soon realized that if I immersed myself in it, by the time I finished reading, my company would be at its end. At that time, Laurie and I could only sustain our livelihood for six months, along with our children Spencer, Madison, and a pet dog. The financial pressure of life necessities forced me to give up deep reading. So I didn't write a business plan. I just had a conversation with Wilf Corrigan. One day he called me and said, "Hey, you left the company, you didn't even tell me what you're doing. I want you to come back and explain it to me." So I went back to explain this to Wilf. In the end, Wilf said, "I don't know what you said. That was one of the worst elevator pitches I've ever heard." Then he called Don Valentine. He called Don and said, "Don, I'm sending a kid over. I want you to give him money. He's one of the best employees LSI Logic has ever had."

What I learned from this experience is that you may have had successful interviews or failed ones, but you can't escape your past. Therefore, strive to shape a good past.

I truly became the best dishwasher at Denny's, I planned my work well and was methodical. I am the kind of person who works in an organized manner. Then I worked hard. After that, they promoted me to Busboy. I am sure I was the best Busboy Denny's has ever had. I never left my position idle, I never returned empty-handed, and I worked with high efficiency. Therefore, I eventually became CEO . I am still striving to be a good CEO .

Against the Wind, Upholding What I Believe

Q: Facing fierce competition, standing out among 89 companies building the same product is indeed not easy. With only 69 months of remaining operating funds, you realized that the initial plan was no longer feasible. In such adversity, how did you decide on the next steps to save the company?

A: We founded this company with the aim of advancing accelerated computing. But the key question was, what would this technology be used for? What is its killer application? Our first major decision, and the reason why Sequoia Capital supported us, was to determine that the first killer application would be 3D graphics technology, with the application scenario being video games. At that time, producing low-cost 3D graphics technology was almost impossible, with image generators produced by companies like Silicon Graphics priced at millions of dollars. Moreover, the video game market at that time was almost non-existent. Thus, we faced a huge challenge: how to combine a difficult-to-commercialize technology with an almost non-existent market. It was this intersection that defined the establishment of our company. I still remember vividly the profound insight Don gave me at the end of my speech. He said, "Startups should not invest in or collaborate with other startups." This implies that NVIDIA's success relies on the success of another startup - Electronic Arts. As I was leaving, he reminded me that Electronic Arts' Chief Technology Officer was only 14 years old and still needed his mom to drive him to work. He wanted to remind me through this story that our success depends on such partnerships. Then, half-jokingly, he warned me that if I lost his money, he would make me pay. This is my most profound memory of that meeting.

Nevertheless, we have achieved some milestones. We spent several years creating the gaming market for personal computers, a long process that we are still continuing today. We realized that not only do we need to create technology, but also invent a new computational graphics method to transform technology worth millions of dollars into something that can be installed in a computer for only three to four hundred dollars, and we also had to create this entirely new market. Therefore, we not only need to create technology but also create a market. This spirit of innovation defines NVIDIA.

Today, almost everything we do revolves around creating technology and markets. That's why people say we have a technology stack, an ecosystem. This has become the core of NVIDIA for the past 30 years: in order to create conditions for people to buy our products, we must invent new markets. This is why we are at the forefront in fields such as autonomous driving, deep learning, and computational drug design and discovery - we are creating technology while also creating markets.

Then, Microsoft introduced the Direct3D standard, giving rise to hundreds of companies. A few years later, we found ourselves competing with almost everyone. The 3D graphics technology we invented was incompatible with Direct3D, putting us in a very awkward position. We faced a choice: either reset the company or go out of business. But we didn't know how to build it the Microsoft way. I remember at a weekend meeting, we discussed how we now had 89 competitors. We knew our way was wrong, but we didn't know what the right way was. Fortunately, we found books on the OpenGL pipeline at Fry's Electronics, defining how computers process graphics. I bought a few books and brought them back to the company, telling my team that I found our future. We implemented the OpenGL pipeline according to the book's methods, creating something unprecedented in the world. Many lessons learned in this process laid the foundation for our company's later success.

