The Washington PostDemocracy Dies in Darkness

Transcript: The Path Forward: COVID-19 with Palantir CEO Alexander Karp

MR. IGNATIUS: Welcome to Washington Post Live. I’m David Ignatius, a columnist for The Post. This morning we’re joined by Alexander Karp, chief executive of Palantir Technologies, a leading data analytics company in Silicon Valley. We’re going to be talking about Palantir’s role in trying to work with the terrible problem of COVID-19 and other issues involving data analytics. But, Alex, first I want to ask you to begin by talking about the news that’s on every viewer’s mind, and that’s the storming of the U.S. Capitol building yesterday. What are your own thoughts?

MR. KARP: Well, you know, we began to build Palantir almost 17 years ago with a focus on strengthening U.S. and Western institutions to support democracy and make all of our lives better. So, by reducing terrorism and increasing civil liberties. And you know, when I see what happened yesterday--and honestly, when I watch the inflammatory tone of some of our elected leaders that led to this, it's quite concerning--I would say also depressing--as a citizen of the U.S. but also as someone who's been pretty actively involved in national security and some of the more important parts of protecting democracy. It severely corrodes trust in the government. It damages the view that our allies and very competent non-allies have of the promise of American democracy. And I really hope that all leaders of all parties get their act together. And there's a disobedience--disagreement is crucial to democracy. It's what makes living in a democracy so special. But there's a line where we begin betrayals that corrode the very institutions that preserve this discourse, and that line was crossed. And those of us who played a crucial role in enforcement of law especially need to speak up. And it's also something, by the way, you know, if--there's a lot that America has to do to bolster its defenses, but if we're going to have these kinds of events, it's very hard to have strong national security because the internal structures just won't support it, and the charisma of the American project will be tarnished and by--was tarnished by these events.

MR. IGNATIUS: That's a good description of the national security costs that we paid for what happened yesterday. Alex, you've been a critic in many instances of Silicon Valley. It was interesting that the big social media companies yesterday, Twitter and Facebook, essentially withdrew their platforms from Donald Trump, at least temporarily, thinking that it just wasn't appropriate to make them available. What's your own thought about that? Should social media companies be allowing the dissemination of what the courts have found are false information by the president, or should we say that's enough, no more?

MR. KARP: Well, you know, in general, first of all, Silicon Valley's in a difficult position because there were--there's really been two phases of Silicon Valley, the first phase which helped us win World War II that led to our increased freedom and liberty, technology, dual-purpose technologies primarily built for the military, repurposed for commercial utilization. And that Silicon Valley bolstered our nation.

I happen to believe the president crossed a really important line in this case in that we all have to take--do what we need to do. And we at Palantir certainly would not support any actions of this kind. The problem is, Silicon Valley has become so divisive, I worry that any action that they take might actually end up bolstering these--this kind of transgressionary behavior. However, you know, what we saw yesterday, just--it's a breaking point. And for once--you know, I've been very critical of Silicon Valley. As you know, we moved our office to Denver. I would say, for once, I'm largely in support of Silicon Valley's actions.

MR. IGNATIUS: That's fascinating given the history--and we'll get into it--of your recent criticism of Silicon Valley. I want to ask you just one more thing, Alex, before we turn to Palantir and the specifics of what it does. In the aftermath of the seizure--storming and seizure for a time--of the Capitol--the business roundtable--a very distinguished group of big American companies--some of the top CEOs, Brad Smith of Microsoft, others--have issued very strong statements, the Business Roundtable essentially condemn what it called unlawful actions to challenge our political transition. Would you endorse those kinds of statements by the business leadership and join in a general business insistence that we now proceed in an orderly transition of power?

MR. KARP: Absolutely. And I would--I think--I mean, look. There's the business leaders as important parts of our society. There's political leaders which is important parts of our society. There's spiritual leaders as important parts of our society. There are academic leaders as important parts of our society. And there are national security experts and former heads of services. And I think we're at a point with what happened where really everyone involved in a leadership position needs to articulate what their position is. I believe that--yeah, that we as business leaders, but also more importantly just as people that may have influence because of accomplishments or who we know, take a--take a really direct open and firm stand. And I'm hopeful that if we're, you know, unified that will have a pretty significant impact on political leaders.

You know, I think just as a larger issue--and it's probably outside the bounds of what I should say--but businesses like--there's a certain modicum of adult leadership necessary to run a significant organization of any kind, whether it's a university, a newspaper, a church, a synagogue, a mosque. This is below that line, and we shouldn't tolerate it.

