What Happens If We Stop Funding Innovation?

Arait Prabhakar: Imagine a future where instead of decades for a new drug we could develop new drugs and months. But if we want the Steve Jobs of the future we have to keep investing in this technology base that makes it possible for those amazing commercial products to happen. The deeper damage that's being done is generational in nature. This administration came in and basically just started gutting this capacity we've built over 80 years to have a federal support for science and technology that then supports the entire country's innovation capacity.

Dan Koh: How does America outcompete the world in science and technology? This question has been top of mind for our next guest for decades. Aarti Prabhakar was the chief science and technology advisor to President Biden and previously led DARPA the Department of Defense Agency responsible for everything from vaccines to the Internet from GPS to touch screens. Director Prabhakar talks to us about how the Trump administration might make us less competitive. What we need to do to turbocharge innovation. And what this all means for artificial intelligence. Let's swear into the people's cabinet Aarti Prabhakar. Director R.K. Prabhakar Welcome to the People's Cabinet. It's so good to be here with you Dan.

Dan Koh: Well we're thrilled to have you because you were President Biden's science and technology advisor. And there's a lot of questions around science and technology these days under this administration. We have a lot to cover. But I want to start with A.I. because I think there's a lot of excitement around it a lot of fear around it. And so my first question to you is are we falling behind China. And my second is how do we win.

Arati Prabhakar: Yeah that's a great place to start. And I'll just I'll say that when I first came to work at the White House when I first got to meet you and come to work for President Biden I showed up in October of 2022 on chat ship hit the world in November of 2022. And so while the science and technology issues are incredibly broad I ended up spending a disproportionate amount of my time on A.I. And I'll tell you President Biden and Vice President Harris got it exactly right because they said this is one of the most powerful technologies of our time and our job not just the administration but our job as a country is to make sure that we manage its risks so that we can seize its benefits. And that's the whole ballgame is how do we navigate this powerful technology. And that's really ultimately that's what it means to win is to make sure that this powerful force that you know people have brought forth in the world that we use it to change people's lives and to improve the human condition. And part of that absolutely means outpacing our our global and geopolitical rivals. If you look at what China and other countries are doing everyone is racing to build a future with A.I. that reflects their values. And unfortunately some of their values are not our long held American values of freedom and individual rights. And so that that just to me really underscores why it's so important that we do the work to make sure I serves the needs of the American people protects against that we do the work to protect against its harms. But ultimately I want to use this technology to build the future that we really need for this country with better health and more opportunities for every child and more robust national security. All of our great aspirations are places where we can harness A.I. but that's the work.

Dan Koh: And can you take us behind the scenes. Because I think most people now have used some version of chat jpt you know perplexity all of these different things. But I don't think people really understand how it works. And I don't want to get into like 30 minutes on what a large language model is. But if you wanted to but if you could explain in layman's terms what's happening behind the scenes because I think most people don't really understand what's happening.

Arati Prabhakar: Yeah. Yeah It's a different kind of information technology that everyone's getting their heads around. At the end of the day what artificial intelligence this generation of artificial intelligence is it's information technology that instead of programing to do specific things you train it on a lot of data. And the technology capability that AI represents is the ability to start seeing patterns in the training data. So you showed a huge amount of text It starts seeing patterns. And what has the astonishing breakthrough of generative I was it got so good at seeing those patterns that it could reproduce text or images depending on what it was trained on that was so realistic that it was that actually was useful to people. So when you're talking to a chat bot actually often I think people feel like they're talking to a chat bot and it's like a search engine It's going and pulling facts or stringing information together. That's actually not what's happening. What it's doing is synthesizing a word at a time or even a syllable at a time an answer that that that that machine is estimating is going to be the best answer to the question that you've just asked it. So it sounds pretty simple and it doesn't seem like it should be that powerful. But of course when you see what it's capable of today it's astonishingly powerful.

Dan Koh: So the interesting part of that is when you look at how China just does business in general and how at least we aspire the United States does business you can imagine the ethics behind taking all of this data. And if China has fewer ethics to do that then you can imagine a scenario where China is saying we're just going to take everyone's private data we're going to ingest all kinds of information that maybe the United States wouldn't be seen as kosher. Wouldn't that dictate potentially a more powerful end product on the Chinese side than our side.

Arati Prabhakar: That's that's that gets right to the issue of what your core values are and what you have to respect in order to build a future that looks like the values that you really care about. And yes what you described is exactly right that if you are indiscriminate and sweeping up all the data and you don't care about protecting people's privacy and in fact you are a surveillance state that that doesn't permit individual privacy See then that's where that road goes. And so the the great challenge for us and the prospect ahead for us is how to build powerful technologies but to use that as well and really responsible ways while rigorously respecting privacy while respecting the fact that we have laws on the books against discrimination in housing and lending and in health care and in criminal justice. And if I is embedding biases from the past and being used to make automated decisions that violate these not just our laws but our fundamental values about equal treatment in the country those are exactly the issues that have to get wrangled so that we can use AI in ways that really And so and so how do we do that if part of this is based on the scale and type of information. Yeah While respecting there are ways to do that and I think that's number one A lot of those issues right now are in the courts and are getting sorted out about what kind of data can companies train on. Is it fair use under copyright to build an AI model that trains on a bunch of copyrighted information. It's a new way of using information. And so it's it's new territory that still has to get sorted out. And today what is happening is AI developers are saying very much what you're saying which is if China and others train on all the data and we can't then that's going to hold us back. It's bad for for international and global competition. And creators of content are saying well wait a minute if you are training on my information and you can now synthesize something that copies me you are undercutting me in a way that was never what copyright was intended for. And that's that's exactly what has to get resolved. I think you know the most hopeful scenario I can see is it gets resolved by finding a way to to provide compensation to creators for the value that companies build on top of their creations with A.I. models. But we're just at the beginning of sorting that out.

