The future of tech: AI, robotics & quantum computing
01:21:21 Share

Welcome, ladies and gentlemen, to Mario's Bistro. The special tonight is the Beef Carpaccio. With the Venmo debit card, you can turn the basketball game tickets your friends paid you back for into a romantic dinner that you can earn up to 5% cash back on. Use your Venmo balance to pay for the things you love to do. Visit venmo.me slash debit to learn more. The Venmo MasterCard is issued by the Bancorp Bank N.A. Pursuant to license by MasterCard International Incorporated. Terms apply. Dosh cash back terms apply.

You're listening to a podcast from Washington Post Live, bringing the newsroom to you live. Well, good morning and welcome to The Washington Post. I'm Kat Zekreski, a White House reporter here at The Post. I spent the last decade covering technology, and I'm so excited to be joined today by the chair of the House Energy and Commerce Committee, Congressman Brett Guthrie of Kentucky. Thanks for having me. I appreciate it very much. Thank you so much for being here. Thank you.

just to kick things off, President Trump has talked about how he wants to be the, how he wants the United States to be the, quote, world capital of artificial intelligence. What do you view as Congress's role in that goal? Well, I think we absolutely have to

It has to be. And it's interesting. So the Energy and Commerce Committee is right in the middle of the role. And so my kind of background was, and not even my professional background, is manufacturing. So it kind of ties it together. But my background in Congress has been chairing the Health Care Subcommittee. So Energy and Commerce has vast jurisdiction, as you know. AI is important to health care. But that's kind of been my background. And as I was looking to be chairman, they have energy, environment, technology, all the other areas that we have.

It just kind of hit that I was visiting out in the Silicon Valley and the West Coast, and you hear people, things like Microsoft, say that our data centers require as much energy, some of them, as the city of Seattle. And so you see the needs of it kind of, our committee's jurisdiction fits. John Dingell, who used to be the chairman, used to say if it moves its energy, if it's not moving its commerce, that's our jurisdiction. And my Dingell core, Larry, I call it is it takes energy to move commerce.

And so then I was, we had Eric Schmidt from Google talk about his book Genesis that he wrote with Henry Kissinger. And he lays out what President Trump said and what you are saying is that we have to be China to be the dominant platform across the world for AI. Eric Schmidt didn't say it but I walked away going it's the same thing as being the dollar the world currency. Who controls the platforms for AI? And the biggest, we have the capital. We

We have the brain power. The biggest limit is energy. And then if you look at what Europe has done with AI, do you over-regulate? And so our job is to find the energy resources to produce AI because AI converts energy to intelligence. Someone told me that. And also not over-regulate. We have to be careful how our data is used. But we also don't want to over-regulate that we destroy the AI industry that seems what Europe has done.

And I want to come back to those energy demands, but you mentioned the concerns about overregulation. We currently are seeing industry executives play an unprecedented role in Washington and in the Trump administration. Are these leaders from industry looking out for the rights and needs of consumers? Well, I think in every era, if you read it, as I always tell someone earlier, everything's repetitive. And so you had the

whenever new industries blossom and grow, there's how big is it? If you look at the 1880s, when Silicon Valley first became around Detroit in the 1900s, how do you regulate it? Somebody wanted to regulate General Motors in the 1950s. There's no way anybody could ever compete with them. And so we have to be careful to know that

We want our data protected, but we also want to make sure that we don't overprotect, that industry moves somewhere else, and that's what we have to be worried about. And so what we've done on Engine Commerce, my vice chair, Dr. John Joyce from Pennsylvania, has put together a working group, and we're going to do what we think is the right thing. There are people in town that think you should regulate like Europe or California regulates, and I think that works.

puts us in a position where we're not competitive. But we also don't want an open free range where we have to know where our data's protected. And what the answer to that is, we're really gonna put a lot of effort into find out. I wish I had all the answers, but we're gonna put people in the place to try to find the best answer. - And because we've seen really with the tech industry, if you look at what happened with social media and the internet,

Largely, we did not see Congress pass data privacy regulations, children's protection since the 90s, Section 230 changes. Why do you think AI might be different? Well, so we're going to do it. So it's not just AI. It is the internet and social media. As a matter of fact, that's the bigger focus. We're having a hearing actually next week on Ted Cruz's Take It Down Act. We're going to look at children's online safety.

So I would project there's going to be a bill out of the House of Representatives sometime this year, if not before the summer. If the Senate passes one first, we'll have to deal with that. But I don't know why it's taking so long. I just became chairman a few weeks ago, a few months ago. But we're going to focus on making sure. But if you over-regulate, like Europe has done, I think Eric Schmidt said it in the Library of Congress meetings, I assume it's public, he said Europe has chosen not to grow.

And it's because of what they've done with energy, where they made their energy so expensive they can't compete. And then they overregulate in the private market. So it's not that I want to protect Silicon Valley and their product. I want to make sure that we win the battle against China. And that's what we have to do.

- On that point of the battle against China, we recently saw China release DeepSeek, a competitor to ChatGPT. How concerned should Americans be about the release of DeepSeek? - It needs to wake us up that we're in a real competition. And we always feel like we're gonna be the winners of everything 'cause we always have been. And we have the brain power to do it. We have, I said, the capital to do it. And so we have to make sure it's just not always gonna be there.

And my understanding with what DeepSeek did is that we have superior chips. So they had to learn how to take inferior chips and be more efficient with them. And that should wake us up. And that we're in a real competition. And the competition is the energy to supply and the right regulatory framework so we can unleash the power that we have. And on that point of energy supply, we actually have a question from the audience. Erin Vanderberg from California asks...

Do you think the U.S. will be able to adapt to the massive demand for energy to power the increasing computing needed for AI? We absolutely have to. And so the first thing we need to do is not take baseload power offline. I know a lot of people won't get a fossil fuels. I'm from Kentucky, so it's a coal state, I understand. But you already have baseload power that's projected to come offline, right?

at the amount that we're projected to grow. So we're projecting to take as much power offline as we're projected to grow in the next decade. And that's just untenable if we want to be China. So that's why the wake-up call is important when we saw what happened with DeepSeek. And so we have to be innovative. You're seeing what's happening. We can't take power offline. And we have to be innovative with new power sources,

Microsoft said that their data center creates as much power as the city of Seattle can. So if you think about the way Seattle grew like any city, a few thousand people moved, then a few tens of thousands, then hundreds of thousands. So the grid grows over time. Now you're going to put down, boom, as much power in one setting that it takes for the whole city to run. And so it completely has to change how we regulate it. And I think it's a good

solution for geothermal. That's another non-carbon way. And then obviously small modular nuclear reactors have to be part of it. And some of these data centers may be powered off, be connected to the grid for backup, but be powered off the grid with their own power source. And so what exactly is your committee doing then to move forward with those needs that you just laid out? Well, almost all of our hearings have been focused on

It's focused particularly in the environment and energy. We have environment and energy subcommittees, is the energy demands of AI. We're going to have an AI hearing coming next week. In a couple of weeks we're going to have just a full committee hearing on AI. And then in our telecom and our commerce we have other subcommittees that deal with privacy to see what is the right lane to make sure we're conscientious of the environment,

But we also don't want to put ourselves, and I hate to keep bringing up Europe but it is so easy. We have an example of what happens when people go too far in their energy policy making themselves dependent on dictators. That is a foreign policy issue. But also price themselves out of being able to compete in this world. We had the same, Europe and America had the same economy.

2008, same size economy, European Union including Britain at the time. And now we are far bigger than they are because of decisions they made. So my job is for us not to go down that path.

but also find the path that's going to beat China to AI. And what is your response to people who raise environmental concerns about this use of energy to power artificial intelligence? Well, Eric Schmidt said that we need to use energy because it's going to produce the solutions to climate change, is what he says. So if the most existential threat to America is climate change, is your view, or is it losing the battle for AI?