MR. IGNATIUS: That is a strong statement, and in my view, not at all outside the line. Thank you for sharing it. Let's talk a little bit about Palantir and what it does. You often like to stress what it doesn't do. It doesn't own data. It's not in the business of sharing data. But it is in the business of analyzing big data. And maybe you could just explain to us what this technology that Palantir applies to data is all about. When somebody retains Palantir, becomes one of your customers, what are they buying from Palantir?

MR. KARP: Well, they're--what they're buying--there's what they're buying and what they're getting. What they're in the end buying is a result, based on their data, so that they can make a decision, not that we can make a decision. And how they make the decision comes down to their level of technical acuity. Some people use ML at the final mile. Some people use data analytics. But what they're actually buying is a platform built years before it was needed that would allow you to replace or in some cases replace parts of your platform that would--you would need to answer seemingly simple questions.

So as an example, if you have hundreds and hundreds of data sets, and you have to ask a simple question of how many people can fit into this hospital in this neighborhood, you've got to integrate all those data sets before you can answer--ask and answer that question. And then you have to have a series of software products. And you may have three of the software products, but you don't have 14 of them. And so, what Palantir does is we essentially collapse the time needed from the time you realize that your infrastructure is not working to the time you can ask and answer the question. And we did that by essentially predicting that times before--five or six years out before we built the product would be very different than when we built the product. And so, our product is used to maintain airplanes, to help people drill for oil and gas. In the national security context, it powers most of the police intelligence work in Europe and because of its data protection. But what you're really getting is an ability to replace or augment the software products you have overnight or within the context of a couple weeks so that you can ask and answer context in--questions in a critical context, either in the commercial or the government context. And that's how we ended up being the engine behind most anti-terror institutions in the West, a number of the largest economic institutions in the West and why Palantir's in 35 countries helping to mitigate the disaster incurred in the wake of the COVID pandemic.

MR. IGNATIUS: And before we ask you about how--why this Palantir approach to understanding data to current healthcare problems, I want to ask about Palantir's origin story, which as you said, is in the counterterrorism CT area. It's said sometimes about Palantir in print that your technology was useful to the Obama administration and to the CIA under Leon Panetta in the hunt for Osama bin Laden. Let me just ask you, was that true?

MR. KARP: Well, you know, Palantir--I mean, it's one of those questions, obviously, when you run a company like mine, you're never allowed to answer. But what I can tell you is, you know, Palantir stopped, disrupted some of the largest terror attacks you've never heard of. And many of the anti-terror events you read about are, in fact, the result not of Palantir but of the incredible people that use Palantir all over the world.

MR. IGNATIUS: All right. We'll stick with that as an answer. Let's turn to the COVID nightmare, which is something causing death and destruction around the world today, and the way in which Palantir is working with the Department of Health and Human Services, I think also with Operation Warp Speed, to try to help get better analytics. Alex, as we look at the development of vaccines, we're all just extraordinarily pleased by how fast that happened. But as we look at the distribution process seeming to break down, we throw up our hands again. What would Palantir be able to do to help us, help HHS in this period going forward, distributing vaccines, getting the country healthy again, preparing for the next pandemic?

MR. KARP: Well, Palantir is involved globally in--and the two places we're most prominently involved are the U.S. and UK, but a number of other countries--in allowing people who are doing their planning to understand exactly what's happening on the ground, distribute PPE. So, for example, which hospitals are over utilized, which hospitals have capacity, which hospitals need PPE, which ones have too much, which ones have too little. And so--and that's--and then on vaccine distribution, most notably in the U.S. and UK--the Operation Ward Speed, which uses Tiberius built on top of Foundry uses--

MR. IGNATIUS: Foundry is your basic Palantir infrastructure, so our viewers will understand that.

MR. KARP: Sorry. And what Foundry allows them to see and decide is they--it's essentially a dashboard unifying the hundreds of data sets that make up the production of vaccines and the distribution of vaccines up to the state level. So how many vaccines are produced, how many can be shipped, and where do they go.

The central issue in America--less so in England because they have a centralized healthcare system--is the delta between vaccines that are delivered to the state and the actual doses that arrive in people's arms. And so, we have a serious issue with that in America--as I think is commonly known and was discussed by your guest yesterday--that there's roughly a 12 million delta between the vaccines and roughly 17 million vaccines delivered to states, roughly 5.4 administered. Palantir brings those vaccines to the states, and then there's--the question is what we as a society can do to resolve that problem.

