Transcript
A (0:00)
Foreign.
B (0:05)
Welcome back, everybody, to what really Matters. I'm Jeremy Stern with you in Los Angeles. I'm here as always with Walter Russell Mead of tablet, the Wall Street Journal, Hudson Institute and the Hamilton School at the University of Florida. We're going to talk a lot about the unfolding operation in Iran in the big conversations today. But first, let's start with this week's news items, all of which are tied to the larger Iran story, but we'll borrow from some of its interesting component parts first. So, first story of the week. A week long clash between Anthropic and the Pentagon came to a head this week when CEO Dario Amadei refused to loosen safety guardrails, preventing Claude's use for mass domestic surveillance or fully autonomous weapons. As a result, President Trump ordered every federal agency to immediately seize using anthropics technology. And Defense Secretary Pete Hegseth designated the company a quote, supply chain risk, a designation typically reserved for US Adversaries, effectively forcing military contractors to avoid Anthropic as well. With a six month phase out of its military applications. Within hours, anthropic's chief rival, OpenAI, announced its own Pentagon deal to deploy its models and classified systems. The dispute has since grown. Messier Amadei sent an internal memo calling Trump's attitude toward Anthropic a demand for, quote, dictator style praise and dismissing OpenAI's Pentagon arrangement as safety theater, which an administration official said threatened any path to reconciliation. As of today, Thursday, March 5, however, Amadei has resumed discussions with the Pentagon. Claude, crucially, is the AI model powering Maven, the data and intelligence system that is behind the current US Campaign in Iran. Walter, is this episode news or Faux News?
A (1:41)
Well, first, let's just note that if you had read those paragraphs out five years ago, no one on the planet would have had any idea what you
B (1:49)
were talking about and any part of it.
A (1:52)
We are in a brave new world and we're having fights about things that didn't exist a very few years ago. I think it is news. I mean, I'm glad that they're still talking because I think the answer is to figure out some way for the Defense Department, excuse me, the Department of War, to be able to use critical technologies that might save our servicemen's lives. We are actually fighting. Let me just say that I for one would approve of that course of action. The deeper issue here is this kind of the way that information and the processing of information is, it has become linked to the strength, security and even existence of the state in that sense, for over 300 years, money has been, and the banking system have been not necessarily part of the state, but have had a very close relationship with it. And central banks, which buy and sell government debt, which do all kinds of things that keep the financial system going. You know, they are, they're independent. At least most of them are independent of the state. Some in former times they were actually often private companies. But the government, the state wouldn't have the power to fight war or to pay its bills if it didn't exist. And so money is both something that ordinary people use in their lives and something that is essential, is almost a constitutive element of government. Well, we're getting to the point where all information is going to be like that. And so information companies are already beginning to have sort of the same kind of importance to government that banks and the financial system have long done. Who knows, in some years we may be talking about the info system or something of that kind and who knows, the central info system, the central info bank or whatever we call it, by the way, I would love to really like get some, you know, angel investment capital and whatever company, you know, that turned into. But, and this is going to cause immense problems up and down the line. And this one of, gosh, you're going to use our poor, innocent, lovey dovey software to kill people overseas. All right, that's, you know, but also that information capability, if you have enough information to defend your country from the bad guys overseas, you also have the information capability to monitor and surveil everybody in the United States. And so does that mean that we're going to be, you know, that our choice is to either just get conquered by China and let the Chinese government have control over the, you know, total control of everybody's information, or do we actually have to build our own monstrous despotism to hold them off? Hopefully not. Let me say we will need a, an infosystem that is as strong and capable as anything that, that an authoritarian government can produce. But somehow we do need guardrails. We, we do need ways to protect our democracy. This is not a new problem for us in the sense that if the government has, has enough force to repel foreign invaders, it also has enough force to quell domestic rebels. So technically this is not, it's neither an insoluble problem nor a completely new one, but it is a really thorny, intricate, complicated one. I would say that right now what we see both on the side of the company and the side of the government is that neither One really has the institutional tools, the legal framework, or the experience to know how these lines should best be drawn. And I expect we're going to see a lot more kind of explosions along these lines. There's another little element in it, which is that the people who work in these companies and who do this design would like some say in how their work product is used. You know, I'm just here. Well, I'm not coding anymore because Claude has taken that part of the job away. But I'm doing whatever is left once all the coding has been outsourced. And I'm pretty sharp, and I'm making a lot of money, and nobody else really can do exactly what I do. And I want to be able to say to the government, you can't make me make something that you then use to destroy my liberties. And I get that. But how much power do unelected employees of tech companies deserve when it comes to national security? Are they really the ones who should get to say, based on whatever their own feelings are at any moment, how this technology can be used? So there's a struggle for power within the companies as well as between the companies and the government. I don't know how it all gets worked out. Everything I'm saying is that this is a big problem. It'll appear in many different aspects over the next few years. And all you can advise folks on both sides is to try to behave with a little bit of wisdom and restraint and humility, realizing that we don't actually have all the answers, even though we do need to make some decisions.
