This Part 2 (with a summary by Joanne Z. Tan) is the second 10-minute segment of a 3-part recording of Jensen Huang’s 30-minute talk in Stanford University. (Here is the link to Part 1.)
To watch Part Two as a 16-minute video
To listen to Part Two as a 16-min podcast
Jensen Huang is the co-founder & CEO of Nvidia. As of today, Nvidia has a four trillion dollar valuation. It is the most valuable company in the world and in human history. Jensen shared insights for 30 minutes with a Stanford audience at the Asian American Pioneer Medal Symposium on July 26, 2025.
Joanne Z. Tan, brand strategist, thought leadership coach at 10 Plus Brand, Inc. was in the audience and recorded Jensen Huang’s entire talk, which is broken down into three parts, with her summaries and comments. (Joanne Tan did NOT use any AI to summarize or comment).
Jensen Huang continued: “Of course, the belief has to come from first principle thinking. You can’t just randomly come up with some crazy belief, and go along with it. The belief has to come from first principle computer science or physics or economics. First principle reasoning, added to it, was inspiration.” He states that principled thinking was really important in arriving at a belief system or vision. “Then you gut check that literally every single day,… continuously re-evaluate our strategies and our decisions, and somehow the assumptions that led up to that strategy, or that decision.”
In light of the performance of GPU over the last 10 years increasing 100,000 times, Jensen Huang explained how Moore’s Law is evidenced in computation, AI advancement, and the dramatic cost reduction. Then he observed that Nvidia’s computing platform can enable changes in algorithm, architecture, system, silicon, and fundamental computation – all at the same time. A decade ago, Jensen predicted that the entire data center will be one unit of a computing network. A single person can write an application that activates every single processor in the data center. This vision led Nvidia to reinvent and reframe itself from a chip company to building entire data centers. Today, they architect a “nominal” computer from the ground up, and build every single piece of technology within it, valued at $50 billion.
As new advancements arrived – such as AlexNet and computer vision – Jensen asked the fundamental questions by breaking down all the way to the first principles, fundamental electronics, in reinventing the way to compute and Nvidia itself.
Going back to Moore’s Law, when AI technology’s cost drops by 100,000 times every ten years, “it becomes a commodity everybody should squander”, when a large language model that can memorize and generalize from all of the data of humanity.
In answering the question if AI will take jobs from humans, Jensen Huang traced that it took 2-3 years for AI to develop from existence proof to superhuman capability; and 2-3 years from computer vision to super-vision to speech recognition; and another 2-3 years to solve math problems; and 2-3 years later, AGI. Now AGI is solving math, reading, and science problems, and all kinds of tests. “It will take less than two years to achieve superhuman levels”.
Huang analogized AI to test taking skills, which is not the same as being fundamentally intelligent. Tasks are inside a job, AI can automate tasks, increase productivity, but “the purpose of our job is different than doing those tasks”. Going back to first principle reasoning: What is AI? “AI is about solving these cognitive, reasoning, planning, and search problems.” “And it comes down to one thing: Do you have more ideas than you wish you could pursue if you were more productive? Do you have more dreams, and they can come true, if you had more time? And if you have those things, productivity is going to result in your growth. But if you don’t, then productivity is going to result in more idle time.” “I think we will be busier than ever.”
(Jensen Huang’s remaining 10-minute talk, with my summary and comments, continues in Part 3.)
©Joanne Z. Tan all rights reserved.
================================
(Part 2 of 3:)
(The interviewer, Xuedong Huang, is currently the Chief Technology Officer of Zoom.)
(Jensen Huang:) If you look at all of the things that we built, you know, all the systems that we’ll talk about, some day, you have to believe what you believe, and you have to pursue that belief.
Now, of course, the belief has to come from first principle thinking. You can’t just randomly come up with some crazy belief, and go along with it. The belief has to come from first principle computer science or physics or economics, first principle reasoning, with inspiration, added to it, inspiration. And that, that, that principled thinking is foundational thinking. It’s really important in everything that we do; when it… when we ultimately come at, arrive at a belief system, a vision, if you will, then you gut check that literally every single day.