Now, you know, one of the things that makes America particularly interesting and complicated and great is that states have a lot of rights and they use local--they use local ways of distributing, and each one has a different infrastructure. What I'm hopeful will happen, especially once we get past this incredibly discouraging and aggressive events of the last couple days, is the new administration will also be in a position to have better discourse with states. States will begin to develop best practices. That's what software could do, and what we obviously would be willing to do, is to help them administer the policies they develop more accurately, quicker.

One of the things that's super interesting about software is it allows you to compress time. So, you know, we--what should not be possible, distributing--we need to get roughly 300 million vaccines to America to get herd immunity--is something that will only happen if we unify as a country and we use the best technology, especially software, available, because the amount of vaccine per day is just--surpasses what we would normally be doing as a country. The British distribution has certain attributes we don't have because they're just more centralized. And so, they're--but in any case, what we're doing is providing a dashboard so that every single person involved in Operation Warp Speed can actually track distribution, how much is being produced, where is it going, and if there's any disruptions between manufacturing and final handoff to the state.

MR. IGNATIUS: So, let me ask you a question that goes to the heart of what we're talking about here with vaccines, ask it to you as a data analyst, not in a political sense. Is it going to be possible to have an orderly system for distribution of this vaccine or future vaccines and future healthcare crises, if we don't have an America, a somewhat more centralized system so that we just have better, cleaner data and we know what's going on?

MR. KARP: Well, you know, another way of answering your question is to posit--which I would--which I think is unfortunately obviously--this will not be the last pandemic. And so, you know, one of the things that's made American software companies dominant in the world is that we look at the mistakes that we've made, that we're making. We try to figure out what they were. We're fairly non-judgmental about them, and we iterate so that the next time around we're dramatically more effective.

Assuming that this will happen again, there's a whole series of things we need to put in place. I mean, just banalities we all know, but they're so large we forget. There's this--right now 361-some Americans have died. That number within the next couple of weeks is going to surpass the number of people that died in World War II. This is already a hundred times larger in terms of human life than what was inflicted upon us on 9/11. This is something that we in the West cannot afford to address in a way that's better than this time.

And so, what would need to happen? Now, there are lessons both in building software companies and doing cybersecurity that are quite helpful. First of all, you need an unbroken chain between attribution. Where did the pandemic begin, how did it begin, under what conditions did it begin? That data has to be something that is shared freely and examined internationally, not just by America, both so that we understand it and it has credibility. That data has to be immediately shared with vaccine producers. We probably need to figure out a way to use the most advanced computational technologies we have, the most advanced software technologies to accelerate vaccine production.

And approval--approval, as you know, is a super important issue in America. And broadly in the West, you have in all Western countries something like 30 percent of the population unwilling to take a vaccine. That's not just an American issue by the way. You have roughly the same numbers in France. Higher in France, roughly the same in Germany and in England. And to get herd immunity, you need to get probably to 80 percent. So, we have a legitimation crisis on vaccines that is not primarily technology. And it can only, in my view, be solved by providing transparency in the actual approval and after it has been administered, because otherwise we [unclear].

Another super hard thing that we have to look at is, we do not want to pretend or obviously attempt to change the basic way in which we live in order to get pandemics under control. And that makes vaccines even more important, because if you're going to expand civil liberties and fight pandemics, you're going to have to do that with vaccine and increased legitimacy. Software, the way in which vaccines are developed, and honestly, more credible political leadership can actually make a huge difference in that problem. But you really need all three, because if the vaccine is delivered too slowly, then you have--you have compounding math working against you. If the vaccine is delivered on time but people don't believe they will take it, you can't get to herd immunity. And then in democracies, we need to convince people, not force people.

And so those--by the way, the element of--obviously what--analytics, data integration, ML, AI will play a role in that whole chain. But they play a role primarily in two ways. One, acceleration of time, and two, creating transparency. By the way, transparency is a specialized issue, specialized concept in software, because you both create transparency of what's happening in the data, but you can anonymize the people involved in acquiring the data. And so that's obviously necessary in correspondence with our generally held civil liberties.

MR. IGNATIUS: Let's turn to an issue that has been a central one for you in an ongoing debate you've had, really, with your colleagues in Silicon Valley, and that's your feeling that in terms of values, there's a kind of breakdown between what you see in the Valley and what you think is appropriate. In your S-1 filing that accompanied some of the paperwork associated with going public, you said, "We seem to share fewer and fewer of the technology sector's values and commitments." Then you went on to say, "We've chosen sides, and we know that our partners value our commitments." And by "chosen sides," you mean you've really chosen sides with United States, your country and with Western democracies as a group. Is--do you feel that other Silicon Valley companies have failed to do that, and what advice would you give to the Google employee who was uncomfortable working on Project Maven or some other Pentagon project, to try to convince them that that's a mistake?

