Speaker 1: Tech stocks selling off today on the emergence of China's new AI app DeepSeek and a new model it made. As news ripples through the markets, its impact could have even bigger implications for AI governance in the US. For more, we're bringing in Matt Calkins, CEO and founder of Appian. Matt, it's good to see you. And I thought of you when I was looking at this tweet from David Sachs, who, of course, is the AI czar and crypto czar for the White House. He said DeepSeek's R1, that's the new model, shows that the AI race will be very competitive. He also said President Trump was right to rescind the Biden executive order, which he said hamstrung American AI companies without asking whether China would do the same. I'm curious what you make of that view and sort of taking the guardrails off, if you will, for AI firms in the United States.

Speaker 2: All right. The war for AI is going to be an asymmetric war between the US and China and anybody else who gets involved. And the things that make us strong are not the same as what's going to make China strong. So we shouldn't worry about trying to match exactly the way they approach this. And I think we have a free hand to make policy. I do believe that the executive order that was rescinded recently, we're better without it. It was overhead. It was bureaucracy. I think, however, we would be wise to create a national regulatory framework and protect the rights of the privacy rights and copyright information rights of Americans. I think those would be a very American way to create a framework for succeeding in AI.

Speaker 3: And talk a little bit more about that, Matt. I put you in charge now. Talk more about that framework. How would you flesh it out? What would be the details of that?

Speaker 2: Yeah, first of all, total transparency. We got to understand AI in order to regulate it properly. So it's important that we all understand whose data is training these models and what data is being used. Once we understand that, we'll be able to make better, more enlightened laws about regulating AI. But we've got to understand, first of all, where the value comes from. This deep-seek announcement is a wake-up call for everybody who thought they knew where the value came from. We thought it was the chips. Well, you know, we need to think about that again. We thought it was the top-level expertise. We need to think about that again. I'm not exactly sure, but in all, it reminds me very much of open AI. A couple of years ago, we would not have guessed that open AI would explode onto the scene as an AI leader. We thought it would be Google. We thought it would be somebody who'd invested more money. It turns out that open AI and deep-seek are both capable of surprising us. And I would expect a bunch more surprises to come in an industry as volatile and innovative as AI. What we need to do is create just a clean framework, a national framework for laws. Don't do this at the state level. Don't put unnecessary overhead on our innovative companies and still protect individual privacy rights. That's the way to do it. And I think David Sachs is on the right track.

Speaker 1: And do you think that, if that's the case, and even without that framework, do you think deep-seek being Chinese is a competitive disadvantage? Do you think that there are companies who will say, I'm not going to use this model because I am concerned about what China, not just this company, but that the government in China might have access to my information?

Speaker 2: I think deep-seek is more important as a demonstration of possibility than it is as a product right now. As a product, I think a lot of people would have reservations about using it because of the Chinese ties. And there's some censorship mixed up there that we could probably get rid of. But the most important thing here is just the possibility that if you could create an AI model, a winning, strong AI model with fewer chips and less money and no big name experts, that's an exhilarating possibility. And so the deep-seek shockwave that's hitting the US this week is the realization that there may be another unanticipated way to create great AI models.

Speaker 3: Matt, I'm just curious whether you downloaded and experimented with the app. And if you did, Matt, are there ways that you do so smartly, prudently, responsibly, given that it is a Chinese company?

Speaker 2: Yeah, you know, I don't download Chinese AI. I just don't. And I would draw the line there. I read a lot about it, but I would choose not to disclose any information.

Speaker 1: But Matt, just to build on what you're talking about, is this sort of a clarion call for US developers that they should be trying to do things more efficiently, at lower cost, et cetera?

Speaker 2: AI is an art and not a science. And the winners who appear to be winning today in AI may not be the winners in the future. We're still guessing. There's a remarkable amount of guesswork, chance, evolution going on in this industry.

Speaker 1: I mean, I don't doubt what you say, but man, there's a lot of money being put to work on stuff that is guesswork and an art and chance, right?

Speaker 2: An incredible amount. $245 billion of CapEx last year by the major AI firms. Just an astonishing amount of build out. And the question they've got to be asking themselves now is, what exactly did I buy? Am I a leader as a result of spending all that money? Or has somebody else found a cheaper way in which, you know, I think they are leaders, to be fair. But I think they've got to ask themselves whether that's money well spent and whether they can convert it into revenue. That's the golden question in AI. We're going to spend all this infrastructure like we're building out the railroads all over again or creating the internet all over again. There's got to be an enormous payout at the other side of that. And these companies that have made the investment, they want to believe they're leaders.

Speaker 3: I'm curious about, would you expect to see companies, and I would expect many startups, you know, younger, you know, counting every nickel, to move toward this app? And yes, it might be a Chinese company, but there are concerns about cost. And they may say, well, it's a Chinese company, but it is meaningfully cheaper than some of these American rivals.

Speaker 2: Well, it's a lot cheaper and also it's open source. So you could just take it and maybe you could even disable the censorship. So yeah, it'll probably get used, definitely. But most importantly, it'll be used as an inspiration for how an upstart firm with a new idea could create a great AI product. And it's a demonstration of the fluidity of this market.

Speaker 1: Matt, I couldn't help but noticing that Appian shares were up today at a time when we saw a lot of the ripple effect negative amongst big tech, up three and a half percent. What are the implications, if any, for you guys from sort of a changing AI landscape?

Speaker 2: Well, you know, we don't create AI models. We merely use them. We're dedicated to making AI valuable in a business context, to implementing it and getting a return on the investment. And so the more accessible AI becomes, perhaps the more attractive becomes our value proposition. That might be the reason.