So if you choose climate change we're not going to produce the energy. We're going to lose the battle to AI. And I will tell you China is producing the energy. They're doing one coal plant every two weeks. So not only are we going to lose the battle for AI, you're also going to see the battle on climate change. And that's why we have to look at both. But we can't go down the path some other continents have done.

I want to ask you a little bit, given this concern of competition with China, we know that they are heavily spending on research and development when it comes to artificial intelligence, drones, quantum computing, these technologies that are key for national security.

Are you concerned at all about the cuts that the U.S. DOGE service is making to scientific research and grants to research universities? - So the grants to research universities, I think we need to look at it, and so my jurisdiction would be under NIH, particularly in healthcare and a lot of that area. And so,

We have to look at what's the proper overhead. I think what DOJA is doing is proper to say how much should a research grant, should 60% go to administration? Or should we want more of it to go to, and they're all different. And so we need to find them. We should have hearings actually looking at what is the proper, how much money that we send to research should go to an institution just running its institution. I think that's a fair thing to look at.

- And can you tell me more about what those hearings would look like? Would they involve representatives from DOGE as well, or would it be primarily focused on the universities and recipients? - Well, you know, we haven't put a hearing together, so we'd have to have that discussion. A lot of stuff in Energy and Commerce, believe it or not, are bipartisan, so I think most people wanna know if I'm gonna send a dollar to an institution, how much is it gonna go into, and there's some costs to administer.

So it's fair to have a portion of that go to pure administration. But you know when you see grants that's 60% and then one, I think one grant, then they say well we had to build a lab for it. Okay, but the renewal of the grant still had 60% overhead research. So I think it's fair to look at that, absolutely fair to look at that.

Given the activity we've seen from Elon Musk and Doge so far, what grade would you give their performance after these first couple of weeks? Well, I think, you know, if you look at just going in for a few months, I think it's too early to grade. I think you've got to see what they're doing, and they're going in and trying to find opportunities to save money within federal government. To make it more efficient, that's the goal for it. And if we can save money through efficiency and waste, it's not going to solve our budget problem, but it's certainly going to be a big step in the right direction. And

- And what's your personal relationship like with Elon Musk? Do you have a direct line with him? - I've only seen him, we've never chatted. - Got it.

- And I wanted to talk to you, you just brought up the topic of the budget and I wanted to ask how realistic are the 880 billion in cuts in the House GOP budget bill? - So they're realistic, we can make those numbers or we wouldn't have voted for that. And so you look at what, we have broad jurisdiction. So you look at things like in the IRA, we have some IRA grant funding we're really concerned about. I know we tried to put them on hold, a court said you can't put on hold

You can look into those if you'd like to. There were some institutions that had a few hundred dollars in their bank account that got a billion dollar grant. And so, I think those should be real concerns for every American. And so, you look at-- But if you look in the healthcare space, there are-- If you want to get-- They'll take up our 20 minutes for me to explain it to you, but you asked, and so I'll explain it to you because it's not a simple question.

When the Affordable Care Act set up expansion of Medicaid they had one traditional Medicaid, they are the most vulnerable, single mothers, disabled. In Kentucky you get the federal government pays 72 cents on the dollar for that group.

expansion population which is healthy adults. And if a disabled child goes to the doctor in Kentucky the federal government pays 72 cents. If a healthy adult goes to the doctor the federal government pays 90 cents.

And we just think that's just unfair. It's just unfair, and it takes away money because now the states are incentivized to take care of this population because it's more money to the hospitals, more money to the healthcare system, which hospitals used to do more. I'll cede that point. But we're seeing an opportunity that we're seeing that it's met a disadvantage of the traditional Medicaid population. And so what we want to do is make it fairer, and I think you can literally...

If you look at the $880 billion number, you can take people on traditional Medicare, single mothers, others, make sure that they're eligible to be on it, and then you protect that group.

And then you can look at making it far more efficient within our expanded population. And so what we're saying is we're going to protect traditional Medicaid and some of the incentives that were developed. Because think if you're a state government and you can get 72 cents for this, if they're in this category, 90 cents in the other category, what are you going to do? You're going to put a lot of people in the other category. So we need to fix that.

They're perverse incentives. But this has really created a lot of concern among Americans that their benefits could be impacted. I mean, what's your message, particularly to Americans who are concerned about the impact on them? So I'll start with that we're going to spend more money on Medicaid every year for the next 10 years.

So this year it is going to go up next year, it is going to go up the year after, it is going to go up the year after. Medicaid currently because of some of the incentives that I was talking about, the perverse incentives is projected to grow beyond inflation

by $900 billion over the next ten years. $900 billion beyond inflation. And so, if the American people don't see there is an opportunity to preserve benefits, to make sure people have the benefits that they have been promised, but also get a handle on the large growth. And I can get into details, I don't have all morning to tell you. So, if you go to a hospital,

you are in the healthy adult population and the state pays you $100, I am going to make this simple for the math, $100 then the state pays $10 and the federal government pays $90. So, the state can go to the hospital, I am going to put a tax on you and it is going to be $20. So, you are going to pay me $20 so the state puts no general funds into the, so there is no skin in the game for the state. But by you giving me $20 I can now give you $180.

So what hospital in the world isn't gonna give $20 in a tax to get $180 instead of getting 90? And those are the things we're trying to fix and not take away any benefits. We're trying to fix those perverse incentives in the law. - But is there a way to get to the $880 billion figure without taking away any benefits? - We can do it with just fixing those issues. We will spend more money every year. And what we'll require is for states and healthcare systems

to govern within, keep the same dollar and manage their systems within inflation, within inflation not manage the level of funding. Funding will go up every year.

to turn to a topic that's in the news right now, which is the signal messages about military plans that were shared with a journalist from The Atlantic. I mean, what was your response to that story? Well, I just saw that last night, so I haven't had a chance to get anything other than what I saw in the headlines. You know, I think it's absolutely concerning that for some reason how a journalist was on a

A text message with, I don't know how that happened. I have no idea how that happened and I think it's something we need to find out. And when you say it's something we need to find out, what's Congress's role in oversight? Well, I think the Intelligence Committee should figure that out and figure out what needs to happen. And I haven't, I mean, I just saw the headlines last evening and haven't been here this morning, so I'll find out, talk to my intelligence friends and see. But that, so we do things in different jurisdictions. If I was on the Intelligence Committee or chair, I would certainly want to find out how that happened.

And I wanna find out how it happened as a member of Congress, but I think that they set up the procedure for members of Congress to figure out how it happened. - Got it. And given just your experience working in the federal government, I mean, what do you think about the fact that those conversations were happening on a commercial platform like Signal in the first place? - I don't know the content of their conversations back and forth.

We have a signal thing with our energy and commerce members that talk about different bills and things going forward. I don't know what was said on that. I just saw the headlines and I'd have to see. I don't know what they were sharing back and forth and see. But I think people communicate. I'm reading a book on Lyndon Johnson. He spent all day on the telephone. People don't talk on the telephone anymore. They communicate back and forth. I think that's important. If anything was secure or needed to be secure, obviously it needed to be through a secure network. I don't know if they shared something that was secure or not. I have no idea.

But I think that's just the way people communicate now. And to say that, well I can't believe these people in the administration were talking through text messages or some app, then, well gosh, they used to talk on the telephone all the time too. People just don't do that anymore. So it's like going back to Linda Johnson never talked on the telephone. I mean that's the thing. That's the way people communicate. You've got to make sure they're communicating the right things in the right systems, the right methods.