MR. KARP: There's a--there's a series of super important questions there. Just starting from the beginning, this country and a lot of the norms that we cherish was built on economic prosperity in part powered by our ability to develop technology that was useful for individuals, not just useful in making a couple of--a small number of people on the West Coast rich. And so, you have what I would call Silicon Valley one, which was the Manhattan Project, a lot of the technologies that are even powering this call. So, computer technologies, internet technologies. Those were built by the military and shared with humanity, both for America and the world's good.

Consumer internet--you can debate it back and forth--it's very hard to define what utility it has. And I do, in part, blame it for creating an enormously tribalistic America, where people are ever more convinced that their ever-narrower opinion is the only one that matters. And that combined with I think great economic disparity in part because we're not investing enough in technology, I think is responsible for some of the problems we have today, not all of them, and in no way excusing the current behavior of the last couple days.

The--so I do think Silicon Valley has to ask very simple questions. That--I don't think everyone has to agree on the answer, but like what is your value proposition, actual value proposition to the average person in America and among our allies in the world.

Then you asked another set of questions, which is a super subtle question. And we had the same question at Palantir, but an employee for whatever reason doesn't feel that their valuable effort should support the military. And that's a completely legitimate opinion that should be part of the discourse. I would only say to that employee, as importantly to the leaders of these companies, your company is worth hundreds of billions of dollars in part because the U.S. government and the West protect your right to engage in international economic behavior. You live in a system, you drive streets--on our streets, and we are prospering disproportionately. And there's--you know, you can't live in a country and not support it.

Last, not least, I would point out to that employee that--and again, without being--you know, there's no--you can't force someone to work against their conscience, but you can remind them America has very serious, very, very competent adversaries that have a very different view of how the world should work and very different views of human rights and economic justice than most people in America. And we are living in a world of complexity where we may need to defend ourself.

Last, not least, I would try to engage in a discourse around what do you think the consequences are going to be? Technologists realizes that technology is nonlinear, so that it's much more--AI is much more like--which Project Maven is working on--is much more like a nuclear bomb than a machine gun. And if one side has a gun and the other side has a nuclear bomb, the side that has a bomb will often decide that their way of seeing the world is the way of seeing the world. And I don't believe America's perfect. I've written a lot of critical things about the West and America in my former years as an academic, and I still believe some of those critical things. But I believe that a world order that is defined by, broadly speaking, Western democratic norms, the rule of law, fairness, separation of church and state, a general concept that things will be adjudicated in a court that will be followed, is a better system than a system that doesn't have those things. And a country which does not have the equivalent of software nuclear warfare because its best people won't work on it is a country that will not be able to have--engage in dialogue of equals on these very important issues.

Last, not least, what I would say to the head of Google or whomever I was talking to is, you know that America has the best software production in the world. You also know that if we do not work on it, the people who are the very best, that countries that are not as good at this will win. We have the best people and the best companies. It does not mean we will have the best military technology if none of us work on it.

And by the way, it's worthy of remembering this is not even an option in any other country. And not just in America. No French company can opt not to work with the French government. It's not--and there are reasons for this. And I think one of the things--and as some of your viewers know, I gave to Biden. And one of the things I'm hopeful about is that it'll be a less politicized environment. And a lot of these problems are actually only solved where people--a wider swath of people are willing to actually say we don't agree on some issues, but we do agree on these issues and we're actually going to put our technology, our spirit, our efforts together and actually do something that can work. And in the national security context, especially technologists who know this is not--this is a nonlinear situation, need to be involved, because they understand this is the whole reason their companies are dominant, because you can't compete with a world-class software company if you're not world-class.

MR. IGNATIUS: So, Alex Karp, I want to thank you for the broadest-ranging conversation on subjects that really matter. A lot of what we were just talking about now goes to the question of the United States and China. We'll ask you to come back I hope before long and talk about how the Biden administration is doing with some of these technology issues. But thank you for joining us today on Washington Post Live.

MR. KARP: Thank you for taking the time.

MR. IGNATIUS: This afternoon at 3:00 my colleague Jonathan Capehart will discuss the storming of the Capitol a day ago. His guests include Congressman Ro Khanna and former Senator Jeff Flake. So, thanks to everybody for joining us, and we'll see you later today on Washington Post Live.

[End recorded session.]