Dan Koh: What about access. So both access on the information side maybe being able to restrict access that China has to feed its models as well as on the hardware side which I know you have done a lot of work.

Arati Prabhakar: Yeah well I want to back up from your question and think about where we are with respect to China because a lot of the focus in the conversation has been about how do we slow down a competitor like China when we our companies design and build the most advanced CPU chips which is what you need to build A.I. models. And because we have had this enormous surge in capability from American companies and I think first of all I'm a little concerned that that that's a model that works the model of I'm going to slow down my my competitor. That model works when you are the only place that's doing leading edge research. And for many many years China did not do leading edge research. And so you know the term R&D has research and development. And for so long for decades America was the only country that had just an unbelievable lead in all areas of research other countries developing economies including China. In that period they did R&D but they mostly did did. They took other people's research our research and they commercialize it. They turned it into businesses and products and they and they put it to work for example in AI and which they used extensively even before this current era. So in that world it's fair to say well you can't have my basic research because you know you're going to outcompete me. But that's not where China is anymore today. If you look at Chinese the country the People's Republic of China in Hull now spends nearly as much on R&D as the United States does. We've been the undisputed leader for many many years now. They're getting close and total spending. And to put it in perspective China spends 40% more than all of the European Union put together on R&D. Right. So a very intensive level of investment within which is a lot of basic research. Last year when the Chinese government said they were going to continue to increase their contribution to R&D they boosted overall R&D 10% they boosted basic research 13%. So they are dead serious about research. They're becoming really good at research in virtually every field. So now it's we're no longer in a world where it works to say you can't have my secret sauce because now we have a competitor that can make its own secret sauce and so we can do export controls. Those things slow China down a little bit and that has happened. But that's not a winning strategy by itself. And so when you have a great competitor if you don't run harder and faster you're just not going to win. And that's exactly what we have been trying to boost over many many years here. We've got a competitive industry but how do we make sure we're educating the students and doing the basic research in our universities. We we're making progress. And that's what's now getting dangerously undercut.

Dan Koh: And so look I think a lot of people know I at this point still on a relative surface level right now. I'll put it in Where should I go for dinner tonight and I'll spit out something or they'll do research basically as a replacement for Google or another search engine.

Arati Prabhakar: Yeah I do that too I do too. In fact for this interview I made sure to do some deep research for the function.

Dan Koh: But my question is what are some of the applications of A.I. especially for science and technology perspective that people may not realize that is potentially revolutionary science.

Arati Prabhakar: Yeah and I think this this to me this is actually one of the areas I'm spending some of my time on now because a lot of the conversation about AI is basically about productivity business productivity. Now I hope that's more exciting than I can write more emails faster than you can use your AI to read my emails faster and the like. Where are we going with that. So and I think that there will be perfectly good business productivity applications. That's what the market is excited about. That's what the market is investing in. That's what companies are building. That's great. I want productivity. It's going to boost our economy. It has to be done in a safe and and responsible way. That's why we have to deal with the safety issues and the discrimination issues and the privacy issues and the workforce impacts. But that's that is nowhere near the most exciting part of AI to me. What I'm excited about is how I can bring the changes that are going to change our lives and really deep ways. I think about the fact that we have thousands of known diseases today for which we have no medical treatment at all and yet we only approve maybe 30 new drugs in a year because it's such a slow difficult process. You have one that comes to mind when you think about oh the one that the one that keeps so many of us up at night is Alzheimer's. Right. And it's just so if we're if you're lucky enough to grow old you have to live with the fear that Alzheimer's will become part of the end of your life. And yet there's no real prospect for anything that's got that's a meaningful treatment for Alzheimer's. That's one of many many examples. There are so many rare diseases that afflict not that many individual people. When you aggregate how many people have rare diseases it's a huge part of the population. And those are they're not profitable to go after because it takes billions of dollars and decades to develop a new drug. So that's where we are today. One of the great visions for A.I. is that A.I. models not that are just trained on text but A.I. models that are trained on biological data and on clinical data about how a body reacts to medical medications of different sorts that those kinds of AI models could help us generate new candidates for drugs and get through the process of developing new drugs. So people imagine a future where instead of decades for a new drug we could develop new drugs and months. That would be astonishing or it would change. I want my girls to grow up and never have to worry about some of these diseases that we just think are a fact of life today. That might be possible. It's not going to happen without public and private sectors both doing the things that they need to do. And so when I think about the really big dreams for A.I. they are things like better drugs faster. They are things like closing educational gaps for our kids or delivering a better weather forecast for every person on the planet. In a time when the climate is changing in these dangerous ways those are huge dreams. They're why I get excited about AI. And I also know that to make those come true we're going to have to do it's not just you know some tech entrepreneur is going to start a company or some scientist is going to train a new model. There are huge public roles in basic research and the foundation of research in regulation in approving drugs and in the the sensor data that it takes to feed weather models. And I'll tell you when I think about this moment that we're in right now Dan for science and technology. Though it's painful to watch the destruction of a lot of our capacity for these kinds of huge advances because my whole professional life I have seen the public sector and the private sector in America do astonishing things things that have that have changed our lives and the way the way we live and flourish today. And so I know how big the possibilities are. And that's what that's what I we've now put at risk.

