The Stanford Emerging Technology Review 2025: Frontier Tech For A New Geopolitical Era

>> Condoleezza Rice: Hello, I'm Condoleezza Rice. I'm the director of the Hoover Institution. And I'm delighted to be here on Capitol Hill to help launch our Stanford Emerging Technology Review, which I'll describe in just a moment. But I really want to thank Senators Young and Hickenlooper for hosting us here.

I have a connection to both of their states because I did my master's degree at Notre Dame. I'm a dormer, and I spend most of my life in Colorado. And I actually am also a part owner of the Denver Broncos. So delighted to have them here. We're going to talk a little bit about the role of government, particularly Congress, in thinking about our country's innovation ecosystem, about how we win the technological arms race that we're in with China, which is integrated into the international system, integrated into the technological ecosystem, and yet adversarial in so many ways.

So this is a complication that we've actually never faced as a country. The Soviet Union was a military giant, but it was an economic and technological midget. That is not the case with China. And so we have a particular challenge, and I think that most of us would say that running fast and running hard in these frontier technologies is our best way to assure our economic security and our national security at a time when these frontier technologies are challenging every aspect of life.

A little later on, after we have a conversation, you're going to hear from a set of folks who've been involved in the Stanford Emerging Technology Review. And what we've done is to go to people who are in the labs, at the bench, actually doing the leading edge work in these technologies.

We're not a group of political scientists talking about technologies that we don't understand. And so we will be able to introduce to you people Fei-Fei Li from computer science, an AI specialist, one of the world's most renowned AI specialists. Allison Okamura, who is a roboticist and works on the most advanced aspects of robotics.

We will have with us. Mark, are you there? Yes. Right. Mark is the chair of Electrical Engineering at Stanford University. And we'll talk to you about chips and also maybe a little bit on quantum as well. And Herb Lin, who is a physicist and a specialist on a number of technologies, including cybersecurity.

And then Amy, Amy Zegart, she is a political scientist and she's going to bring together these thoughts. But the Stanford Emerging Technology review is for us, science and technology forward. In other words, we need to understand the technologies before we begin to try to think about what policies we need as they change every aspect of our life.

So Senators, thank you very much. And I'm going to just start with a question to each of you, which is how do you think about the world that we find ourselves in with these emerging technologies, these transformative technologies? And how do you think about the role of Congress at this particular time?

So may I start with you, Senator Young?

>> Todd Young: Of course. So thank you, Condi, for the question and it's great to visit with you again. Great to visit with all of you. And congratulations to Stanford for producing this year's updated version of the Emergent Technology Review. I'll tell you, I've consistently read it since its first publication and these sorts of issues are both top of mind and the center of focus for me and my office.

The convergence between geopolitics and technology, they've always been important. In fact, one can trace GDP and geopolitical power to out competing their adversaries, out competing other nations to technological first innovation and then adoption. We don't talk as much about the latter, but I think increasingly that is something we need to discuss here on the Hill, how we can de risk and encourage more adoption of these technologies.

Also, I think we need to continue to nurture our ecosystem in various ways. I think of artificial intelligence, which is so central already to our way of life. And will find its way within seemingly every facet of our life and every area of our economy in fairly short order, I believe.

We have world's broadest, deepest capital markets. We have an amazing system of education, particularly higher education. We have some of the world's best startup companies and big tech companies alike. We can't take these advantages for granted. In fact, now on account of the brute force economics of our adversaries, we really need to figure out how to optimize our existing system and account for its shortcomings too.

And so in all those areas I mentioned, those will be critical to AI. Emerging biotech is another area that it's a platform or a general purpose technology. Here in the United States, most of us think of either bioag or biopharma, both very important in my state of Indiana, but we think less about bioindustrial innovation and applications.

I'm chairing a National Security Commission on emerging biotechnology right now and we'll in the next month or two produce our final report. And within that report we really emphasize that the United States is about to be passed up by the Chinese because they have recognized the importance of bioindustrial applications to biotech and I think most of the changes we need to make there will be regulatory in nature and also enabling more biomanufacturing to occur in this country.

Another area where I've been quite active in recent years is semiconductors. I think Congress has done quite well in showing leadership and being attentive to our semiconductor needs. In the last few years, through the Chips and Science Act, we've had roughly $30 billion of taxpayer monies flow, that's real money.