That moment gave us great confidence. It proved that even when you know nothing about something, it is still possible to successfully create the future.

That's my attitude towards everything now. When someone introduces me to something I've never heard of, my first reaction is always, "How difficult is this?" Often, it's just a textbook or a research paper away. Therefore, I spent a lot of time reading research papers. If this is true, you can definitely learn how others do things, then go back to the original principles and ask yourself, given today's conditions, my motivation, tools, and the changes in the situation, how would I do this again? How would I reinvent the whole thing? If I were to build a car today, would I start improving it gradually from the 1950s and 1900s? How would I build a computer today? How would I write software today? So, I always go back to the original principles, even in today's company, I would reposition myself because the world has changed. The way we used to write software in the past was holistic, designed for supercomputers, but now it's decentralized. How do we view software today, how do we view computers today, how do we view things, always bring your company, bring yourself back to the original principles, which creates a lot of opportunities.

If you send me something and ask for my help in reviewing it, I will do my best and show you how I would handle it. In this process, of course, I have also learned a lot from you. Right? You have given me many seeds of information. I have learned a lot. Therefore, this process is beneficial to me. Sometimes, it does take a lot of energy because, to add value to others, and they are very smart from the beginning, I am surrounded by very smart people, you have to at least reach their level. You have to enter their mindset, which is really difficult. It requires a huge emotional and intellectual effort. So, when I deal with these matters, I feel tired. I am surrounded by many great people.

In theory, a CEO should have the most direct subordinates because those reporting to the CEO require the least management. For me, it doesn't make sense for a CEO to have very few direct subordinates, except for one fact I know to be true. That is, the CEO's knowledge and information are considered extremely valuable and highly confidential. You can only share it with two or three people. Their information is so priceless, so confidential, that they can only share it with a few more people.

I don't believe in a culture and environment where having information equals power. I hope all of us can contribute to the company, and our position in the company should be related to our ability to solve complex problems, guide others to achieve greatness, inspire, empower, and support others. This is the reason for the management team's existence, to serve all other employees in the company, create conditions for all these amazing people to voluntarily work for you. Compared to all the top tech companies in the world, they have chosen to work for you voluntarily. Therefore, you should create conditions for them to engage in the work of their lifetime, that is my mission. You may have heard me say it very clearly. I believe my job is simply to create conditions for you to engage in the work of your lifetime. So, how can I do this? What are these conditions like? What results should these conditions bring? Great empowerment.

You can only be empowered when you understand the context, right? Right? You must understand the context of your situation so that you can come up with great ideas. Therefore, I must create an environment for you to understand the context, which means you must be informed. And the best way I've been told is to minimize the level of distortion in the information between us.

That's why I often reason things out in situations like this or in front of an audience. I start by saying, these are the initial facts we have. This is the data we have. This is how I will reason. These are some assumptions, these are some unknowns, these are some knowns. So, as you reason, you've now created a highly empowered organization. NVIDIA has 30,000 employees, making it the smallest large company in the world. We are a small company, but each employee is given tremendous power, making smart decisions on my behalf every day.

This is because they understand my position. I am very transparent with people. I believe I can trust you with the information.

Now, many times the information is hard to hear, the situation is complex, but I believe you can handle it. You, you know, many people listen to me, you are all adults. You can handle this. Sometimes they are not really adults. They have just graduated. I'm just kidding.

I know when I first graduated, I was hardly an adult. But I, I was fortunate to be trusted with handling important information. So I want to do that kind of thing. I want to create conditions for people.

NVIDIA's Core: Pioneering a New Way of Computing

Q: Yes, the way you have applied this technology has proven to be revolutionary. You gained all the momentum needed for an IPO because in the next four years, your revenue grew ninefold. But in all this success, you decided to slightly shift NVIDIA's focus on innovation, based on a phone call with a chemistry professor. Can you tell us about the content of that call and how you connected what you heard with your subsequent actions?