Congressman, I mean, from the reporting, we know that they were discussing plans, military plans for a tactical mission. They were talking about... Right, and so what I said, I don't know if what they were saying was in secret or confidential or things like that. I know they were talking back and forth. They should not have had a reporter, and I'm agreeing with you, reporter on the Signal app

But what I don't know is specifically what, I haven't seen their signal discussions to give you an answer. I just can't give you an answer. I don't know. Okay, well, I will share the Atlantic article with you, which details some screenshots from those messages and describes exactly. I mean, you can sit here and try to make me look like a fool if you want to. I saw the headline. I'm watching this. I've been here this morning because you invited me to come down here. And so I'm here.

But, I mean, they can laugh if they want to. I just haven't seen it. And I'm not going to comment on something I haven't seen because headlines are far different than what's usually in the article. Well, Congressman, I'm certainly not trying to make you look like a fool, and we appreciate you being here today. But it does seem like this is a major national security concern and even, you know, the White House national security concerns. And I should do my homework before I comment on it. Got it. Okay.

And Congressman, I mean, just looking forward to the next year, we've only got a minute left. I want to ask you, I mean, what is your top priority when it comes to AI policy in the coming months? We have to make sure we get the right level on regulation. And we also need to make sure we have the energy to produce. We have to make sure we have the energy so we can produce the data centers to generate the AI.

Okay, well Congressman, we'll have to leave it there. Thank you so much for joining us today at Washington Post Live. Thank you, appreciate it. Thanks. The following segment was produced and paid for by Washington Post Live event sponsor. The Washington Post newsroom was not involved in the production of this content. Hello everyone. Thank you so much for joining us. I'm Linda Moore and I'm the CEO of TechNet. I want to just thank the Washington Post for their partnership on this event. Thank you for those who are watching us live via stream and for those of you who are in the audience.

We're going to talk a little bit about the industry perspective on a lot of the questions that we just heard from Representative Guthrie answering from the congressional perspective. And here with me today, I'm so excited to have Aparna Bawa. She's the COO of Zoom. And Aparna joined Zoom in 2018. And then a little thing called the pandemic made Zoom an overnight sensation. And then Kent Walker has been...

the president of Global Affairs of Google. You've been with Google for about 20 years now, I think, Kent. But a lot of things, one thing that people don't know about Kent is in his early days he was assistant U.S. attorney for DOJ in D.C. and also in San Francisco. So thank you both for being with us. We have other TechNet member company executives in town. We're going to meet with administration officials and congressional leaders tomorrow about a lot of the issues we're going to discuss today.

I'm going to pose the first question to you, Kent. So we heard Representative Guthrie talk about the need to beat China and other adversaries. We've led in the tech race for decades now, and we tend to think that that's going to happen forever. So what do you think is the most concern when you think about America leading the world, and what should we be most worried about if we fall behind? Yeah, right.

It is the key question of our time in technology. It's important to go back and remember that in what was called the long century between the late 1700s and the early 1900s, Britain led the world geopolitically because it led the world technologically in technologies like coal and steam. In the 20th century, America led the world largely because we led in mass production material science. The open question is who leads the world in the 21st century?

And there's some concerning trends out there. ASPE, the Australian Strategic Policy Institute, does a study every few years about technological leadership. Roughly in 2000, America led the world in 60 of 64 different key technologies, everything from batteries to chemistry. Today, China leads the world in 57 of those 64 technologies. So that's a challenge in front of

But we do think that AI and quantum and other new technologies that are just starting to come to the fore are really the key to turning that around. AI is not just a scientific breakthrough, it's a breakthrough in how we make breakthroughs. And it really has the opportunity to keep America with a lasting technological edge. Now, there are obviously things we need to do to get there, and we can talk more about that. But the stakes are extraordinarily high. China has more than 200,000 new AI companies.

President Xi has made it a strategic imperative to have widespread AI deployment. They have the AI Plus program where they are investing in new technologies. And I'm sure we'll come back and talk about Deep Sea because as Chairman Guthrie did a moment ago, there's an awful lot of competition at the moment right now and we really have to keep pace.

Thanks, Kent. And, Aubrey, I have a question for you that's a follow-up to that. You know, lots of times it's the small and medium-sized companies that really drive innovation and take it to the next level. How can we create an environment that allows those kind of small and medium-sized companies to scale, especially on AI? You know, it's such a good question. Obviously, even when I look at the news, I turn on the TV this morning, I'm

I hear about my good friend Kent's company over here. I hear about the top seven, the big seven in tech.

I just want to remind everyone that the beauty of technology in the United States has been a combination of not just industry leading players like these top seven, but the vast number of small and medium-sized businesses that drive innovation every day. For those businesses which TechNet supports greatly,

Resources are a challenge. So we have to make it such that these players can also have a competitive advantage. And that means taking sort of a risk-based approach where we don't focus on as much on theoretical harms, but you actually focus on actual priorities

And also policy makers taking a sort of a business industry partnership approach so that you can get the benefit as people and other companies discover innovations. You can get the benefit of that and leverage that partnership to help these small and medium sized businesses get ahead. The last thing I will say is, you know, resources can be challenged. It's very easy to think of all tech as big pockets. It's not the case.

And so now just take an example of privacy, for example, the representative that was here before talked about sort of privacy legislation and the divergence between federal and state policy. It is really important that we have one overarching architecture that we can follow and that we're not sort of stymied by following multiple different frameworks, etc. That you just don't, you cannot compete, you cannot focus on best of breed innovation when you're pulled in so many different directions. And so I would say,

And these are the main things that I think we as a society in the United States can help our small and medium sized businesses to innovate around AI. Thank you.

And Kent, I'm going to toss it back to you. What do you think are the greatest opportunities for AI, but also the greatest challenges? And how are companies thinking about those? And also, how should policymakers think about it? We are living in an age of wonders right now. The potential for AI is just unimaginably fast. We tend to think of AI as a chatbot because that's the way it came up. That's a small percentage of the promise.

As we have these scientific breakthroughs, we are seeing remarkable innovations in material science, tens of thousands of useful new materials being discovered. In healthcare, our AlphaFold project folded every protein in the human body, 200 million proteins, which each would have taken a biology PhD student a full thesis, three or four years to fold one protein. We did 200 million of them in a summer and open sourced them.

It's like taking every man, woman, and child in the country, training them to be a biology PhD researcher, and having them do nothing but fold proteins for three or four years. So now we're about to see the first AI-generated drugs go into clinical trials before the end of the year. And that's just one slice. Imagine progress in nuclear fusion, in quantum combinatorial innovation, where AI makes quantum faster and quantum speeds up AI.

This is a really remarkable time we're living in. So the opportunities are vast and the importance of continuing to lead are equally important. Now on the challenges side, again you heard Chairman Guthrie talk about the importance of energy to power data centers. That's certainly right, both creating new kinds of energy, small modular reactors, taking advantage of wind power, geothermal, solar, natural gas, a whole variety of different opportunities.

Transmission, we're hopeful that Congress will actually pass permitting reform to allow us to connect the sources of power to the uses of power across the country. The United States has a huge strategic advantage. Our power is significantly less expensive than it is in most other parts of the world. We can't lose that advantage. But then the other side, of course, is the regulatory piece. We talk about the three I's, the importance of innovation, infrastructure, and investment.

Infrastructure I already talked about for a moment. The innovation, we need to have balanced frameworks around privacy, around copyright, around data use more generally to make sure that America continues to be a great place to train and develop new AI tools. And then lastly, investment. We need a workforce that is continuing up the learning curve. AI is the easiest tool you've ever worked with. It's the only tool I've ever seen that will teach you how to use it.