However, if you're unlocking roughly $450 billion in private sector investment, getting access to new innovations and new talent that you didn't previously have, and making more resilient your supply chains in the process, I think that establishes a decent model for some other technologies where we need to increase our resilience and bake a national security premium into.

To some of the things that we buy. So in short, I think Congress has a very active role to play. This is institutionally going to create some challenges for Congress. Congress needs to act. Congress needs to hold more hearings. Congress needs to realize that in each of the areas and many others, tech areas that I mentioned, we need upstream investment in research.

We don't need to be cutting research right now. We actually need to be planting more seed corn for innovation and growth. And so it remains to be seen whether we step up and meet this moment as an institution, but I believe we will. Our minds are focusing. It's recently become fashionable to work on what is now, I think this is an accepted term to work on economic security, a term that was not particularly embraced by many folks on my side of the aisle just a couple of years ago.

And so I, I think that will incentivize more members of Congress to activate their interest in this field.

>> John Hickenlooper: What he said, I don't, I'm not. There's little I can add. I will recognize that Senator Young has been a leader in the entire Senate, entire Congress on this and really was raising issues around our investments in scientific research and chips, and the supply chains are necessary for these things.

Big, complex creatures that if not nourished, become weak and have points of, of failure. So anyway, I want to recognize him as my senior in the Senate, which is very, very important. And I love the fact that Indiana is, is so well known as a place of innovation on so many different levels.

I was just actually interviewing for a confirmation hearing, the soon to be secretary of the Air Force, who is a technologist, really a sharp guy, shouldn't be saying these things out of school, but he has been judging robotics competitions for 12 years. And there are two points about that that I think are worth mentioning here.

One is that he never sees someone in the competitions who's a senior, right? In other words, if they don't catch kids at a younger age and get them involved in the robotics, they're lost and they're not going to come join the team when they're a senior or pretty unlikely even as a junior.

And the second part of that is, and I think we see this all over, that kids coming out of rural backgrounds where they're still making things and fixing things themselves, you give them the training, the skills, the mathematics, the engineering applications, they're going to rise to the occasion again and again and you're going to get a much higher probability of kids that are going to rise to that top level.

Research, again, that's a very rare person. But we gotta get a lot more input to be able to get those real giants. One of the great advantages we have over our Chinese rivals. I don't like using the word adversaries yet, but it's pretty real, is we have a market system and a system of research that is unmatched.

But we haven't done a great job of making sure everybody's committed to this. I'm old enough to remember when it was Sputnik, you know, the Russians launched, put something in space and we weren't even close, we weren't even thinking about it. And yet this country came together with the business community, with government, with our research capabilities, and in five years had 2 million people working on getting someone into space and then to the moon.

I think we need… We've got a Sputnik moment right now. And I think, I think about things like the agricultural extension services. Almost every county has a building or two and has some work that goes on where kids are baking things and learning how to solve rural agricultural problems.

We should be ramping up investments there to make sure that we get more kids when they're in sixth grade and seventh grade and eighth grade and getting them excited about space and science and technology and what that's going to look like, because the level of urgency that this country should feel isn't quite there.

And I think Todd and others in the Senate have done a great job of fanning those flames. We've got a strong way to go. I look at AI. I mean, we've seen what happens if we let these really large companies, and I'm thinking specifically of some of the social networks where we know that they are taking advantage of our kids and in many ways damaging young, young, young women and young men with their self image just by feeding them things that they know are harmful.

But it's a business now of attention. So attention is a product. And I think AI has the potential to be even worse than that. And we need to make sure that we. I think, and Todd's done a lot of work on this, I shouldn't even speak for this, but this notion that we can get some transparency, make sure that people can tell what is generated by artificial intelligence, what's not.

We need a data privacy law that's national right now. We've got different states going out and deciding what is, who does own your data and how do you control that. One thing we don't talk about enough is alliances again to compete Successfully, whether we're talking about AI or quantum computing is a big, you know, I know that there's a big hub in Indiana, there's a big hub in Boulder, Colorado.

These hubs are going to be very, very successful and powerful, but they're only going to be successful if there are alliances that they create as hubs with other hubs and with other countries and the European Union. Some of the research that's going on in Europe is vital to our successful or our ability to successfully address this issue.