A: I remember, the core of the company is pioneering a new way of computing. Computer graphics was the first application, but we always knew there would be other applications. So, image processing came, particle physics came, fluid dynamics came, and so on, all these interesting things we wanted to do. We made the processors more programmable so we could express more algorithms. Then one day, we invented programmable shaders, which made all forms of imaging and computer graphics programmable. It was a great breakthrough. So, we invented it. Based on this, we tried to find ways to express more complex algorithms that could be computed on our processors, which is very different from CPUs. So, we created this thing called CG. I think it was around 2003, C stands for GPU. It was about three years before CUDA.

The same person who saved the company, Mark Kilgard, wrote that textbook. CG was very cool. We wrote the textbook, started teaching people how to use it. We developed tools, rules, and so on. Then some researchers discovered it. Many Stanford graduate students were using it. Many people who later became NVIDIA engineers were playing with it. I, a doctor from a hospital in Massachusetts, picked it up and used it for CT reconstruction. So I flew over to see them and asked, "What are you guys doing?" They told me. Then, a computational quantum chemist used it to express his algorithm. So I realized that there might be some evidence that people might want to use this. This gave us more confidence that we had to do this, that this computational form could solve problems that ordinary computers truly couldn't solve, and it strengthened our belief to keep moving forward.

Whenever you hear about something new, you seem to cherish that sense of surprise. This seems to be a theme in your leadership position at NVIDIA. It's like you've always been betting long before the technological turning point, standing there in your black leather jacket ready to catch it when the apple falls from the tree. How did you do it?

You always seem to be like a deep-sea catcher. You act according to core beliefs. We firmly believe that we can create a computer that can solve problems that ordinary processors cannot solve. What CPUs can do is limited, and what general computing can do is limited. Then there are some interesting problems we can solve. The question always is, are these interesting problems just interesting, or can they also be interesting markets? Because if they are not interesting markets, it is not sustainable. NVIDIA went through about a decade of investing in this future, and the market did not exist. At that time, there was only one market, which was computer graphics, for 10 to 15 years. The market that drives NVIDIA today simply did not exist.

So, when everyone around you, our company and NVIDIA's management team, all the amazing engineers who have worked with me to create this future, your shareholders, your board of directors, all your partners, you move forward with everyone, but there is no evidence of the market, that is a real challenge. The fact that technology can solve problems, and the research papers that may result from it, are interesting. But you are always looking for that market. However, before the market exists, you still need early indicators of future success. Our company has a saying called Key Performance Indicators (KPI). Unfortunately, KPIs are hard to understand. I find KPIs hard to understand, what makes a good KPI. Many people, when we look for KPIs, look at gross profit margin. That's not a KPI, that's a result. You are looking for something that can indicate early positive results as soon as possible, okay? The reason is that you want to show early signs that you are on the right track. Therefore, we have this saying called EIOFS, Early Indicators of Future Success. This is helpful for people because I have always used it to bring hope to the company, hey, look, we solved this problem, we solved that problem, we solved this problem. The market doesn't exist, but these are all important issues, that's the company's purpose. In order to solve these problems, we want to be sustainable. Therefore, the market must eventually exist.

However, you want to separate the results from the evidence that you are doing the right thing, okay? This is how you address the issue of investing in distant things and having the belief to continue, defining as early as possible the indicators that you are doing the right thing. Therefore, it all starts with a core belief. Unless something changes your mind, keep believing in it and look for early indicators of future success.

Question: What early success indicators does NVIDIA's product team use?

Answer: Various ones. I, even before seeing the paper, encountered some people who needed my help with something called deep learning. At that time, I didn't even know what deep learning was. They needed us to create a language specific to a certain field so that all their algorithms could easily be expressed on our processors. We created something called cuDNN, which is essentially SQL in storage computation. This is neural network computation. We created a language for this, you could say it's the OpenGL of deep learning. So, they needed us to do this so they could express their math, and they didn't understand CUDA, but they understood their deep learning. So, we created this thing for them in between.