A hammer doesn't teach you how to hit a nail, but AI will tell you how to ask a better prompt. It also upskills the least skilled the most. Helps everybody, but the people who are just learning a skill get up to average much faster than they would have before. But we need to have skilling across our workforce, and of course, we have to take advantage of the remarkable skills of people around the world. America, through H1Bs and others, has attracted a remarkably

capable workforce of the best and the brightest across the world. China graduates four times more computer scientists than the US does every year. But the United States, and China has 1.3, 1.4 billion people. But if we play our cards right, America has 7 billion people on our side because we have the opportunity to draw from the remarkable technical talent that we're seeing around the world. The attention is all you need paper. It's a paper we published that opened up this new generation of large language models.

Seven of the eight authors were born outside the United States and the eighth was the son of immigrants to the United States. Shame on us if we don't take advantage of the opportunity of talented folks who want to come to our country and contribute to our economic growth and our economic progress. I want to follow up on that. How important is a workforce recruitment to both of your companies? Has AI had any effect on that? Incredibly important. I will echo what Kent says.

I break it up into a couple different categories. First of all, it's the existing, getting the existing talent pool in the United States up to speed on AI. I'm gonna give you an example. I have two kids. I have an older son who is very AI fluent. I think he was, I think he downloaded, he was one of the first early adopters of ChatGPT. I think I didn't know what it was when he came home and said, "Mommy, I'm using this." And I said, "Please don't cheat."

And then I have a second son who is also very proficient. And, you know, he says, Mom, I said, you know, you have a Spanish test. Why don't you learn the we don't speak Spanish. He has a Spanish test. I said, why don't you put the vocab words and note cards and write them out? He said, Mom, that is not how it's done anymore. I put the content into ChatGPT. It gives me quizzes. I take the test. And, you know, I checked and that's not cheating. That's studying. And so, you know, I think it's it's it's

It's having a growth mindset. I'm 47 years old and I need to open my brain. I need to adopt AI. I need to think about new ways of doing things. And I think back to when the internet came upon us and I think my parents had the same view. And so it's sort of having Americans today adopt new technology, not be scared to adopt new technology

And as Kent says, you know, one of the biggest things you will hear from technology companies, and we see this every day, it's, you know, we have so many wonderful graduates, you know, PhDs, scientists, engineers that we educate in this country, and we have them go home.

And it is so sad. I will tell you, it's a real story that every day we'll have engineers that are junior engineers have to leave because we cannot offer them that next step to stay in this country. And it's leveraging that type of openness that can drive the innovation for us in the United States.

And the last thing I would say is I also think that enabling better tax policy, which is one of the things that TechNet cares deeply about, to encourage companies to invest

in local talent, invest in sort of building out your local workforces, we would welcome that significantly. And again, there's a myriad of companies. It's not just large top seven companies with seemingly unlimited bank accounts. It is the small and medium-sized businesses that really need this help. Mm-hmm.

- Let me build on that if I could. We have programs, digital certificate programs that are available broadly. Khan Academy is doing some remarkable work on these fronts. A lot of the times we want to partner with community colleges, existing groups that are out there. But as Parnas says, the key is really making sure that there's a carrot at the end of those programs, a job offer or job opportunity.

So we partnered with more than 150 companies across the United States, from big companies like Walmart to smaller companies like Smucker's, Jams and Jellies, to say, if you've gotten one of these certificates, you have an opportunity to open the door, to interview for a job. My most popular tweet ever was when I announced that Google would take one of these certificates in lieu of a college degree for some of our IT jobs.

So there's a real appetite out there for people who want to get more up to speed on these issues. Right now, a quarter of the code we create at Google was assisted by AI. So this is really a coming trend and we do need to change the way we work. We need to change the way we educate as her story illustrates. There is a risk of cheating, but there's also the opportunity to have every kid in the country have their own personalized AI tutor.

and really change the way they interact, the way they learn. Learning to use the tools, learning to challenge the tools. So it's a really, it's a big opportunity, big challenge all wrapped into one.

Thank you. I'm going to switch to a different topic now. So Aparna, I want to ask you about cross-border data flows. You know, that's really important when it comes to the ability to develop new technologies. So I wanted to ask you how much of a concern is that for you and how do we create a better environment for cross-border data flows?

This is such an important topic and I think it should be an important topic for all of you because it does, it sounds so high level but it actually impacts us as individuals on a day-to-day basis and here's how. So, you know, I'll take it one level.

generative AI models in order to get better, they need to have access to good quality data. So they need to understand different languages. They need to be able to interpret the way someone speaks in different languages and their things in different languages, etc. 96% of the global population sits outside of the United States.

So you can imagine that if you restricted data transfer, you need robust models to sort of generate this higher level order thinking, and you will not have the data for that if you artificially or constrictively restrict data flows.

The second thing I would say is the biggest use of AI, the biggest economic gain that we can have in this country is to have our businesses build AI into workflows to make our human beings, not eliminate human beings or not eliminate jobs, but to make us

better and to make us more efficient, make us have the ability to do more. And so businesses are incorporating workflows into their processes. But if you restrict data across borders, then you can have faulty workflows or incomplete multinational corporations that can drive economic loss, if you will. And so for us, cross-border data flows are so important

And there's some aspect of being responsible about this. Of course, we have to take care of privacy concerns, take care of national security concerns, et cetera. But the vast majority of this data is benign. And so I think some of these things that different countries say that you must localize, you must keep all your data in my jurisdiction, and then you need to do X, Y, and Z, have this many data centers in that jurisdiction. Well, guess what happens?

The small tech company cannot be putting data centers in every single jurisdiction. There's not enough money. You know, the energy cost that it takes, the real estate cost that it takes, the bare metal cost that it takes. So you end up, again, impeding innovation. Impeding innovation from small and medium-sized businesses, not the big giants, but small and medium-sized businesses. And that is the lifeline of the United States for tech innovation. Thank you.

I want to switch to a different topic that you brought up earlier, Kent. You mentioned deep seek and Representative Guthrie talked about that for a minute too. So how big of a concern was that for you and Google? How should we all be thinking about deep seek as a positive thing or a negative thing and what it means in terms of competition with China and other adversaries? And how should policymakers think about it?

I think the chairman was right to refer to it as a wake-up call. It's clearly the most sophisticated model we've seen coming out of China.

There may have been some exaggeration around the margins in terms of the market reaction. As the story has played out, I think there may have been more nuance about how much it was trained on, what kind of chips it was using, how much it was so-called distilling answers from US models. Basically asking a million questions and getting a million answers and using that to fine tune its own model weights.

But all that said, I think if we thought we had a lead of a couple of years six months ago, now we think we have a lead of a couple of months. There are clearly very talented and focused folks in China who have recognized this opportunity and are investing very heavily.

And so as we think about the potential and what's at stake, it's important to remember that you look at folks like Goldman Sachs or Morgan Stanley have estimated that beyond the scientific advances I was talking about earlier, the economic sides of things that Apinor was talking earlier is potentially just as big. The estimates are as large as $22 trillion annually within the course of the next 10 years. That's larger just to scale that. That's larger than the economy of the U.S.,

So if you could imagine another US economy in the world and the benefits of that for people here and around the world, what do we need to get right from a policy perspective, from an energy perspective, from a business and product perspective to seize that opportunity? - Thank you. And just a really quick lightning round here as we close out. So when you meet with policymakers tomorrow in the administration and the Congress, what's gonna be your main takeaway for them? - I would say, you know,

As Kent was talking about the deep seek phenomenon, my reaction, and I feel like it's a very American reaction, is bring it. Let's compete. When we have something to rally around and compete for, we bring our best. I think my biggest ask and my biggest request for policymakers is to say, make this place as easy as possible to innovate.

responsibly, we as players in that innovation need to come with some proposals for responsible innovation as well. We do have to care about the privacy, security, data protection, and even sort of livelihoods of our citizens, but make this place a collaborative environment that we can compete so we can bring it back. Like, you know, that's...