And then we also need independent valuations of all the as we go through these changes with AI. And I think it's true of almost all the technologies, microbiology and all the genetic work that's going on. So much is happening, it is the most amazing time. Sometimes I see people get depressed because they feel that the change is happening so fast and they feel that they're losing control.

This is the time when we have to redouble our optimism and our efforts to make sure that we that we don't let that happen and that we get more kids excited and really involved and make it work for them in whatever way that takes. And I also want to just thank Condoleezza.

It's a funny story, but she and Madeleine Albright both had essentially the same father in a funny way, that's not a fair way to say it. But they both studied at the University of Denver and learned their international diplomacy, sharpened their skills at Colorado. So we claim her proudly, even pre-football.

But thank you to Condoleezza here and the Hoover Institution, obviously Stanford, one of the greatest schools on earth.

>> Condoleezza Rice: Thank you so very much. And thank you for really, I think, sharpening now the issues that we need to address as a country. And I will say I'm very glad to say that a couple of years ago, I think there wasn't as much attention here on Capitol Hill.

I do believe people are beginning to do that. I will say I agree with you about Senator Young going all the way back to Endless Frontier when we, we talked about that. So really having a cohort of senators, congress people who care about these issues because they very much are interwoven with education, as you've mentioned, Senator Hickenlooper They're woven with the issues of our economic growth and they are woven into issues of national security.

And so it's all the aspects of our lives, not to mention healthcare and agriculture and so forth. And so I think this has been a wonderful opening for a session. I just wanna say that you're about to hear from, in addition to Amy and Herb, three of Stanford's most eminent scientist engineers I mentioned before, Fei-Fei Li from computer science, and Allison Okamura who works in robotics, and Mark Horowitz who also does double duty as the chair of electrical engineering.

But because we are going to bring them up, I just want to make one other point, which is that universities have been core to our innovation ecosystem. It is why we are who we are in the leadership that we have. After World War II, a man named Vannevar Bush came up with what would be an absolutely brilliant idea, which was that you would leverage universities which had students and it had faculty, and the federal government would fund the very best ideas at the fundamental research level.

So these were not ideas that would necessarily be commercializable. Some of them would never pan out. But an awful lot of them would lead to innovation, would lead to breakthroughs, and some of them would lead to the founding of whole new industries. And the federal government, I think, has gotten the payback that it would have been looking for for that funding.

If you just look out there in the ecosystem, particularly where I live in Silicon Valley or in your Tex center or in your biomedical centers, there's probably a university fingerprint on almost all of it. And so let's not forget that as we go forward to think about federal funding.

So with that, I'm going to call forward our scientific group, plus Amy the political scientist. So Amy Zegart and Fei-Fei Li and Allison Okamura and Mark Horowitz and Herb Lynn, we will now call them forward.

>> Todd Young: Thank you.

>> Condoleezza Rice: Thank you.

>> Amy Zegart: Glad a political scientist could make this panel.

I'm Amy Zegart. I have the great pleasure of being the co chair of the Stanford Emerging Technology Review along with Condi Rice and our Dean of Engineering, Jennifer Whittem. Just a minute about what this effort is and what our mission is. And then I want to have a conversation with our great experts here.

This is the first ever collaboration across the entire university, the School of Engineering and the Hoover Institution. We have 100 faculty across 40 different departments and institutes in engineering, social science and policy experts working together in a multidisciplinary team to better understand what's happening in our labs, what's happening in our companies at the speed of relevance?

What are the policy implications? And how can we better inform policymakers here in Washington, policymakers in states across the country, policymakers in the private sector about what these technologies are, what's hype, what's not, what's next, and to do it in a way that can help policymakers make better decisions.

So this is our second flagship report. We encourage you to read it. We're not done yet, so there's a lot more to come. We have a podcast that dropped last week in partnership with the Council on Foreign Relations called the Interconnect. You'll hear more from Mark Horowitz, who's our first guest on that podcast.

So we are here to help. I know in Washington, people always ask, what's your ask? Here's our ask. How can we help? How can we help you better understand what's happening so you can make better policies that anticipate what the opportunities are, not just the risks? And with that, I wanna start by kicking off with our panel discussion, putting some news into some perspective.

So if you've been following the news, and I'm sure you have, you'll know there's a lot of deep freak about Deep Seq, number one. And Quantum has made the news a lot in the past couple of weeks with Microsoft's announcement of a quantum breakthrough. So I wanna start by asking Fei-Fei and Mark put those events into some perspective.