We did this because even if these researchers had no money, this is a great skill of our company. Are you willing to do something even if there is no financial return, or it may be very distant, as long as you believe in it, we ask ourselves, is this work worth doing? Does it advance some important scientific field? Note, this is what I have been talking about from the beginning, we don't find inspiration from the size of the market, but from the importance of the work. Because the importance of the work is an early indicator of the future market. No one needs to make a business case, no one needs to show me a profit and loss statement, no one needs to show me financial forecasts.

The only question is, is this important work? If we don't do it, will it happen without us? Now, if we don't do something? And things will still happen without us. This actually brings me great joy. The reason is, you can imagine the world getting better. You don't even need to raise a hand. This is the definition of ultimate laziness. In many ways, you want to have this habit. The reason is, you want the company to always be lazy about things others can do. If others can do it, let them do it. We should choose things that will collapse if we don't do them. You have to convince yourself that if I don't do this, it won't be done. This is important. If the work is difficult, far-reaching, and important, it will give you a sense of purpose. Does that make sense? So, our company has always been choosing these projects. Deep learning is just one of them. And the first sign of success was the fuzzy cat proposed by Andrew Anne. Then Alex Kerchewski detected the cat, not successful every time, but successful enough to make people feel like it might take us somewhere. We reasoned about the structure of deep learning. We are computer scientists, we know how to make things work. So, we convinced ourselves that this could change everything. Anyway, that's just an example. Question: Therefore, the choices you made have brought significant returns both literally and metaphorically. However, you had to guide the company through some very challenging periods, such as during the financial crisis, when Wall Street didn't believe in your bet on machine learning, causing the company's market value to drop by 80%. How did you lead the company and keep employees motivated during such times?

Answer: My reaction during that period was the same as my reaction this week. You, earlier today, asked me about this week. My pulse is exactly the same. This week is no different from last week or the weeks before.

When your stock price drops by 80%, you just want to wear a T-shirt that says, "It's not my fault." But more importantly, you just don't want to leave your bed. You don't want to leave your house. All of that is true.

But then you get back to doing your job. Wake up at the same time in the same way, prioritize my day in the same way, and get back to what I believe in. You have to do a core check and always come back to the core. What is the most important thing, and check them off one by one. Sometimes it's helpful: my family loves me, okay, check; double-check, right? So, you just check it off, get back to your core, and then get back to work.

And every time you talk, come back to the core, keep the company focused on the core. Do you believe it? What has changed, the stock price has changed, but what else has changed, have the laws of physics changed, has gravity changed? What has changed of what we assumed, what we believed, what led us to make decisions? Because if those things have changed, you have to change everything. But if those things haven't changed, you don't change anything. Keep moving forward. Yes, that's how you do it.

Question: When talking to your employees, they said you were trying.

Answer: Avoiding the public, including employees.

I'm just scared. Unfortunately, leaders have to be seen. That's the difficult part. I was a student of electrical engineering, and I was very young when I went to school. I was only 16 when I went to college. So, I was very young in everything I did. Therefore, I was a bit introverted, a bit shy.

I don't like public speaking. I'm happy to be here. I'm not saying, but it's not something I naturally do. And when the situation is challenging, standing in front of the people you care about is not easy. The reason is, can you imagine a company meeting where our stock price has dropped by 80%? And as the CEO, the most important thing is to come face to face with you and explain it. Part of the reason you're not sure why, part of the reason you're not sure how bad it will be, you just don't know these things. But you still have to explain it, face all these people, you know what they're thinking. Some may think we're doomed to fail. Some may think you're a fool. Some may be thinking other things. There are many things people are thinking, you know what they're thinking, but you still have to stand in front of them, deal with it, do the hard work.

Q: They might be thinking about those things, but at times like this, no one in your leadership team leaves.

A: That's what I always remind them. I'm just kidding. I'm surrounded by geniuses. Yes, other geniuses. NVIDIA is well known for having the world's best management team. It's the deepest technology management team in the world. I'm surrounded by many of them, they are all geniuses. The business team, marketing team, sales team, they are all incredible.