You know, I’m always, I’m always reevaluating our decisions, is something continuously re-evaluating our strategies and our decisions and somehow the assumptions that led up to that, that strategy or that decision or to change, then you have to reconsider reevaluating.
And so, that was, that was a big decision, and we believe what we believe in, and we went with it.
(Xuedong Huang:) That’s fantastic. So, this is Asian American Pioneer Reward award ceremony. So I want to take this opportunity to ask, as an Asian American, and you led your company to achieve a four trillion dollar valuation, the most valuable company in the world. And you have amazing, amazing breakthroughs in the last ten years. The Performance of GPU has increased by 100,000 times. So what is the human side of the story behind this amazing achievement in life story?
(Jensen Huang:) Yes. So, several things. Moore’s Law is, ten times over five years, so 100 times over ten years. And in the last ten years, we’ve advanced artificial intelligence by about a hundred thousand times. The first, the first observation is that our computing platform can, can enable us to change the algorithm, change the architecture, and change the system, change the silicon, and change how we do fundamental computation, basically all at the same time.
And we’ve expanded how we think about computation.
About a decade ago, I told the company that, that the entire data center is going to be one unit of computing network. And, and, a single person will be able to write an application that activates literally every single processor in the data center, that idea never existed before.
And we, we, we reason about, okay, that (if) future were to happen, what would, what would we have to reinvent, what are the things that we are going have to become good at? How would, how would you think about our own work, and notice we went from a chip company to now we build entire data centers.
And, you know, a nominal computer for us these days could be $50 billion dollars. And the $50 billion dollars is just a massive undertaking.
And we, wer’are a company today that can start from the ground, I can imagine a blank sheet of paper in front of me and literally architect the entire giant computer from the ground up. And we built every single piece of technology within it.
That reframing of our company was important to getting there.
Now the hundred thousand…What does that mean?
You know, oftentimes when something happens, in the case of AlexNet, we all saw computer vision become quite effective. Everybody was, was interested in it. A lot of people were impressed by it.
In our case, what we asked ourselves is, what was achieved? Why was it achieved? What does it mean to compute, to computing, and where can it go?
And so those fundamental questions when broken down, all the way down to, you know, fundamental electronics, allowed us to reinvent everything from the way we compute, the way we do software, the way we developed. Our company has changed.
And so you want to reason everything all the way back to first principles.
Now 100,000. What does 100,000 mean?
Let’s break down back to the first principles.
Remember, the reason why Moore’s Law was so powerful is it was the single, the greatest technology force, deflationary force, depreciating force, the world’s ever known.
That’s a big idea, that something that cost a dollar, a dollar 10 years ago, is a penny today.
Yeah. How would that change everything that you do if something important cost a dollar 10 years ago, cost a penny today?
Well, obviously you could do work that’s a hundred times greater scale at the same price, that’s one way of thinking about it; Or you can think about it from another perspective, which is computation is a, is effectively a commodity you don’t have to worry about.
And so, if you’re trying to solve a problem, if computation is involved in that somehow, well, its cost won’t matter by the time that you figure out how how to apply it, it would have dropped in price and cost by a hundred times.
Well, now apply 100,000 times. So what does it mean when computation and intelligence drops in cost by 100,000 times every ten years, it becomes a commodity that nobody, EVERYBODY should squander. You should squander.
What… it led to, this observation led to somebody saying, let’s take all of the data in the world, and let’s put it into an artificial intelligence neural, neural model, and compress to memorize and generalize from all of the data of humanity.
That’s a big deal.
Let’s do with large language models.
That, that observation, comes from realizing it’s a hundred thousand times cheaper now.
(Xuedong Huang:) So this is an amazing story behind all the AI advancement. If you think about 10 years ago, the industry has achieved the first speech recognition and human parity. Today, both Google and OpenAI have achieved Olympics gold medal milestone, ten years, all using GPU you and your company created. So tell us all, what do you think AI is going to evolve?
Is AI really going to take all the labor, and all the jobs away from all of us? And if that is the case, what is it you say to your grand, grand kids, about the sense of purpose for humanity?