What's actually a breakthrough? How do you think about it? What are the takeaways of Deep SEQ and the quantum discovery that was announced by Microsoft? Fei-Fei?

>> Fei-Fei Li: Thank you. It's quite an honor to be here with my colleagues. And yes, there is a lot of news about AI.

And Amy didn't miss a beat. Just start with Deep Seek. I think that the context of this is that there has been a lot of progress in AI. I was at the Paris AI Summit last week. I share with the audience that AI is, in my opinion, a civilizational technology because it's the new compute.

And when we say it's the new computer, is that wherever there's a chip which Mark can comment about, whether it's as small as a light bulb, as big as an airplane, there is compute. Compute basically takes data and then turn it into insight, information, decision and action. And AI is the new language of compute.

It'll prevail across all industries, whether it's healthcare, education, agriculture, manufacturing, energy, and all that. And then what is really exciting the past few years, especially with the large language models, is that this is the first coming of age of this technology that's about more than half a century old that is reaching to the hands of consumers and industry.

Large language model has taken the world by storm that suddenly, whether you're a software engineer or just someone at home wanting to look up a recipe can use this language. And many, many, many businesses now are embracing large language-based products and services. So in that context, Deep Seq is a large language model that was open sourced I think now a month or a month and a half ago coming from China.

And what I think is really interesting here Amy, is actually the word open source. This has been a very interesting debate actually is should we open source models, especially these critical models, or not? A lot of company have taken the approach of closed models for good reasons. Because commercial world is competitive, having good closed source models are important for the business.

But an open source model coming not from the US we have had open source model from Europe as well. But really speak of this global moment, of how incredibly competitive as well as exciting this technology is. And it really cuts across borders. It also reminds us of all the not only.

Commercial, but geopolitical issues. So I think that really the context of DeepSeek is the story of an open source story.

>> Amy Zegart: Great, thanks, Mark, quantum.

>> Mark Horowitz: Quantum, o the thing about emerging technologies and new technologies is there's often many hurdles you have to get to before you get to be really useful.

So quantum computing has an enormous promise, but manufacturing them and making them useful is actually a very difficult task. So if you look at Microsoft's announcement that came out recently about their topological qubit was an amazing advance. They've been trying to get this phenomenon demonstrated for many, many years and haven't been able to do it.

So they published a paper in Nature saying, we see it, it's actually there and we can manipulate it. Okay, so it's in Nature, and I'm an engineer and we joke. Engineers joke about papers in Nature about devices that don't really completely work yet because you're just getting the basic physics to work.

And so it's an amazing advance, truly impressive. Don't take that to mean that tomorrow they'll have this amazing quantum computer, because from getting a device and the phenomenon located, there's a lot of work that you have to do to basically build that up. Right now, there are many different possible base technologies in which to build quantum computers.

There's neutral atoms, sometimes called code atoms, there's superconducting, and there's trapped ions, and then there's another group trying to do photonic quantum computing, right? At this point, it's not clear which of those base technologies is really gonna be the most successful, okay? Topological qubits are another way of doing it.

In theory, they have much better properties if you can make them reliable. And they, Microsoft just made this major advance, but they still have to demonstrate the reliability and the other part of engineering, which is still stuff to go. So, like, be careful about taking a scientific advance in one step and extrapolating to where it's going to be, because oftentimes there are other hurdles that happen along the way.

>> Amy Zegart: So I wanna pick up on something that Condi said at the end, which is that let's not forget the innovation model from the 1940s that made the United States an innovation superpower. So, Alison, if you could comment on this. So oftentimes, we don't even realize the fundamental research on which the things we use every day relied from decades ago, the seed corn that was planted decades ago, giving rise to commercial innovations that we use today.

So one example is the cryptography that secures data for better or for worse. On the Internet, relied on decades of academic research funded by the federal government in pure math. That's the story of cubesats, it's the story of artificial intelligence, it's the story of the COVID 19 vaccine.

Decades of academic research funded by the federal government, then giving rise to commercial investment and innovation in the private sector. So, Alison, you run a really interesting lab and you're involved in really interesting robotics projects at the forefront of your field, can you give us a sense of, what is this thing called fundamental research?

How does federal funding actually work in your life?

