JUANA SUMMERS, HOST:
We begin this hour looking ahead to the high-level meeting between President Trump and Chinese leader Xi Jinping tomorrow at the White House. AI is on the agenda. And it's also been the subject of a lot of dire forecasts recently. Leading AI industry voices have thrown out their best estimates of the probability that humanity will be wiped out by powerful AI. One Anthropic employee says it's more than 10% within the next decade, 70% within three years from one at OpenAI. Jensen Huang offered a different number in an interview with CBS News over the weekend.
(SOUNDBITE OF TV SHOW, "CBS NEWS SUNDAY MORNING")
JENSEN HUANG: 2030 is not going to be the end of the world. There is 0% chance that's going to be the end of the world.
SUMMERS: Huang is the CEO of the chipmaker Nvidia, which is arguably the most important company in AI. And he's also one of the key figures influencing President Trump's thinking on AI. To talk about what that means for the industry and for all of us, I'm joined now by Stephen Witt. He wrote a biography about Huang and joins me now. Welcome to the program.
STEPHEN WITT: Thank you for having me.
SUMMERS: So Stephen, when President Trump says that the idea of AI taking over the world or destroying humanity is a hoax, that we don't need guardrails and that AI will, instead, be the greatest economic development engine in history, how similar is that to Jensen Huang's view?
WITT: Those are exactly Jensen's talking points. Jensen is now the president's most influential adviser on technology issues. The two have become quite close. They call each other on the phone all the time. Jensen even caught a ride on Air Force One to China. So this is really - the president is really listening to Jensen on this.
SUMMERS: Do you have a sense of how it happened that they got so close? I mean, I just - I remember seeing a clip of Huang taking a call from President Trump onstage during a live event earlier this month.
WITT: Yeah. I think that was a little bit of a kind of stage-managed photo-op.
SUMMERS: Sure. Sure.
WITT: But it did happen. Trump wants to be close to Jensen because Jensen is a winner. Nvidia, Jensen's company, is the most valuable corporation in American history, even when you adjust for inflation. It is the biggest. Jensen, I think, wants to be near Trump because Trump has so much to offer him.
When I interviewed Jensen years ago, he told me that every problem in the world is actually an engineering problem. Now, that's true if it's a personal problem or a business problem or a political problem. You study the inputs and then you modify the inputs to get the outputs that you want. In this case, I think Jensen has realized that the inputs for Trump are, like, flattery - like, public flattery, nonstop public flattery. And the outputs are, you know, lifting restrictions on sales of Nvidia equipment to China and zero regulation on AI.
SUMMERS: Hearing Huang's views, they sound pretty optimistic. I wonder, is he an outlier when it comes to the broader AI industry?
WITT: He's not an outlier. He represents an entire camp of people who think this way. I should say, I interviewed tons of AI people. Most of the hardware people, most of the people who designed the networking equipment and the microchips, feel like Jensen does. They tend to have a very low sense of risk regarding this technology. They just don't see it.
Most of the software people, and especially the more abstract software people who work on neural nets, as well as philosophers and politicians - and the public, I should say - do see the risk. So there's a huge split here. This debate has been going on for at least 10 years. And we actually now have a lot of data on how AI performs and what it does. And there's still no clarity at all. Neither camp has moved at all.
SUMMERS: I mean, we can't have this conversation without talking about the dollars and cents of it all. Obviously, Jensen Huang has a strong financial interest in the continued development of frontier AI. I mean, he's selling the chips that power all of these models.
WITT: Yeah. I mean, sure. If we did a pause on AI, as some have suggested, Nvidia's stock price would go down, probably a lot. There's also the risk that, you know, China can just pull ahead if we stop developing AI. It's hard to get unilateral action.
SUMMERS: Yeah. Let's talk about the China piece of this. Earlier this year, President Trump authorized the sale of some advanced Nvidia chips to China, which had been banned under the Biden administration. And that would seem at odds with one aspect of AI that Trump does seem to be worried about, which is staying ahead of China in the race for more powerful models. Did Jensen Huang influence that decision to the degree that we know?
WITT: Oh, for sure. That's Jensen's doing. I mean, Jensen is really the only one (laughter) who wants to sell his equipment to China. Most of the U.S. kind of State Department and national security apparatus are opposed to it. Jensen has been remarkably good at getting what he wants out of the president.
SUMMERS: Do you get the sense that he's more influential than someone like a David Sacks, who was Trump's AI czar?
WITT: Well, you said the word was (laughter). I mean, yeah. Jensen's on the phone constantly with Trump. He's at the White House. He'll be at this state dinner with Xi Jinping later this week, and he was at the other state dinner in China in July. And I think Trump's point of view is - look, there was this very famous quote from the 1950s, which was, what's good for General Motors is good for America. I think now Jensen has convinced Trump that what's good for Nvidia is good for America.
SUMMERS: Thinking ahead to the Trump-Xi visit, what's your opinion? How likely do you think we are to see any sort of ground rules or guardrails agreed to, or even any just, like, substantive dialogue on this issue?
WITT: They'll talk about it. They're talking about a hotline. I don't think we're going to see much progress. I think even if Trump wanted to do it, I think it would be a struggle. And I don't think Trump wants to do it. You know, he's very afraid of losing the AI race to China. We're far ahead right now. But it is a real concern. And if we did pause, China would start to catch up to us. And I don't blame the government for not wanting that to happen.
So in the end, any kind of regulation we do have to be done in a way that is basically we act, but it also benefits China, even without their cooperation. This is why I advocate for something like the NTSB, which does air crash investigations. If we start publishing, you know, like, reports on how our AIs are going haywire, we can translate those into Chinese and make AI safer around the world. But unfortunately, I'm pessimistic, at least right now, about any kind of international cooperation on this issue.
SUMMERS: Stephen Witt is the author of "The Thinking Machine: Jensen Huang, Nvidia, And The World's Most Coveted Microchip." Thanks so much for coming on the show.
WITT: Thank you for having me. Transcript provided by NPR, Copyright NPR.
NPR transcripts are created on a rush deadline by an NPR contractor. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.