(Jensen Huang:) Yeah. Obviously, very important question and, and, and again let’s reason through it, okay. The first observation, the first observation is, is, in technology, and now more, more so than ever, there’s the first point, which is that you can solve the problem kind through existence proof.
From existence proof to superhuman capability, you know, call it 2 or 3 years.
So, computer vision, super vision, 2 or 3 years later; speech recognition, superhuman, 2 or 3 years later; solving, you know, math problems, 2 or 3 years later; AGI. Right.
So we know that AGI is now (at) superhuman levels of solving math problems, reading problems, science problems, You know, every single test you can imagine, solving tests, you’re going to achieve superhuman levels in probably less than two years.
Okay. And so, so I just said something, solving tasks.
Now you got… you all know that taking tests is a skill. It doesn’t make you fundamentally intelligent. Successful people and people who do well in, in….
So when I met my wife, and I made her a bold claim, I was 17 years old, and I said, you know, if you, if you study with me every weekend, I promise you, you will get straight A’s in the next quarter. And that was a big, big, bold claim.
Well, it happened. And the reason for that, as you know, is doing well in school is a skill.
It doesn’t necessarily make you successful. There are many people who were unsuccessful in school and very successful in life. Okay. So my point is that it’s just a skill.
It’s no different than twirling a pencil. There is a lot of Asian American people do. Yes. And these days it’s like, (whir whir whirrrrr…) you know, and so and that’s just the skill, okay. You’re not going to do get a job, that won’t lead… unless there is a job for twirling pencils. That’s not going to make you successful, but it is a skill. Taking, taking tests well is solving problems as you know, it’s a skill.
And so we are going to … AI is going to do all this stuff instantly faster.
Now go back to the first principles and ask: what is AI?
AI is about solving these cognitive, reasoning, planning, search problems.
You know, if you could if you can frame a problem in that context, we could now solve it incredibly fast, at almost no cost, with incredible capabilities.
And so, for a lot of companies, what that means is that the productivity of your company is going up, of your people, are going to be incredibly high.
And you have to realize that you and I have a job, and inside our job, there are tasks. And those tasks can be automated.
Are… those tasks are being automated, but jobs are not. The purpose of our job is different than doing those tasks. It involves those tasks. And so those tasks are going to be done incredibly well now. Companies are going to be very, very productive.Society is going to be very productive.
And it comes down to one thing: Do you have more ideas than you wish you could pursue if you were more productive?
Do you have more, you know, more dreams that you could pursue? And they can come true, if you had more time?
You know, so it comes down to:
Do you have more dreams?
Do you have more aspirations?
Do you have more ideas?
And if you have those things, the productivity is going to result in your growth.
But if you don’t, then productivity is going to result in more idle time.
And so I think that, that’s the bottom line.
I think it is more likely that we will be busier than ever.
It is likely will be busier, and along with the, the thought experiment, suppose, you know as, as a, you’re, you’re, you’re, you’re working in collaborative team or maybe you’re, a manager or, maybe you are, you know, a, a professor, or you have a bunch of research assistants, suppose you give the research assistant a problem and the answer comes back, not in a week, but it came back in second.
So my question is: Are you going to be busier, or less busy?
You know, I sent off questions to my company, and problems for my company to help solve that are pillars of, you know, larger problems and larger systems I’m trying to solve. And I get the benefit of them having to work on it for a week, but if they come back in a minute, then I gotta go, right, I’ve got to go understand what the impact is, and reason about it some more, and think and dream some more, and so, so I think that the answer is, is because we have so many dreams and problems and challenges and hopes that, you know all that, ideas, I think we will be busier than ever.
===================================================
(To be continued in Part 3.)
To read the full transcript of Jensen Huang’s entire 30-minute talk (coming soon)
Please don’t forget to like it, comment, or better, SHARE IT WITH OTHERS!
– To stay in the loop, subscribe to our Newsletter
(About 10 Plus Brand: In addition to the “whole 10 yards” of brand building, digital marketing, and content creation for business and personal brands. To contact us: 1-888-288-4533.)
– Visit our Websites:
Phone: 888-288-4533
– Find us online by clicking or follow these hashtags:
