AMY GOODMAN: This is Democracy Now!, democracynow.org, The War and Peace Report. I’m Amy Goodman, as we bring you Part 2 of our discussion with Sharon Weinberger, the longtime national security reporter and editor. Her new book is titled Valley of Death: How Big Tech Is Fueling the Future of War.

In a moment, Sharon, I want to talk with you, of course, about artificial intelligence and how it’s transforming warfare. But first, I want to ask you about Silicon Valley moving into the business of war and what that means for our collective future. Can you talk about the tech companies, from Peter Thiel’s to Elon Musk’s to Jeff Bezos’s Amazon, how they are insinuating themselves into the U.S. military complex?

SHARON WEINBERGER: Well, the pitch that Silicon Valley is making to the Pentagon and the government is that the types of thing that Silicon Valley excels in — artificial intelligence, autonomy, software — are the types of things that the Pentagon needs for modern warfare, particularly for a potential war against China, but also, as Silicon Valley would like to argue, useful for conflicts all around the world. That is their pitch.

What I think differentiates them from the defense company of old is that they really have— it’s a combination of power over the technologies that they’re building in terms of the ownership, but they’re also very, very involved in politics in a way that the defense-industrial complex old never was. You typically wouldn’t have the head of Lockheed Martin or Northrop Grumman expressing political views as you do here out of Peter Thiel, out of Alex Karp and out of some of these tech billionaires. It’s a very different and, I argue, much more powerful defense-industrial complex that’s being built in Silicon Valley now.

AMY GOODMAN: And talk about that. Talk about what Elon Musk says and what Alex Karp says — right? — the founder of Palantir, and where he comes from, right to his Haverford yearbook.

SHARON WEINBERGER: So, I’ll talk a little bit about Elon Musk, but I feel like he’s better known. So, of course, Musk came out very, very vocally in favor of Trump in the most recent election, you know, giving hundreds of millions of dollars to the campaign, campaigning for Trump, and then, of course, rose to sort of lead DOGE and the cuts in government. He has also, I mean, expressed a variety of political views, some bordering on the absolute bizarre, many border — I mean, sort of, you know, whipping up sort of race-related arguments, sort of all these bizarre political arguments, which, again, I think, for a private citizen, beliefs, good or bad, fine, but for someone who has such control over military technology, I think that is very dangerous. It is a problem for accountability for our democracy.

But let’s also take Alex Karp, the head of Palantir, which has been at various times, based on its stock market valuation, the most valuable defense company in the world. Alex Karp is interesting. He started off in college as sort of a social justice activist. His mother is Black. His father is Jewish. And in college, he identified as Black. He was very vocal, complaining about the Reagan administration, accusing it of contributing to racism. And one of the things that’s been described and that I found for my book was a picture. It’s something that wouldn’t be so out of nature now. He’s in his senior yearbook photo for Haverford College, a very liberal college. He’s very prominently draped in a keffiyeh.

And today, the Alex Karp who heads, in some respects, the most valuable defense company in the world has been very, very vocal both in taking a position in support of Israel, but also very, very publicly critical of the protests that were taking place on American college campuses after the war in Gaza started. And again, I don’t have a problem with anyone’s personal political beliefs, but when you have the head of a defense company that controls technology that’s critical to the Pentagon staking out a position on speech on American political — on American college campuses, I think that’s a very dangerous combination of power, and not something — I never thought I’d be in a position of defending older defense companies, and I’m not — I think they have their own problems with the revolving door — but that is something new that we’re seeing now, and I think not great for democracy.

AMY GOODMAN: Tell us about the Palantir commercial that aired during the 2024 Army-Navy football game. You write about this in your book’s epilogue, saying, quote, “The depiction of a bloodless war fought by robots is still fanciful … the type of autonomous technologies featured in the commercial have not yet been fielded, and Palantir isn’t building them. That fact didn’t matter to investors betting on vibes. Palantir’s astonishing market capitalization in late 2025 was more than that of any other defense company in the world, despite having significantly less revenue than the prime contractors.” Explain.

SHARON WEINBERGER: So, this is sort of the twin fallacy of Palantir. So, this commercial — and it’s sort of the criticism I would make of many defense commercials, and not just of Palantir, but for Palantir it was particularly strike — particularly stark.

So, it depicted what appeared to be a potential conflict in the Indo-Pacific, a potential conflict with China. And it shows these unmanned boats going to war, and then there’s literally a conductor who is sort of directing, like with a conductor stick, conducting where these ships go. And they take everything out with AI weapons, and there’s no death, and it’s bloodless.

So, there’s a couple problems with this. This is not warfare as we know it. We can see this in Ukraine. We can see this in the Middle East. The idea of a bloodless war just doesn’t happen.

And also, this goes to the fundamental contradiction of Palantir. So, essentially, what it is providing is an enterprise software system that hosts other companies’ LLMs, their models. They’re not building AI weapons. They’re not building anything. They are essentially a software company. And they have gotten so much mileage out of the narrative that Alex Karp talks about of defending Western civilization, that it’s almost hard to look at the fundamentals of the company. And the fundamentals of the company are that they are growing, they are making inroads into the Pentagon, but they have a fraction of the revenues of other defense companies, of defense companies that are building big weapon systems.

And what’s interesting about Palantir is they have a very high proportion of institutional investors. These are retail investors who often — you know, these are often young men who like the message of Palantir. They like the — Alex Karp’s message, the sort of Lord of the Rings defending the West. They often aren’t looking at the business fundamentals. And I think this is why you have seen over the years a lot of people saying this company is vastly overvalued.

AMY GOODMAN: I want to go to Secretary of Defense Pete Hegseth speaking at Elon Musk’s SpaceX facility in Brownsville, Texas, back in January.

DEFENSE SECRETARY PETE HEGSETH: Gone are the days of equitable AI and other DEI and social justice infusions that constrain and confuse our employment of this technology. Effective immediately, responsible AI at the War Department means objectively truthful AI capabilities employed securely and within the laws governing the activities of the department. We will not employ AI models that won’t allow you to fight wars.

AMY GOODMAN: So, if you can talk about the significance of what Hegseth is saying, the fact that he’s saying it at Elon Musk’s SpaceX in Brownsville? Elon Musk, a member of what’s called the PayPal Mafia, the former executives of the online payment firm that went on to create other companies. They sold their startup to eBay and then went on to start SpaceX as part of his dream to take humanity to Mars, but hoping that working for the Pentagon could help pay for his Mars plan, as you reveal in this book, Valley of Death, Sharon.

SHARON WEINBERGER: So, that was a message. I mean, very significant that it was delivered at SpaceX, but the message was actually squarely aimed at Anthropic. And so, what was going on at the time was at the beginning of the year — so, initially, Anthropic was the only frontier AI company whose models were able to be deployed on classified networks in the Pentagon. So Anthropic was in a really good position. So, after the Maduro raid, Anthropic raised questions with Palantir, because Palantir was this enterprise software that the model was deployed on, and expressed concerns about how their models were used in the raid. And the reason why was Anthropic had two red lines. One was that their models couldn’t be used for mass surveillance of Americans, and, second, that they couldn’t be used for fully autonomous weapons.

Well, Palantir took those concerns and went straight to the Pentagon. The Pentagon leadership went berserk and basically said, “Who is this private company to raise these issues? And, you know, oh, as well, like, we’re not using it for mass surveillance of Americans. We’re not using it right now for fully autonomous weapons.” But it created this huge conflict between Anthropic and the Pentagon. What happened at the end was the Pentagon dropped Anthropic, labeled them a supply chain risk, to try to get them totally kicked out of any government work, and then made deals with the other AI companies, to include Elon Musk’s xAI, to deploy those models on classified networks.

And what’s really interesting about that conflict is, I take away a slightly different lesson from it than I think others do. On one hand, it sounds great that Anthropic had these high ideals, that it had these red lines for its models. But I think what we’re seeing is a real struggle between government, which doesn’t — they can’t just order companies to do things. They can’t with Elon Musk. They can’t with Anthropic. And so, although I don’t think the Department of War, as they call themselves, quite took the right position by designating Anthropic as a supply chain risk, I suspect that they’ll probably lose that battle in court eventually. Maybe not. But really, even if you agree with Dario Amodei and Anthropic, what happens if the next CEO, like Elon Musk, has positions you don’t agree with? Do you really want to be in this position of the government having to ask tech companies and their CEO for permissions to do certain things? So, however that conflict resolves — I don’t think it’ll resolve. I think we’ll increasingly see these conflicts between the Pentagon and between these tech companies over their technology and who controls it.

AMY GOODMAN: Dario Amodei did not attend last week’s state dinner with the Chinese President Xi Jinping, while the other major tech CEOs did. But apparently, last night, Amodei and Trump had dinner together at the White House. Can you talk about the significance of this?

SHARON WEINBERGER: So, I think the significance of it is people can talk a big talk, but this speaks to two things. Trump doing that dinner, doing these meetings, because they really need Anthropic — Anthropic still remains, by most accounts, ahead of the other frontier AI companies in what they’re offering. The government needs it. It also speaks to the fact that Amodei and Anthropic, they need the government. They want these contracts. They want to work with the Pentagon. This is a lot, on the Anthropic side, about profit motive. And they want to be — because they’re making this argument that we’re in an arms race with China, they need to be part of that race. Both sides — I hate to be cynical, but both sides have a motive for making some sort of peace.