>> Allison Okamura: Yeah, sure. Thanks so much, Amy. And it's really an honor to be here. So, yes, I run a robotics lab. And a lot of the work that. We do is in the field of medical robotics. And we have funding from a combination of sources, a little bit of funding from industry, but often that funding is very short term, it maybe lasts one year.

And as soon as you get such a grant, you're thinking about where does the next year of funding come. Most of my lab is sponsored by the National Science foundation of the National Institutes of Health, which would give us multi year projects, sometimes up to four or five years, to allow graduate students and even undergraduates to engage in research that is a little more long term and forward thinking.

So one example I can give of a project that is a large collaboration within different faculty and it's very interdisciplinary at Stanford is an ARPA H sponsored project to 3D print a human heart. Now an interesting story behind this is a heart seems like a complicated thing to 3D print, but that's actually where you begin because it's one of the easiest organs in terms of the number of different types of cells that you have to have and the structure that you need to create.

But that still is a big challenge. And so what faculty at institutions like Stanford and around the world do is we gather interdisciplinary expertise, which you can do across universities and across departments and schools, so that you can bring people in different fields together in ways that might be very hard to do at companies or in other types of environments.

And so, for example, for 3D printing a heart, we need to produce millions of cells and have new technology and bioreactors that actually generate the cells that you need to print. You need mechanical eng, like myself, where the role is to actually create multiple printheads that can very quickly, within an hour, print out all of these cells, extrude those cells out and put them in the correct structure so that they can form a beating heart before those cells die.

And then you need Doctors, actually surgeons and other clinicians who understand how do you actually, how does this affect the human body? How can you actually move forward to clinical applications? And all of this takes this interdisciplinary expertise that you can find within universities and the ability to do longer term work.

We're not going to have a commercial product within four years, but what we are doing is planting these seeds that Dr. Ziegert referred to that are creating innovative solutions to these kinds of problems, so that further on down the line we have these outcomes that will get commercialized and will get used by people.

And one last thing I'll mention is that not every project is going to be successful. Sometimes you have to take risks in order to create breakthroughs. And so some of these projects will fail. You'll read about projects that maybe at universities weren't successful, but that's part of the model.

We have to actually take risks and invest in creative ideas, and those are the things that academics can do. And eventually, even if it's a small percentage of them, if they come to fruition, that original funding can help. And I'll leave with one last example that everyone has hopefully encountered.

There was a government sponsored digital libraries project many years ago, and you may not have heard of this, but everyone has heard of Google, which used those search technologies, developed search and indexing technologies developed under digital libraries to eventually become something that has really improved the lives and touched Americans of every type. Thanks.

>> Amy Zegart: So, Herbert, I wanna turn to you next. So if you pick up a copy of the report, and we have a lot of them, so I encourage you to take one with. With you on your way out the door, if you haven't already, you'll see that Our report covers 10 major emerging technology areas.

Now, you can draw the lines in different ways, and we anticipate that those 10 will change over time because science never sleeps, right? But there are 10, and one of the. As a starting point, one of the key things in this report, and Herb really wrote this section, was that you can't look at individual technologies in isolation.

We have to think, think about how they fit together. This is a moment of convergence. There are approaches to technological innovation we need to understand across these technologies. We call that chapter creatively cross cutting themes. So, Herb, talk about some of the most important insights that you got in looking across all these 10 technologies that folks here should be thinking about as they're thinking about what kinds of policies to make.

>> Herbert Lin: Thanks. Professor Okamura already teed that one up. The interdisciplinary nature of the work that she does. To do hard, you need to draw an expertise from a large number of fields. We see this as we look at the different emerging technologies in the report, one of the most striking things is how they all interact with each other.

Not everyone, for everyone, but. So, for example, if you look at energy technologies, for example, AI is an important thing. Material science is an important thing, space technologies feed that. If you look at robotics, we just heard a spiel on how different technologies and scientific expertise play into that.

So the idea that you can focus on one area and say, this is the important area and neglect the rest, that's not a way of winning any kind of race. I think a second thing that we looked at was how to sustain innovation across all of these fields.

And overwhelmingly, the thing that comes out from talking to our faculty and others on it is the importance of talent. Talent can't be produced on demand. It takes a long time to develop talent for the United States to. To obtain talent. It comes from two places. It comes from domestic sources, and it comes from foreign sources, right?