Change that months to years again. I mean, that's what makes this place so great. It's a hotbed for innovation and we need to find those ingredients and keep them. - Yeah, that notion of a pro-innovation agenda couldn't be more important. You hear this on the right and you hear this on the left.

On the right, Vice President Vance talked recently about the importance of technological leadership, working together, technology and American workers. American ingenuity, we can lead the world. On the left, you're seeing Ezra Klein, Derek Thompson's book around abundance. How do we make things work again? How do we align the interest and the efforts of government and the private sector to really get our mojo back, to really capture the energy and the capacity that we have? We know we can do this if we focus and if we get the policy side right.

I would say, Linda, I take it back to kids because lots of us have them.

When we teach our kids to have a growth mindset, we kind of need to do the same thing. And I think that's part of the goal of engagement here in Washington. - Awesome, thank you both. I really appreciate it. And thank you all for watching. And I'm gonna turn it over to the Washington Post now. Thanks. - And now, back to Washington Post Live. - Good morning, everybody. I'm David Ignatius, a columnist at the Post. I'm joined by Martin Schmidt, who's the president of Rensselaer Polytechnic Institute.

private university known as RPI. Marty, welcome to The Washington Post. Good to have you. Thanks. It's a pleasure. So I want to begin by talking about quantum computing. And I have to just make a note to this audience that I wrote a novel in 2018 called The Quantum Spy. I've now written 12 novels. But it was way back when quantum computing was still kind of a gleam in the eye. And people said that

maybe in five years we'll have a fully programmable quantum computer that will allow researchers to do

extraordinarily powerful things. You, Marty, at RPI have a quantum computer of your own. Tell us about what it does, how it works, and some of the ways that your students and faculty are using it to open up new pathways. Yeah, well, so it's interesting.

To a lot of people, quantum computing sounds like magic when you start getting this actual quantum physics of how it works. But we had a panel down at South by Southwest. And my wife, who is not a scientist, said that the one panelist gave the best description of quantum computer, which is that if you think of a classic computer as a light switch, which is either off or on,

think of a quantum computer as a dimmer, meaning that you can have an infinite number of states from fully on to fully off. And so what that means is you can encode a lot more information in, say, a single bit of a quantum computer versus a classical computer. And that ability to put more information in means you get greater computing power. So that's one thing. The other thing is that

A lot of the world we live in is defined by quantum mechanics, how material properties are created based upon the way in which atoms interact in that material. That's a fundamental quantum phenomena. And so being able to model that behavior using something that is quantum gives us much more power

in solving problems than we might have in a classic computer. So those are the, so when you think about what is it that's different about quantum computing, it's fundamentally that fact that you can model quantum mechanical phenomena in a way, in a much more authentic way than you can in a classic computer. The other thing

is that if you think about Moore's Law and classic computers, every time you doubled the number of bits in a computer, you doubled power. Every time you doubled the clock speed, you doubled power. And Moore's Law was really all about scaling things down to get more of those transistors on, and that's how you got the exponential growth.

But we're running out of scass on that. What's interesting about quantum computing is when you double the number of qubits, the power grows exponentially. So we're about to enter a computing world where we're going to see exponential growth in the power of these computers, which is really quite astonishing. So let me ask you, it sounds technical, but it's a useful metric. How many qubits does your quantum computer at RPI have? 127. 127.

So, just in terms of the league charts, that's pretty good. IBM was bragging about having a 56-qubit computer a while ago. There are some that are higher than that. Let me ask about the thing that fascinated me the most, and I know researchers really struggle with, and maybe you could explain how your research is focused on this.

The problem with these wonderful qubits that are zero and one simultaneously and allow omnidirectional computing is that they don't last very long. They last just fractions of a second.

they decohere is the term, and any kind of electronic, magnetic, other noise makes them decohere even more quickly. So they're very fragile, and that's been, from what little I know, the biggest challenge is how to make these qubits last longer so they'll be more robust and can do more computing.

Everybody who's in this field is focused in part on that. What are you doing in that area? How are you getting your qubits to last long enough to do real computation? - Right, so it's a great question. And you're right, they can be noisy.

And so when you set up a problem that you solve in a quantum computer, you run it, you'll get a result, you run it again, you get a result that's maybe a little bit different because some of the noise was in the system. And so one of the things people do is they run it a bunch of times and then average the result, simplistic way of describing it.

There's some basic work which is being done in IBM and other places in terms of designing better qubits, using different materials, things like that. But the other thing that's happening, and we recently did some work with IBM on this, is connecting the quantum computer to a high-performance computer.

and running information back and forth between the two, where you can do error corrections and error mitigations by saddling up a quantum computer next to a high-performance computer.

We're blessed in the sense that we not only have a quantum computer on our campus, we also have one of the most powerful high performance computers on any university campus in the country. And so having those two close together has allowed us to do some really exciting things that I think is gonna shorten that roadmap to where we have the opportunity to solve things with a quantum computer.

In the intelligence community that often drives high-end research, the fascination with quantum computers has been in part the expected ability they'll have to decrypt information

any super large number that's used to drive encryption. So there's this notion that once a quantum computer really is operable, programmable, no encryption system, whether it's code for the

National Security Agency or whether it's the way you do transactions with your Visa card, nothing that's encrypted will be safe from a quantum computer. Is that overstating things? And just talk, this is a morning where we're all thinking about, geez, what's your signal? Is it safe? What the heck is Mike Waltz doing with signal anyway? Talk about the way in which quantum

quantum computing and other aspects of the advance in crypto and computing threaten fundamental security of transactions. So encryption to this point has been done using what's called RSS encryption. And a gentleman named Peter Shore created something called Shore's algorithm, where he demonstrated that if you had a quantum computer with adequate performance, you could basically crack the RSS encryption.

However, the good news is that recently a number of researchers have come out with quantum hard encryption technologies. So going forward, we need to move to those quantum hard encryption technologies. The current concern is have our adversaries recorded all of our encrypted conversations and will they be able to decrypt them in that moment? And the answer to that is absolutely.

So that's your bet, that the Chinese have been recording everything that matters for so long anyway, that they're just sitting waiting for the computer to-- - I think we just have to assume that's gonna happen and be prepared for it. - So one final question in this area, but it applies to all the technologies that you work on at RPI. Give us a sense of the Chinese competition

Xi Jinping has spoken about his desire to control the commanding heights of every significant technology going forward. From what you know, how are they doing? Are we overestimating their prowess, underestimating it? What do you think?

So I think in quantum, the US has a strong lead. When you look at the concentration of companies working on the development of quantum computers and the pace of advancement, it's really centered in the US and maybe to some extent in some of our allies. Either they're not talking about it or there's not a lot relative to where we are in China today. But I would just point out that I think I would have made that statement a decade ago about AI.