AMY GOODMAN: Can you talk about Peter Thiel, the billionaire co-founder of PayPal, first outside investor in Facebook, launching the company, working on data mining software for the CIA and the Pentagon, as you write about in the prologue of Valley of Death?

SHARON WEINBERGER: Sure. So, Peter Thiel is a really interesting person who, I sometimes wonder, may be going off the deep end now. He saw — so, he was part of this PayPal Mafia, together with Elon Musk. Musk, after the PayPal sale, he decided to go to space, and Peter Thiel went a different direction. They took an idea that arose out of PayPal, which was fraud detection, using algorithms to try to detect fraud. And after 9/11, Peter Thiel saw a real potential market with government: What if we use these same algorithms to try to spot terrorists? They really didn’t have much technology at the time. It was more just sort of a visual sort of a user interface. But they pitched this around Washington, and they eventually got money from In-Q-Tel, which is the venture arm of the CIA, which both gave them some money, but also it kind of gave them an entrée into the national security world. It took them quite a while to get government contracts, but they did. And, you know, I think there’s a number of reasons.

Peter Thiel tends to be a contrarian. He’s been described, you know, as a conservative, as right-wing, as libertarian. I think he has been all of those things. But I think he has changed a lot over the years. If you look at his early statements about PayPal, he talked about being able to have financial transactions beyond the control of the state. And then he goes on to form a software company to basically do the exact opposite, to allow the state to monitor transactions to try to find patterns among people to spot potential terrorists. So, I think he’s a lot more ideologically flexible than perhaps he thinks.

He has become even more controversial today as he kind of pairs his business interests with sometimes very bizarre philosophical ideas. These days, he’s giving lectures on the Antichrist, which he thinks will come in the form of sort of tech and left-wing figures. I always joke that if you read Peter Thiel’s book Zero to One about entrepreneurship, it’s actually quite — you can agree or disagree with it, but it’s quite good. It’s quite clear. It’s well written. Peter Thiel, like a lot of these tech billionaires, wants to be a philosopher. I don’t think his philosophical writing is all that good or all that clear.

AMY GOODMAN: I’m wondering if you can talk more about SpaceX, Palantir collaborating with actually shaping U.S. national defense. As you write, they’re providing spies and soldiers with AI and cloud computing tools, as well as actual military hardware, such as microwave weapons, hypersonic missiles and advanced drones.

SHARON WEINBERGER: So, this is something that’s actually, I think, becoming a bit of a tension with the Trump administration, which is fascinating to watch. So, the argument of Silicon Valley, and actually one of the reasons that they moved into this particular area — so, if you look back at the start, the real sort of flow of venture capital money into defense, which started to really get going in 2016, 2017, that was around the same time that Silicon Valley was finding the China market. You know, what was their big push in the early aughts? Facebook, Uber, all of these tech companies looked at that consumer market in China, and they thought that was going to be huge. By 2016, ’17, most of the tech companies were realizing they were not going to make a lot of money in China. China was shutting the door to American tech companies.

And at the same time, the Pentagon had been, for several years then, saying that China was the next big strategic threat. And Silicon Valley saw a business opportunity, particularly for a potential war with China. What Silicon Valley has to offer — AI, AI-driven weapons — was very much part of what you would want for a battle with China. And so, what you started to see was suddenly Silicon Valley was all talking about the war in China. When I visited Anduril’s headquarters in California, Palmer Luckey told me that they — a lot of their teams at the company, they would tell them, “If the weapon won’t be ready for a 2027 war with China, we’re not going to work on it.”

Well, the funny thing is, is the Trump administration actually hasn’t been as focused on China as a military threat as you would think. Trump is trying to make these trade deals with President Xi. There was even talk from the U.S. ambassador that Trump talked about selling weapons to China. Kind of strange stuff. So, Silicon Valley is sort of trying to repivot: “Like, what do we do? We pitched this war with China.” But they have the same motives as the defense industry of old. The problem is, if you are financially motivated to build AI-powered weapons, then you’re kind of looking for conflicts where those can be used. You can use them in other conflicts, but China is kind of the number one.

AMY GOODMAN: I mean, it’s so interesting how you address the issue of China becoming the argument that ends almost every debate, whether it’s autonomous weapons, drones or AI: “If we don’t build it, China will.” President Trump repeatedly talks about beating China. And so, the question is: How will the Congress, how will governments regulate AI when you have this constant threat? And talk about the power of these AI companies, I mean, the image of the tech billionaire CEOs addressing the U.N. national — the United Nations Security Council last week. Their budgets are often larger than countries. They are as if they are the heads of nation-states.