Those are the only two places that, that they, that they come from, and we're failing on both of those. On the domestic side, we have STEM education that doesn't do a great job of producing people that are fit to advance the, that are capable of advancing the frontiers in the future.

Very far, we are. I think the number that you've provided was 30, 34. The US is 34th in the world. 34th in the world on the various STEM proficiency indicators. Massachusetts is the number one state in the country. If it were a country, it would be number 16.

So if everybody were as good as the best state in the United States, we would still be just number 16. That's not a position of leadership, so that's one. And in terms of foreign talent, we make a practice of importing graduate students. We get graduate students from all over the world and when they get their PhDs, we kick them out.

We go out of their way to kick them out. They wanna stay. And even worse, the attractiveness of the United States as a place to come study science and technology is still high, but it's diminishing. You have a student, Dr. Ziegert, about this.

>> Amy Zegart: So I have a researcher who just compiled exhaustive data on every author of the Deep SEQ paper.

And what she did is she also looked at the authors of all five papers released by Deep Seq and she tracked how many authors were on how many papers and what their backgrounds were, where did they study, where did they work, what countries were the, were they in.

And what she found was in looking. This is all unclassified data. What she found in looking at them is there is a homegrown talent story here to DeepSeq, which is over 100 of the 200 authors on that last DeepSeq paper studied and worked only in China, nowhere else.

So the myth is that foreign talent comes to the United States, gets trained up and then leaves and takes that knowledge with them. Homegrown talent in China, because students in China do not need to come to the United States. There are first class universities, technical universities in China that can train them.

There's a lot of other interesting data in there. The ones that got away, how many of them actually came to the United States and then left? How many of them came to the United States and then stayed? To Dr. Lin's point about, we kick them out once they come here.

But one of the myths of innovation is that the world's best talent will always come here. They get a voice and they get a vote and they have options. So we shouldn't assume we're always going to get the world's best.

>> Herbert Lin: And the last thing I wanted to talk about is that as we think about investment strategies and how to promote innovation and the like, we see what we call in the report a frontier bias.

Now it is absolutely important to try to keep on funding and supporting research at the frontiers on designs and technology. But we also want to note that there are many examples of important world changing innovation that can come out with technologies that are not on the frontier. Let me just give three examples.

What you're looking at in the Russian Ukraine war right now is both sides using commercial drones for their weaponry. The Russians are even using chips that you can find in dishwashers to power their drones, okay, to power their missiles. That is, it's not state of the art technology, but their drones are effective, just ask the Ukrainians, okay?

In the history of the United States and in the world, the assembly line was probably one of the most important inventions of the industrial age, enabling mass production. But assembly lines didn't use any new technology at all. What it did was it used technology that was already around and rearranged.

It changed processes to create a whole new way of manufacturing. And lest you think that this is all just in the past, look at SpaceX. SpaceX has a good chance of being able to deliver on its promise of reducing launch costs to outer space by a hundredfold. Why is it able to do that?

Because they have, quote, figured out that if you don't throw away the rocket that launches the stuff and you bring it back down to Earth safely, you can reuse it and you can reduce costs that way. Well, but they didn't invent that idea. That idea has been known for 50 years.

Everyone has known that. So what has SpaceX done? It's adopted a variety of incrementally improved technologies from a variety of different fields. To enable rocket to come back down and land itself. Dr. Horowitz has a great story about some of the technology in there.

>> Mark Horowitz: Well, a colleague of mine, Stephen Boyd, has been working on sort of more mathematical aspects of electrical engineering.

And many a couple decades ago, he got very interested in a particular kind of optimization. That's convex optimization. That was based on math done by some Russians a couple decades earlier. And he showed how many kinds of problems that you need to solve could be formulated as a convex optimization problem.

And if you can formulate it that way, you can build a very efficient solver. So that was stuff done a while ago. More recently he showed that if you have to solve the same problem over and over again, he used that mathematical knowledge and some compiler technology to be able to build a very efficient solvers for that particular problem.

And so what lands the SpaceX rocket when it comes back and has to align to the lander, is that software being run inside to control the gimbaling of all the engines, to basically get back with the minimum amount of fuel, because we don't have very much fuel at the end of the flight to basically land.

So, it's some technology that was invented, 50 years ago, that was kind of made engineering ready 20 years ago, that got tuned up about 10 years ago, that's now part of that whole system.