Empowerment is crucial in corporate organizations

Q: Incredible. Your employees say your leadership style is highly engaging. You have 50 direct subordinates. You encourage people from various parts of the organization to send you the top five things on their minds, and you constantly remind people that no task is beneath you. Can you tell us, why did you intentionally design such a flat organization? How should we think about designing future organizations?

A: No task is beneath me. Because remember, I used to be a dishwasher, I'm serious. I've cleaned toilets before. I'm serious, I've cleaned a lot of toilets, more than all of you combined. Some toilets are invisible. I don't know how to tell you. You know, that's life. So, you can't show me a task that is beneath my ability. The reason I don't do it is just because, is it beneath me.

If you send me something, and you want me to provide feedback on it, I can serve you, in the process of reviewing it, share with you how I reasoned. I contributed to you. I enable you to see how I reason about something. By reasoning, understanding how someone reasons about something, gives you the ability. He says, oh my god, this is how you reason about these things. It's not as complicated as it looks. This is how you reason about very vague things. This is how you reason about things that cannot be calculated. This is how you reason about things that look very scary. Do you understand? So, I always show people how to reason about things. Strategic matters, you know, how to predict something, how to break down a problem. You're just empowering people everywhere. So, that's how I look at it.

The future of computing will heavily rely on generation rather than retrieval

Q: Now let's talk about the topic everyone is most concerned about - AI. Last week, you mentioned that generative AI and accelerated computing have reached a turning point. As this technology gradually enters the mainstream, what applications are you most looking forward to?

A: We need to go back to basics and think about what generative AI is. Through extensive learning and data analysis, we can now understand the meaning behind this data. Not only do we understand the meanings in various forms, but we can also achieve transformations between them. This transformation is essentially the generation of information. This means that in the future, the way information is processed will fundamentally change, and so will the way software development and application processing are done.

In the past, we relied on a retrieval-based model, where information was pre-recorded and retrieved based on algorithms. However, in the future, generative AI will start from a seed of information and generate more content through prompts, making computing highly dependent on generation rather than retrieval.

For example, in our current conversation, most of the information is generated in real-time, rather than simply retrieved. That's intelligence. In the future, our computers will largely operate in this generative way, rather than relying on retrieval.

For entrepreneurs, this shift means the need to reconsider which industries will be disrupted, our perspectives on networking, storage, and how our use of internet traffic will change.

From an organizational perspective, we should build organizations based on fundamental principles. If we are creating different things, why should our organizational structure be exactly the same? Whether it's building computers or providing medical services, our organizational structure should reflect the environmental characteristics we are in.

Q: I'd like to leave time for the audience to ask questions. This year, your theme is "Redefining Tomorrow." The question we pose to all guests is: As the co-founder and CEO of NVIDIA, if you could magically change one thing about tomorrow, what would it be?

Q: Should we have considered this question before? I might give an unsatisfactory answer. I believe it's not just one thing. There are many things beyond our control. Your mission is to make a unique contribution, live a meaningful life, do things that others in the world can't do, so that people can say after you're gone, the world is better because of you. That's how I live my life.

Q: I always try to imagine the future first, and then look back. Your question is the opposite of my way of thinking. I never look forward from the current position, but imagine myself in the future, and then review the past. It's easier this way. Looking back is like reading your own history. We did this, we did that, we overcame this challenge. Does that make sense? It's like your way of solving problems, first determine the desired outcome, then work backward to achieve it. I dream that NVIDIA can make a unique contribution to driving the future of computing, which is one of the most important tools for humanity.

This is not to say how great we are, but this is indeed what we are good at, and it's very difficult to do. We believe we can make a truly unique contribution. It took us 31 years to get to where we are today, but our journey is just beginning. It's extremely challenging. When I look back, I believe we will be remembered as a company that changed everything, not by what we say, but by persistently doing something we are very good at and passionate about.