SHARON WEINBERGER: It’s crazy. And I’m not saying that AI companies or their leadership have nothing to contribute to the discussion. But again, can you imagine the head of Northrop Grumman or Lockheed Martin or Boeing addressing the U.N. on the safety risks of hypersonic weapons? It just wouldn’t happen. And it speaks to this tremendous power and the money that these AI companies have.

I am, sadly, a little bit pessimistic about what’s going on. The AI companies are pushing to self-regulate. I don’t know where congressional oversight is in all this. I don’t remember the last time I’ve seen a real congressional hearing trying to dig into defense issues. And there are so many issues to dig into now, like the use of AI in military targeting. I don’t see the expertise that we need. There are some good people inside the Pentagon, but the National Security Council and the AI expertise you would expect to see resident there just isn’t there.

One of the problems that you have is pretty much any AI expert that you have in government will be looking to their next job in Anthropic, in OpenAI. I think it would take a real push by government to say we need AI experts, we need them in the National Security Council, we need congressional staffers who are expert in this, and then we have to have revolving-door and conflict-of-interest rules, so these people can’t go and turn around and work for the AI companies.

There is going to be some sort of regulation of AI. I mean, we’re already seeing that. There are export control regulations. And the only question is: Who do you want to set the rules — the AI company themselves and the people affiliated with them, or people in government who have the best interests of the public at heart?

AMY GOODMAN: Sharon Weinberger, you have written for The New York Times. You’ve written for The Wall Street Journal, that latest piece you did, “The Hidden Debt That Apple Owes to the CIA.” But now you are editor-in-chief of the Bulletin of Atomic Scientists, founded in 1945 by former scientists involved with the Manhattan Project. Right? The Manhattan Project, that secret Pentagon project to build atomic weapons, which it succeeded in doing. Can you talk about how artificial intelligence, how AI, is already impacting nuclear weapons decision-making? Could AI make a nuclear war more likely?

SHARON WEINBERGER: So, I think we’re at the beginning stages both of trying to understand that. We’re already seeing AI being used in targeting in military operations. There was a very good Bloomberg investigation that came out recently about the — about how analysts sort of overrelied on the Palantir smart system on which AI is deployed to come up with targets, and that contributed to the targeting of the children’s school in Iran.

So, in nuclear command and control, you’re now seeing some of the experts in the area starting to look at how it’s being used. I certainly don’t think — we’re not at the point where AI is, you know, going to autonomously launch a nuclear war. But to the extent that AI is being integrated into targeting, it will inevitably be integrated into the way nuclear weapons — into the planning, and likely, eventually, into the use — again, not that they’re going to be used to launch nuclear weapons.

We’re only at the very starting point of that debate and trying to understand that, which is why it is so critical to have some sort of congressional oversight, because what you’re seeing now is outside analysts raise the questions, but you actually need answers from government and from the Pentagon on how they plan, how they are planning, to use or using AI in nuclear command and control.

The Bulletin of the Atomic Scientists is completely about existential risks, things that could lead to the possible end of civilization, to the end of humanity. And AI is a part of it, not just, I mean, AI itself, but, as you mentioned, how AI could be incorporated into nuclear war.

AMY GOODMAN: Finally, you’ve written this new book, Valley of Death: How Big Tech Is Fueling the Future of War. What surprised you most, or, should I say, what frightened you most?

SHARON WEINBERGER: I think what surprised me most was the story of Steve Jobs and the CIA, because I really thought it was, you know, maybe something of a D.C. urban myth at the start, or an apocryphal story. And the fact that it ended up being true, and that, in fact, people said it did in many ways save his career, enabled him to leapfrog back to Apple, that was surprising, because I think we sort of accept some of the self-mythologizing that Silicon Valley does about itself, about the counterculture. There was a real counterculture in Silicon Valley. I don’t doubt that. But that doesn’t mean they weren’t willing to work with the military or the CIA. So, that was the most surprising thing.

The most dangerous thing was, again, that idea that — the study that was done in 2013 about, you know, a billionaire building nuclear weapons, where, you know, when I interviewed some of the participants, some of them said, like, “Yeah, you know, we worked on it, but we thought maybe the premise was far-fetched.” And here we are in 2026, and Palmer Luckey, the head of one of the most prominent defense tech startups in the world, is saying, “I would build nuclear weapons. I would build fusion weapons. I think they’re a stabilizing force.” I think that’s frightening.

AMY GOODMAN: Well, I want to thank you, Sharon Weinberger, for joining us. Sharon Weinberger is a longtime national security reporter and editor, previously at The Wall Street Journal, currently editor-in-chief of the Bulletin of the Atomic Scientists. Her new book is titled Valley of Death: How Big Tech Is Fueling the Future of War. To see Part 1 of our discussion, go to democracynow.org. I’m Amy Goodman. Thanks so much for joining us.