>> Herbert Lin: And none of that technology, that essential enabling technology, ever made it to the front page of the New York Times.

>> Amy Zegart: Transformational, but not on the frontier.

>> Herbert Lin: Right,

>> Amy Zegart: Right. So I want to make sure we leave time for questions from all of you. So let me end by asking each of you, we talked about, we promised to look at sort of what's next. So if you're thinking of preview of coming attractions in your technological areas in 2025, what should we be looking for? Fei, Fei, why don't we start with you?

>> Fei-Fei Li: I think AI will continue to be very exciting, especially the application of AI models, especially the language models will start to blossom in many, many areas. And we see it to be augmentative to so much work that is out there beyond language.

I think we're going to see very interesting technology coming in. Spatial intelligence for everything that is complementary to words and sentences. It's those models that could understand and reason and generate pixels, videos, 3D worlds, and that will empower so many applications in robotics, in manufacturing, in entertainment and all that.

So it's still a young field, but it's already coming of age and showing so much application and we'll continue to see that growth.

>> Allison Okamura: Great, and I'LL pick up that thread and mention some things about robotics both on the brain side and the body side. One way you can think about robots is their physical intelligence.

They're building on the type of AI that Fei Fei spoke of in order to do useful work in the physical world. Some of the questions are what should the nature of those robots be? So a very popular topic and there's a lot of investment going on right now.

Are a body of robot that is in a humanoid form. And that's very interesting and will be impactful in the long term because it addresses some really difficult technical problems. But I think what we'll also see is the growth of robots that are not anthropomorphic. They are designed to do more a narrower set of tasks, ones that people are specifically looking for and ones that ultimately do tasks that no human can do currently like the 3D heart printing example.

The goal there is to use robots as augmentation or assistance or to do novel tasks that have never done before as opposed to replacement. And I think we'll see more and more growth in those types of robotic activities in the coming years.

>> Mark Horowitz: So, I think on the semiconductor front, you'll see continued evolution and improvement in the base computing platforms.

I think much of the push in the next couple years is going to be on communication because there's a lot of issue on how you get the data back and forth. There's a looming or there's a present energy problem that the solutions we're doing are consuming too much energy.

I think that's going to be a joint application hardware co optimization that's going to move to try to push those energy costs down. So I think we're going to likely see a bunch of sort of cooperative work between the software and hardware designers to try to build more efficient systems that we need for the future.

>> Amy Zegart: Herb, did you want to add anything?

>> Herbert Lin: I did want to add one thing. Just because the person who should be up here doing it isn't here. It's Drew Endy. Drew Endy is the creator of the biology content of the report, and I urge you to read it.

The argument there is that the promise of biology now for the future is sort of where semiconductors were in 1960 or so. And so, if you can imagine that there's a huge opportunity to be exploited by taking advantage of biology, synthetic biology and biotechnology. Only if we take advantage of that.

Only if we take, take, take the opportunity. We heard earlier from a senatorial panel that that the Chinese are, are, are making that bet, investing huge amounts in Biotechnology as a. Has not, not just for, for health but, or agriculture, but across the board significant investment in biotechnology and our investments on that kind of building that kind of biological infrastructure for progress in the future, you know, sort of pales by comparison.

So there's a lot, there are many number of studies which suggest that over 50% of the GDP could be affected by biotechnology. And this is a big deal, but it's aspirational. That is, we could get there, but if we blow it, we won't.

>> Amy Zegart: And from a geopolitical perspective, just to end on a political science note, before we open up for questions from a geopolitical perspective, what this means is you can have localized supply chains.

If you can biomanufacture something, you don't have to have a factory someplace else. As Drew likes to point out when we walk across campus, he's like, aren't trees amazing? They can grow anywhere we can. The leaf development system of a tree is unique to a particular place. Now imagine you can biomanufacture anything anywhere.

We're not talking about supply chain vulnerabilities to China anymore. That's the kind of forward thinking opportunity that the bioeconomy could have. I think I channeled Drew Endy appropriately there. Okay, so with that, let's open it up for questions. If you have a question, I think there are microphones running around.

Malaysia has a microphone, so. And we have some other. And Jack has a microphone too. So why don't we start with this gentleman in the front and then I'll try to alternate over on this side too.

>> Guest 1: One of you must have an answer. I've been annoyed with electric cars coming out and they don't recharge themselves.