And we saw China move extraordinarily fast into the AI space. And so I think keeping up a fast pace of innovation is going to be really vital. Let's turn to AI. Tell us maybe first what your students and faculty are doing, some of the research in AI that you're proudest of at RPI. And then we'll turn to some other questions that flow from the

rapid advance of AI that we all know about? Yeah, well, I would say what's happening for us with it, I'd say it's a couple things. One is when AI, you know, when we had access to large data sets and the ability to manipulate them, which is really what AI provided us with, RPI introduced in 2018 a program called Data Dexterity. So we require every student at RPI to take a data analytics course and

and then a course in their major that has them apply those skills in the major. So from an educational perspective, we're trying to make sure that our graduates have data dexterity at the undergraduate level. From the research side, it's really a lot about how do I use these tools to transform the way the research is done in my field? Maybe it's in materials discovery. So we've got some great work using AI to do

modeling of molecules and when you think about taking those learnings and then connecting them to the quantum computer. But also using AI in modeling financial markets and things like that. So when I was at MIT, we created the Schwarzman College of Computing and the whole goal was really to say, how do we drive the proliferation of AI into all disciplines? And so I think that's what every university is doing today.

Having state-of-the-art hardware so that you can really do it is really important, and that's one of the challenges for universities when you look at what Meta or others are spending on AI hardware in order to understand it. We're fortunate. Governor Hochul made a commitment through Empire AI, which is a partnership that's bringing that kind of capability to a lot of the research universities in New York State.

So I think one of the things many of us wonder about is what will happen to employment, to the range of work skills that people do in a law firm, in a business?

goodness, any kind of research establishment. My wife has a doctorate in computer science. You'd think she's safe, but she spends her day writing code, and it's possible that a large language model will just write better code than my wife can pretty darn soon. So give us your sense, as you think about

reaching out into our economy, what changes we're likely to see over the next, say, five to 10 years as AI is adopted by the great financial institutions, law firms,

software companies. Yeah. So what I would say to your wife is that it will write software, it will write the code. It may not be better, but it may be good enough. And I think that's kind of the sense. Are you still, just parenthetically, so you've got all these students, once upon a time they would have come to RPI, they want to learn code, how many languages can I learn? Are you still

stepping back off the pedal on that because those skills actually will not be as valuable in the future for your students? Not at all. Because I think if you think about...

it's about the abstraction. So if I can get AI to do some of my coding, maybe I can step back to a higher level of what's the system look like and how do I choreograph that? So I think having that fundamental knowledge is important because in essence, the AI then becomes your employee in developing that code. You have to verify that it's good and so on and so forth. So I think not in the least, in fact, I'd say even more so, but I think that...

One of my friends back at MIT is David Autor, an economist. And David is, as you surely know, economists generally aren't viewed as optimists. But on AI, David had a piece out recently where he took a fairly optimistic view of how AI will allow certain people to be upskilled by bringing the capability to them. So if I'm a...

a medical doctor or maybe a diagnostic nurse, that the ability to sort of use an AI agent to help me figure out what might be going on, that kind of upskilling I think is a viable path. And what I would say, and what I say to the people at RPI is think about how you can use AI in everything you do. I was meeting with an alumni last week

And we had a conversation. He works in the finance sector. He said, you know, these AI agents are really powerful now where you can get your AI agent, which is customized to you, can start talking to other AIs and generate information. And so I think we all need to be thinking about how can AI help me do my job? I'm going to turn to a question that's very much in the news, and that is a sharp

and very unpredictable cuts in government funding that's available for research. NIH funding cuts have been the most obvious, hitting medical and scientific research. One of my children is a very good, young research doctor. I'm just seeing...

people falling by the hundreds, thousands around her at Johns Hopkins where she teaches. What effect do you see these research cuts having first in general at universities

as they think about how are we gonna pay for next year or the next five years. And then I'd be interested specifically at RPI if you've got some areas where government funding cutbacks are affecting you and what you're doing about it.

So I would say for the most part, we're not seeing cuts today. We're concerned because we're seeing what's happening at other institutions. Hard to tell whether those are in order to influence certain behaviors and then the funds are going to be released. That's hard to know. But I would say that it is concerning without a doubt. And, you know, when you think about the role universities have been having,

played in driving economic development in this nation. I mean, look at, you talk about AI. In 1956, there was a workshop in Dartmouth where the term AI was coined that led to the creation of a field of AI that got a lot of support, but then people got disillusioned about its impact and we went through what we call an AI winter.

But you had people like Jeff Hinton, who was trained in the UK but came to the United States because of the investments the United States was making in these fields back in the 80s. So that steady commitment to basic research put us in the pole position in AI as a nation.

Talk about the sequencing of the human genome that started in, I think it was late 80s or early 90s, that led to our ability to think about gene editing and led to the kind of opportunities we have all in the United States. So I think the commitment to basic research has paid great dividends for this nation and I think it's important for us to take the long view of that. - So two questions follow from that for me.

One is the obvious one. Are we, in effect, eating our seed corn or maybe just throwing our seed corn out the window? This is such a helter-skelter process. And the second part of this is what can the research establishment, presence of major universities like yours,

you know, the field as a whole do to put up some warning signs and say, this is dangerous to our country. Be careful what you're doing. What about that? Well, I think let's talk about the Chips and Science Act. Okay. So there's a bipartisan act where there was an explicit recognition of the importance of reshoring semiconductors, but also basic investments in science. So I think

did my PhD in the semiconductor field. That's been my research area for decades. And what I can say is as those factories went offshore,

universities stopped doing basic research in semiconductors because the PhD students weren't interested in pursuing that as a research topic because they didn't see the jobs, the research jobs for them necessarily at scale in the US. And so I think we lost our leadership in some of that. And I think bringing that back is really important. I don't think we want to lose our leadership in other fields. And so I think we have to keep our focus on what are these emerging fields and are we really properly supporting them?

And just, again, ask the sharp edge of that. Do you think we're in danger of not having the seed corn and the corn later to harvest and losing the edge that you described? Not yet, but I think if we did make significant cuts in basic research at universities and other institutions, I think it would be a very serious concern. And the other thing I would say is that

When you look at particularly graduate schools in the US in science and technology, a lot of the students doing that research come from outside the United States, but they stay in the United States. And so I think making sure that we're continuing to bring in the brightest minds from around the world, which make us stronger.

and making it a place where they can continue to thrive because we're committed to these things and we create that kind of innovation environment in this nation. I think that's the key thing.

Last question follows from that. It is true we get these brilliant students from overseas, superbly trained in STEM skills, but we wonder what about STEM education in the United States when you look at the international tests? We keep falling further behind some of our peers. What are your brief thoughts about how to improve STEM education in the United States?

Well, I think part of it is sharing the excitement. So we have at RPI a program called Engineering Ambassadors, where our undergraduates go and talk to K through 12 students and just try and get them excited to drive that curiosity about how does that work and how can I actually have an impact on it. So I think it's a boots on the ground effort across the board, but I think it's also creating opportunities.

When we got the quantum computer at RPI, we announced we'd get it. I found out that there were 40 students at RPI who had spontaneously formed the Quantum Computing Club a semester ago. I didn't even know about it. So they were really excited to find out they were going to get their hands on one. Well, guess what? A year later, there's 400 students in that Quantum Computing Club because they're just excited. And they get to play with this tool and hopefully create new careers.

Well, your conversation with us, and I'm sure with your students when you go back, does generate that interest in STEM. Marty, thank you so much for joining us today, being part of this conversation. Thanks very much. Pleasure. Thank you.

Please stand your seats. Up next, my colleague Jonathan Capehart will be joined by robotics startup founder Jeff Cardenas. Thank you, Marty. Good morning and welcome. I'm Jonathan Capehart, associate editor at The Washington Post. Joining me is not the humanoid you just saw, but the human you just saw in the intro video, Jeff Cardenas, CEO and co-founder of the robotics startup Aptronic. I always...

Get all the letters mixed up. Apptronic. Jeff, welcome to The Washington Post. Thank you for having me. You've been working in robotics for more than a decade. How has AI impacted its development? AI has had a significant impact on robotics overall. I think it's really the final building block that's going to enable robots to really become an everyday part of our life.