NVIDIA Faces Two Challenges in the Next 10 Years

Q: I will graduate in 2023. My question is, how do you see the development of your company in the next ten years, what challenges do you think the company will face, and how are you preparing for them?

A: First of all, let me share the thoughts that come to my mind. When you mention challenges, a long list of challenges flashes through my mind, and I wonder which one to choose. To be honest, your question mostly makes me think of technical challenges because that was the topic of my morning today. If you had asked me yesterday, it might have been the challenges created by the market. There are some markets that I really want to explore. But we cannot fight alone.

NVIDIA is a technology platform company, we serve many other companies, helping them achieve the hopes and dreams they entrust to us. I hope the field of biology can reach a level similar to the chip design field 40 years ago when Electronic Design Automation (EDA) paved the way for everything we can do today. I believe we will enable them to achieve computer-aided drug design. We can now represent genes, proteins, and even cells, almost understanding the meaning of a cell, what does a cell really mean? If we can understand a cell like we understand a paragraph, imagine what we can do. So, I feel both anxious and excited about this. There are some technologies that I am particularly looking forward to, such as humanoid robots, we are at a turning point. If we can understand language, why can't we understand operations? Once you solve one problem, you will ask yourself, why can't we solve another problem? I am very excited about these advancements; it is a delightful challenge.

Of course, the challenges we face are not only technical but also industrial, geopolitical, and social. For example, global social and geopolitical issues, why can't we get along peacefully? Why do we amplify these issues? Why do we judge others so harshly in this world? You all already know this; I don't need to repeat it.

Expectations and Concerns Coexist for the Development of AI

Q: I am Jose, a graduate from the business school in 2023. My question is, are you concerned about the speed at which we are developing AI? Do you think some form of regulation is needed? Thank you.

A: Yes, the answer is both yes and no. We do need you. A recent major breakthrough in AI is deep learning, which has driven significant progress. Another important breakthrough is the invention of a new language model based on human feedback for reinforcement learning. I practice this reinforcement learning every day. It is my job, and for the parents in the room, you are also constantly providing this learning feedback.

Now, we have found ways to achieve this at a system level for artificial intelligence. Many other technologies are necessary for setting up protective measures, fine-tuning, and ensuring that the system complies with actual physical laws. For example, how do we generate codes that adhere to the laws of physics? Currently, some codes seem to float in space, not following physical laws. This requires technology to achieve.

In terms of regulation, there are two types: social regulation and product service regulation. I am not very clear about social regulation, but for product service regulation, we know exactly what to do. Agencies like the FDA and NHTSA have established regulations for products and services for specific purposes. Please do not introduce a super-regulation that spans all areas. The agency responsible for accounting regulation should not regulate doctors.

But I overlooked an important issue: How should we deal with the social impact of AI? Although I don't have a perfect solution, enough people are discussing it. It is important to break down these issues so that we do not overlook many routine tasks that we could have done because of excessive focus on one thing, resulting in harm. We should ensure that we are doing the right things there.

The person with the most black leather jackets in the world

Q: We have some quick questions and answers as usual for you.

A: Well, I was trying to avoid this. Okay, let's get started.

Q: Your first job was at Denny's. They now have a dedicated booth in your honor. What is your fondest memory? Also, what was your second job?

A: By the way, is there a booth for me at AMD, where I worked as my second job? Just kidding. I really enjoyed my time there. AMD is a great company.

Q: If there was a shortage of black leather jackets worldwide, what would we see you wearing?

A: Oh, don't worry, I have plenty in stock. I would be the only one not worrying.

Q: If you were to write a textbook, what would its title be?

A: Write a book? That's an impossible question you're asking.

Q: Lastly, if you could give one piece of advice to Stanford University students, what would it be?

A: It's not a word, but I would say: Have a core belief. Test it every day, pursue it with all your might, and stick with it for the long term. Surround yourself with people you love and enjoy the journey together. That's the NVIDIA story.

Q: Jensen, the past hour has been a delight. Thank you for sharing. Thank you very much.