Jensen Huang: NVIDIA - The $4 Trillion Company & the AI Revolution | Lex Fridman Podcast #494

0:11:27 · watch on youtube.com↗

Our reading

The install base is the single most important part of a computing architecture, and everything else about it is secondary.

Their wordsThey come to a computing platform because the install base is large. Because a developer, like anybody else, wants to develop software that reaches a lot of people. So, the install base is, in fact, the single most important part of an architecture.

0:24:47 · watch on youtube.com↗

Our reading

Training is no longer limited by data but by compute, because most of the data models learn from is now synthetic.

Their wordsThe amount of data that we use to train models is going to continue to scale to the point where we're no longer limited… Training is no longer limited by… Data is now limited by compute. And the reason for that is most of the data is synthetic.

0:25:32 · watch on youtube.com↗

Our reading

Inference was never going to be the easy, cheap half of AI, because inference is thinking, and thinking is far harder than reading.

Their wordsthat was always illogical to me because inference is thinking, and I think thinking is hard. Thinking is way harder than reading.

0:27:44 · watch on youtube.com↗

Our reading

After pre-training, post-training and test-time scaling, the fourth scaling law is agentic: multiplying AI by spawning agents, and the whole loop scales on one thing, compute.

Their wordsAnd so the next scaling law is the agentic scaling law. It's kind of like multiplying AI. Multiplying AI, we could spin off agents as fast as you want to spin off agents. And so, you know, I… You know, I have four scaling laws.

0:28:12 · watch on youtube.com↗

Our reading

Intelligence is going to scale by exactly one thing, and that thing is compute.

Their wordsAnd so this loop, this cycle, is gonna go on and on and on. It kinda comes down to basically intelligence is gonna scale by one thing, and that's compute.

0:33:13 · watch on youtube.com↗

Our reading

The idea that AI will destroy software and make tools unnecessary is ridiculous: an agent, like a humanoid in your kitchen, will use the tools that already exist.

Their wordsYou know, a lot of people would say, "You know AI is gonna completely destroy software. We don't need software anymore. We don't even need tools anymore." That's ridiculous.

0:35:26 · watch on youtube.com↗

Our reading

OpenClaw did for agentic systems what ChatGPT did for generative ones.

Their wordsAnd I think OpenClaw did for agentic systems what ChatGPT did for generative systems. And I just think it's a very big deal.

0:38:51 · watch on youtube.com↗

Our reading

The cost of a token is falling by an order of magnitude every year, even as the price of the computers that make them goes up.

Their wordsYou know, our computer price is going up, but our token generation effectiveness is going up so much faster that token cost is coming down. It's just coming down an order of magnitude every year.

0:50:28 · watch on youtube.com↗

Our reading

Data centres should be built to degrade gracefully when the grid needs its power back, instead of demanding perfect uptime from utilities that have idle capacity most of the year.

Their wordsNow, the second thing is we have to build data centers that gracefully degrade.

0:58:01 · watch on youtube.com↗

Our reading

Engineer from the physical limits first, then improve, rather than shaving a few days off a 74-day process by continuous improvement.

Their wordsI don't love the other methods, which is continuous improvement. The problem with continuous improvement, it… First of all, you should engineer something from first principles at the speed, you know, with speed of light thinking. Limit it only by physical limits, and physics limits. And after that, of course you would improve it over time.

1:03:47 · watch on youtube.com↗

Our reading

China is the fastest-innovating country in the world today, on the strength of its talent, its open-source habit and its ferocious internal competition.

Their wordsSo you get this rapid, incredible great talent, rapid innovation because of open source and just, you know, the nature of friends, and, and insane competition. Among the company, what emerges is incredible stuff. And so this is the fastest innovating country in the world today

1:07:37 · watch on youtube.com↗

Our reading

Open-source models are fundamentally necessary for most industries, countries and researchers to join the AI revolution at all.

Their wordsAnd if everything is proprietary, it's hard to do research and it's hard to innovate on top of, around, with. And so… Open source is fundamentally necessary for many industries to join the AI revolution.

1:28:10 · watch on youtube.com↗

Our reading

Someone paying a thousand dollars per million tokens is just around the corner, because intelligence is a product that segments by grade like phones do.

Their wordsYou know, the idea that somebody's willing to pay $1000 per million tokens is just around the corner. It's not if, it's only when.

1:29:08 · watch on youtube.com↗

Our reading

World GDP growth is certain to accelerate, and the share of GDP spent on computation will be a hundred times what it was, because a computer is now a factory rather than a warehouse.

Their wordsAnd so when you take these things in combination, I am absolutely certain that the world's GDP is going to accelerate in growth. I'm absolutely certain the percentage of that GDP that will be used for computation will be 100 times more than the past

1:29:58 · watch on youtube.com↗

Our reading

NVIDIA can be a three-trillion-dollar-revenue company in the near future, because nothing physical stands in the way.

Their wordsAnd then the rest of it, to me, is: is it possible for NVIDIA to be a, you know, $3 trillion revenue company in the near future? The answer is, of course, yes. And the reason for that is because it's not limited by any physical limits.

1:56:31 · watch on youtube.com↗

Our reading

Asked when an AI could build a billion-dollar business on its own: AGI has already arrived.

Their wordsI think it's now. I think we've achieved AGI.

2:00:38 · watch on youtube.com↗

Our reading

The number of software engineers at NVIDIA will grow, not shrink, because the purpose of a job and the tasks it uses are related but not the same, as radiology already showed.

Their wordsThe number of software engineers at NVIDIA is gonna grow, not decline. And the reason for that is because the purpose of a software engineer and the task of a software engineer coding are related, not the same.

2:01:45 · watch on youtube.com↗

Our reading

Coding is now specification, and the number of people who can do it just went from thirty million to a billion.

Their wordsSo the question is, how many people could do that? Describe a specification for a computer to go… telling the computer what to go build. How many people? I think we just went from 30 million to probably 1 billion. And so every carpenter in the future will be a coder

2:07:20 · watch on youtube.com↗

Our reading

Between two graduates in any field, he would hire the one who is expert at using AI.

Their wordsIf we were to hire a new college graduate today, and I have a choice between two, one that has no clue what AI is and one that is expert in using AI, I would hire the one who's expert in using AI.

2:13:46 · watch on youtube.com↗

Our reading

Intelligence is a commodity, and it is not the same word as humanity.

Their wordsI don't over-fantasize about, and I don't over-romanticize about intelligence. Intelligence is… And people have heard me say it before, I actually think intelligence is a commodity.

2:17:42 · watch on youtube.com↗

Our reading

He does not believe in succession planning; the real work is passing on knowledge continuously, and he hopes to die on the job.

Their wordsAnd so some of the things that of course are practical things, like how do we think about succession planning? And I'm famous in saying that I don't believe in succession planning.