The way that I like to explain what's happening in AI is we used to have this ceiling in robotics where the progress we were making was limited by what we could do with the robots, how you program them, how rigid they were to new environments. What's happened with AI is that ceiling has now been removed and there's a lot of excitement now overall about the future potential of robots and the way that they can impact the way that we live and work. Now, your company has partnerships with Mercedes-Benz, GXO Logistics, and the manufacturing company DLM.

Jay Bill, how are your robots being used by these companies? So it's still early days for the type of robots that we build. The way I like to explain this to most folks is that you can think of this like personal computers in the early 1980s. So think of industrial robots that we have traditionally in automotive and other areas of manufacturing like mainframes. And this is the personal computer. This is a much more versatile, general-purpose platform that can do a wide range of things.

but it's early days overall. We just had these breakthroughs in generative AI in the last few years. So right now what we're doing with folks like Mercedes is we're in the pilot stage. We're figuring out what applications are the robots good at and really testing them out. And I think largely for the industry as a whole, we're going to be in the pilot stage for the next couple of years. You'll start to see real commercialization happening in 26 and beyond.

- And Tony, that anticipates the next question I was gonna ask. When does this move from being an industrial force to being a part of our everyday lives? When will that be? Are we talking within 10 years? Within five years? - I think within 10 years you'll start to see them enter the home. At some point I will say everyone will have a robot in their home. The question will be what form factor. I think many of them will be humanoid in nature, but they'll be different, the sort of,

The analogy I use is do you want C3PO or R2D2? Some people want R2 instead of C3PO, but there will be robots in every single home. They're going to happen in three phases in my opinion. So phase one is in the industrial base, logistics to manufacturing. This is where we can get robot density very high. This is where we can also put robots around experts, people that are highly trained to be around the robots until we really solve safety.

The next stage is going to be in places like hospitality, retail and healthcare, which is an area I'm very interested in overall. I think these robots will have a big opportunity to make an impact in healthcare. And the final stage is going to be the home and assistive care. And for me, this has been the North star from the beginning is, uh,

assistive care and elder care, which I think has the potential to impact humanity in the biggest way. But those won't happen necessarily sequentially, but in terms of total uptake, this is how it's going to play out. And so we're at the beginning. We're in the industrial stage, and you'll see this accelerate over time. Well, since you mentioned health care or home assistive care, talk more specifically about what –

What the humanoid robot or humanoid, is that redundant, humanoid robot? What do you expect the humanoid to do? So I think it'll do a range of things. I think it'll be different in healthcare than it will be in assistive care. So I'll explain both. In healthcare, we all know there's a nursing shortage, right? This is something that's been widely publicized. One stat I've seen is by 2030, we'll be short 10 million nurses.

So maybe an older way of thinking about this is, well, we'll need robot nurses. But do we want robot nurses? No, what we want is we want better patient care. My dad and granddad were hospital administrators. My mom works for a hospital chain, so I grew up in hospitals. And if you see what's happening around the country today, they're putting screens inside of all the patient rooms because the nurses can't get around to all the rooms.

So my view of an optimistic version of this future is that you pair nurses with the robot. If you look at what nurses do, an interesting question is why do we have a nursing shortage? 40 to 60% of the job a nurse does has nothing to do with their clinical training.

And this is driving nurses out of it. They're basically the support staff for the entire hospital and kind of everything falls on them. You compare a nurse with a robot, such that the nurse or the robot does the fetching the supplies from the closet. It's running all the errands around. And that frees nurses up to do much better patient care where you have humans taking care of humans in a much better way. So that's how I think health care is going to play out. Think of these robots like the base layer support staff for the hospital to enable better patient care.

In elder care, this is one I'm particularly close to. So I was very close to both of my granddads. They both lived into their 90s. One stayed in his house. The other one had to go to a home. And they both had unique stories. And the gist of what happened to both of them is as they aged, they lost their dignity.

So they never relied on anybody for anything. Both of them were war veterans and war heroes. And at the end of their life, they had to rely on people for everything. Had one granddad that had colon cancer. He lost control of his certain abilities that he had to have other people take care of him, clean him, take a shower, do things that were pretty humiliating for him. He had his brain was there, but he had to rely on other people.

I sort of dreamed of this idea that one day he would have a robot that would carry his secrets, that would allow him to keep his dignity, that he could rely on, that would be a human helper. And this is sort of in the limit what I think is possible. I think this is the extension of computing. There's these different sort of ideas about computing early on, this idea that computers will replace us or computers will augment us.

My view, my vision for the future is that robots will be human helpers. They'll help us in all aspects of our life and maybe nowhere more important than as we age and as we grow older. So we got a taste of what your Google DeepMind partnership looks like. I think I may have even missed a question to ask you about that, but you can answer the question. After we show this video, just watch this video. What do you see on the table?

I see some colorful letter tiles on the table. Now could you spell me something that could be found in a deck of cards? Okay, how about the word ace? I can slide the tiles to spell it out. Can you pack me some trail mix? Certainly. I have packed the trail mix for you. Hey, can you pack the orange for me? Of course. Wait, what? You're just gonna- *laughs* Ending the video there? Um...

Okay, who taught the robot? Because that was pretty incredible to see it spell out the word "ace."

So that's a highlight of our collaboration with Google DeepMind. And that's work that we've done directly with the Google DeepMind team. And that's work that they've really been pioneering for many years. So a lot of the core AIs that have been invented have really started in the Google four walls. And we're excited to partner with them to really push robotics to the next stage.

Okay, so we were talking about this backstage, but I have to ask this after showing that video. It reminded me of the 2004 movie, I, Robot. And I'm just going to read the synopsis of that movie that I found on the interwebs. In 2035, highly intelligent robots fill public service positions throughout the world, operating under three rules to keep humans safe.

Despite his dark history with robotics, Detective Del Spooner, played by Will Smith, investigates the alleged suicide of U.S. robotics founder Alfred Lanning, played by James Cromwell, and believes that a human-like robot, Alan Kudyk, murdered him. With the help of a robot expert, Bridget Monahan, Spooner discovers a conspiracy that may enslave the human race. Now, I did not write this, but this gets to what I think a lot of people wonder about,

about robotics and humanoid robotics is, you know, you say they could either be human helpers or people think they'll be helpers or replace humans. Disabuse those of us who might be thinking, you know, the robots could take over. Yeah, I mean, I think certainly in the West with things like Terminator and iRobot and all these...

movies, there's more fear. This isn't actually true globally. If you look at other cultures, they're really embracing robotics. I think that's an important topic for us to think about because this is going to be the key to national competitiveness long term, period.

I think, you know, overall, we've been afraid of every major technology that has, you know, come into society. We were afraid of railroads. There was these ideas if trains moved too fast, it could scramble your brains. We were afraid of electricity. And we were certainly afraid of computers. You can think of these debates in robotics as basically just an extension of the same debates that have waged on in computing from the very beginning.

And we saw this play out very differently in computing, right? There was this idea of you'd have a four-person company, you'd have a computer that would sit at the middle, and this is all you would need. There was a competing idea that humans are tool makers, and you could pair man with machine, and this would supercharge human productivity, and this would be better for everyone. And this is my view of robotics.

So there's different viewpoints in terms of how this plays out. The lab that we came out of at the University of Texas is called the Human-Centered Robotics Lab. So not just as a company, but from a research standpoint, we've been thinking about this idea of man and machine for a long time. And I think that it has the potential to be extremely positive. I think we have to be mindful of the way in which we do this.