How did we do that? We had cars that had alternators, recharge batteries. You have the drive end of the motor. The other end could be using, being used to generate free electricity to power the other side. It has to have applications across everything besides electric cars, that we now have the ability, with batteries or storage and smart energy, to actually have the same motor provide the power for the motor.

>> Mark Horowitz: Yeah, so basically, electric cars do regenerative braking, which means that when you're trying to slow down, you use the same motors to regenerate energy back into the pack. But the laws of physics are kind of obnoxious, and they basically say you can't build a perpetual motion machine because there's always losses in the system.

So, by starting the motor up, you actually regain the momentum you put in when you slow down, but you've lost energy along the way. And so you could say, great, but why don't we just put solar panels on the top of the car? Because, the sun's coming down and we could do it that way.

But if you do a calculation of the energy that's coming from the sun on the size of this top of the car, you'll quickly realize that you can't recoup the energy. So for now, we end up having something that the energy that you do in a trip is somewhat consumable until we come up with a better idea.

>> Amy Zegart: There was a hand over here. Yes, woman in the front. Just wait for the microphone.

>> Guest 2: This is for Dr Horowitz about quantum computing. I understand that Finland is one of the countries at the top in terms of quantum computing, but also here in the United States, I read that Google already reached the level of 1,000 qubits.

And they did it because they were, initially the problem of getting those qubits is because the lack of stability, and once you reach the best, probably lowest temperature, they're already going -40 something. So they have been able to stabilize the qubit, and therefore they reach already 1000. So I would like you to comment on that in terms of whether we can get more, of course, more qubits in the system and what you can do with 1000 qubits so far in terms of quantum computing, thanks.

>> Mark Horowitz: It's a great question. So in quantum computing, basically the measure of sort of the complexity of the machine you have is something called a qubit. It's kind of analogous to a bit in a computer but different. It's quantum and that makes it way more powerful. So today many people talk about different size quantum computers and people have talked about.

And IBM has built machines that have very large number of quantum bits and other people have advertised this. But there's another issue that in addition to the number of bits you have to care about, which is what's the fidelity of the operations you can do on those bits.

And you need the combination. If I have a thousand bits but every time I do an operation has a 1% error, I can't do more than a hundred operations before I've kind of gotten a hundred percent error. Well, it doesn't like for people who do math, it doesn't work that way.

But I just don't but well, so it, it turns out that I've been tracking kind of the best quantum computers that have been announced by various people. So announced, I don't know what's going on, secret. So right now the highest fidelity operations that have been reported are 99.9% and they're in computers that have about 30 qubits, okay?

Google has announced and demonstrated chips that have on the order of 100 qubits and their fidelities are 99 point I think five or seven something. I don't yeah, I don't have the data in front of me. People have demonstrated trapped no, people have demonstrated cold or neutral atom configurations that have 1,000 qubits.

So I don't think that was Google. I could be wrong, but I don't think so. But they haven't been able to do operations on those qubits yet. So there has been advances and I've been tracking this since 2018. I can show you a little plot of the progress, right?

It is up and to the right. We're making progress. It's not quite as fast as some of the advocates have said we're going, but there is steady progress. But we're still at the point where the machines are relatively small and it's not like those machines can do anything useful yet.

So that's kind of where we are today.

>> Amy Zegart: So our experts are gonna stick around for drinks and food so you can ask them your qubit and other questions at our reception. I just want to end by saying one thing which we didn't talk about yet, which is if we're thinking about what is the most important thing to power all of the scientific advances that these.

And other experts are working on today, that other than talent, it's compute, computational power. And I'll leave you with two numbers, 300 and 350,000. Princeton last year had to go into its endowment to buy 300 of the most advanced Nvidia chips, which power scientific discovery the same year.

Last year, Meta announced they were buying 350,000 of the same chips. You want to know what the national infrastructure is for national security and economic development today? It's compute. It's what the highway system was in the 1950s that Eisenhower developed, national security and economics. It's what the Strategic Petroleum Reserve was in the 1970s, economics and national security.

The national infrastructure to guarantee both our security and our economic competitiveness through scientific discovery today is national compute. And if you want to know what Congress can do to ensure that the United States stays at the forefront of leadership in science and technology in the world, it's compute.

I'm gonna end there, and thank you all so much for coming and all you're doing.