My point of view as a naive grad student coming into all this was look this is gonna happen I think it could be very positive and I want to work on shaping it for the benefit of humanity overall So my view is that the positive optimistic version of this is you do have this intelligent expert in your home Imagine as you're aging and you say hey look at this on my arm Should I be worried about this you have a problem with your two-year-old you have a robot in the house that you can talk to and

You can already do that to your phone, but the next stage of this will be embodying this where we have physical AI, where you actually have this little character that's like a pet or your friend that you're working with around. And I think that's a more optimistic version of the future

And we've got to be smart about how it plays out. So on a practical level, how much does it cost to build a robot like the one we saw? And is that Apollo? That's Apollo that you saw. That's version one of Apollo. That's the first prototype of Apollo. The target price point that I've said from the beginning is sub $50,000. But I think these will be much more affordable over time. You can think of these like cars.

And if you compare them to a car, there's roughly 4% the raw material by weight in a humanoid robot as compared to a car. So I think they'll get very cheap over time, like single digit thousands of dollars for some versions of them. And there'll be different versions. There'll be industrial grade versions that are working 24 hours a day that'll be more expensive. And then there'll be the lighter weight home version of the robot that'll be much more affordable. But I think they'll be accessible to everyone at different levels.

And they'll be a big part of the way we live and work. - How soon do you think they'll be accessible? What's it gonna take to scale up? - You're already starting to see robots breaking below the $50,000 price point.

They're limited in terms of capability and functionality. So I'd say we already have quote-unquote affordable robots today. But this is the very beginning of this. Generally, these robots are in the thousands of unit quantity, the type of robot that I'm talking about. So it's the very beginning. If you compare where we're at in computing at the same level of volume or smartphones or any other technology that we use, cars,

We're in the very, very early innings and they'll get much cheaper over time. It's early days for the home, I should say. Everyone wants to know when these robots are going to fold their laundry. I want this as well. And we're still a ways out from that. You've said your goal is to have robots build robots. So how close are we to that reality? We're at the beginning of that. And I think what we're seeing is we have a big labor shortage all across the industrial base.

And so what we're doing initially, and we're trying to be thoughtful about this, is we're having the robots help with support tasks, similar to what I talked about with nurses, where the robots are doing things like material handling and inspection, things that robots are better at. And so they're a part of robots building robots, but humans are still a big part of this and a big driver of this moving ahead. In your company's latest funding series, you raised $350 million. What does it mean to have the support of a big company?

tech company like Google? Well, for me, I sort of have dreamed about all this stuff since I was a kid, and I dream about this future. I think in some regard we've lost our sense of wonder about the future that we had in the 1950s. And so I'm excited about the future, but I do see the responsibility that we have in front of us and this goal to get it right.

And I always dreamed about working with Google. I felt like they were really focused on trying to get this right, trying to be thoughtful about AI for the benefit of humanity. And so I think that it allows us to have the resources to actually really think about things like safety and really do this right in a big way. And I think it also gives us the opportunity to really push this forward for the benefit of everyone.

You've used the word safety a couple of times now. Can you define what do you mean by safety in this conversation, in this context? Well, there's two types of safety. So there's the robot safety, which is an area that we've spent a lot of time on. How do you make the robot safe to work with and around?

In 2004, we introduced in robotics something called collaborative robots. This was the first time we could actually have robots that were safe to operate around people, that had a sense of touch. Generally, it's called force feedback, so force-controlled robots.

So how do you have robots that are safe to be around and safe to interact, but also capable? So these are trade-offs. The safest robot would be one that didn't move at all. But we want them to do all sorts of things. So we have to make them safe to be around people, safe to be around your two-year-old child or your 90-year-old grandmother. So that's one aspect of safety. The other aspect is AI safety.

This is something that I'm not an expert in, but the folks at Google are really pushing forward and they're really thinking a lot about how to get this right. So those are two pieces of the puzzle when we talk about safety. There are a lot of competitors in this space. So I'm curious, how do you measure up against, say, Tesla and Elon Musk?

My view is that this is the space race of our time. So these robots will redefine the way we live and work. And I think it's anyone's race to win. My wife used to joke early on when we started the company, she said, well, if anybody ever wants humanoid robots, you guys are the guys, but nobody wants humanoid robots.

in 2016. So I think this is anyone's race to win. This is something that I've spent most of my adult life thinking about. And I think we have a unique point of view that I think is going to be really important as we move ahead. And that's about human-centered robotics and human-centered design.

So the way I like to think about it is we want to bring the same point of view that Apple brought to computing, this idea that humans are tool makers and robots are the evolution of our tool making, and we want these robots to work with us and around us and to truly be human helpers to take us into the next stage. And so this is going to be really important about the design of the robot, the things that the robots are used on, and this is something that we've been thinking about for a long time. And I think that hopefully that will be an optimistic voice

in this race that's in front of us. You just said this is anyone's race to win, but China has a national robotics strategy. One, does the United States have a national robotics strategy? And if we don't, what should it look like?

Yes, so this is really important. That's actually why I'm here in D.C. is I'm working with the Association for Advancing Automation to really push forward a national robotics strategy. It's been something that the robotics industry has been working on for a long time, and we think that right now is the time that we need to get this right.

overall. Yes, China does have a national robotics strategy. They've had one since 2021. They also just announced in the last week that they're going to invest about $138 billion in their domestic robotics industry, and they want to lead the world specifically in humanoid robots.

There's an interesting history of robotics that's important to point out in this. So the US invented the very first industrial robot. It's called the Unimate Arm. The company was called Unimation. It went into a General Motors factory in 1961. This was the very first robot to enter the industrial space.

Even though we invented the first robot, that company ended up losing support, losing funding by the 1980s. We pulled out of that. For a lot of the fears that we have today about robots and labor and all of the constraints we have, you can see what that did to the manufacturing base. So we fumbled the ball in the first inning of the industrial automation space or the automation race.

It's important that we get this right. So I appreciate all of the challenges and the seriousness of this type of technology, but this is not something that we can afford to sit out. This is something that we have to lead and shape just like the Space Race. We want to get there first and we want to be able to set the tone for the rest of the world. And so this is something that we're pushing forward and hopefully we have the support of the folks here to make this happen.

- And so how do you measure up against your competition, your competitors in China? - So China is great at industrializing. The interesting thing about China is they have four times the population as the US.

but they have much more uptake of robotics than, they have more uptake in robots in China than the entire rest of the world combined with four times the population of the US. So they're very good in industrializing and they're very good at hardware. I think we have an edge in AI. An interesting point about sort of US backing this, what's created the modern robotics industry, in my opinion, is something called the DARPA Robotics Challenge.

So our work started at NASA and this was funded. There was a big competition called the DARPA Robotics Challenge. This is where the famous Atlas robot, the back flipping robot that everyone sort of points to. This was funded for the DARPA Robotics Challenge. So the US invested heavily in robotics in the 2013 to 2015 timeframe, but has really pulled out since then. And so this is something that we really need to lean forward in. And I think, like I said, it's anyone's race to win. There's different advantages on both sides.

And I think that we need more domestic manufacturing. I think this will be an important key to the future of our competitiveness across the globe. Jeff Cardenas, CEO and co-founder of Apptronic, thank you for joining us today at The Washington Post. Thanks for listening. For more information on our upcoming programs, go to WashingtonPostLive.com.

You listen because you know the power of good journalism. And The Washington Post is there for you 24/7. When you become a Washington Post subscriber, you get exclusive reporting you can't find anywhere else. You also get sharp advice columns, delicious recipes, TV and music reviews, and so much more. Right now, you can get all of that for just $4 every four weeks. That's for an entire year.

After that, it's just $12 every four weeks, and you can cancel any time. Add to your knowledge and discover all the Post has to offer. Go to WashingtonPost.com slash subscribe. That's WashingtonPost.com slash subscribe.