Dan Koh: Okay. So I think most people understand kind of surface level ChatGPT is meeting most needs that people have day to day. It still has you know some they call it hallucinations when they're presenting the wrong data or something like that. But if you were to kind of think about the way I think about it is it's probably out of six or seven out of ten of what people actually want it to do on a daily basis. Where are we If you apply a 1 to 10 scale on on the medicine or the health research that you're applying.

Arati Prabhakar: Yeah Much much much earlier. These are these are visions for a future that's going to take a huge amount of research. So you know on a scale of 1 to 10 where one or two. But but we're not at zero. And this is I think the key point is when you see the advances in AI technology in large language models and text and images and then you look at Alpha Fold four which won a Nobel Prize last year that was a that's an AI model. That's about proteins that actually can predict how proteins fold which was one of the long unsolved prior problems and research. So it's a scientific application but it starts showing you the power of AI not just for text and images but for things like molecules. Right now it's starting to get interesting for medicine and you can start to see the direction the research could go and the barriers right. So if this is going to have to if we really want drugs in months rather than decades we're going to need to find a way to train AI models on a bunch of clinical data that is tightly controlled by pharmaceutical companies. So that's a barrier to overcome. If you want to just come up with cool molecules that's not so hard. If you want a medicine you have to go through clinical trials and the FDA has to sign off and say certify for the American people that it's safe and effective. That's a big deal. And so the FDA has to be part of believing that AI technology can be part of this. So these were the pieces of the puzzle that we started putting together over the last couple of years. And again we're this is something that's you know it's going to I think it'll happen in my lifetime but I think it's really going to benefit the next generation. But you've got to start now because the path ahead is long.

Dan Koh: You talk a little bit about I don't want to get deep into FDA approval processes but I think most people both understand that there needs to be some safeguards and process to make sure whatever drug someone's taking is safe but also probably doesn't have a lot of faith in federal government in general about not being innovative and streamlining things. Right. So as you point out I can imagine a world in which we have ready the technology to not make it take however many. How many what's the average drug cycle let's say a decade. Okay. Yeah that maybe could take a year or two or maybe whatever. So you talk a bit about the work that you did under the administration to prepare for that. And do you see any continuation of that under the Trump administration.

Arati Prabhakar: Yeah I very much hope it continues. President Biden started a new health research agency called ARPA H. It's modeled on DARPA in the Defense Department. And he said look we have to change American health outcomes and be serious about innovation to make that happen as we are about national security. And I thought it was just one of the most powerful moves that he could possibly make. So ARPA H did exactly this work that we're talking about. They said this prospect of A.I. for drug development is out there but how do we actually get the ball rolling. And what I thought it was actually genius because what they did they took one intractable problem which was getting industry to allow training on their their sacred data that they hold so tight. And then the other intractable problem was you know it doesn't count if you don't persuade FDA. So what they did was they started working with FDA on not the whole drug development process but just the the very start phase one clinical trials are trials where it's the first time the drugs are given to human beings not to animals but to human beings. And phase one trials are just to establish whether or not these these new medicines are toxic. How do they function in the body. How does your body absorb this medicine and does it do any harm. You got to know that before you go on to the later stages that are about how effective it is at treating the disease. Well even that first stage today can take a year or two years of a new of a company with a new molecule arguing with the FDA about whether there's enough confidence to even go into humans in the first place. And I'm glad the FDA is arguing to keep us safe to go into a safety trial. But that's Part of what makes it so slow. So what what arpa h did was get our fda to the table. Got some really forward looking people at FDA to agree that they want to work towards an AI model that that that could be proven to be good enough at predicting toxicity that after you would use it instead of lab experiments they would use this AI model as a quick start into the phase one process. So that could cut out let's say a year of the process. That's a pretty big bite to start with. And the magic of it to me was as soon as FDA said they might really go down this path all of a sudden pharma companies became a lot more interested in sharing their data because the one thing that they can't ever figure out how to make progress on is making FDA go faster. Right. And so when you sometimes when you throw impossible problems at each other they start canceling out in constructive ways. This might be one of those examples but none of that happens without the government right. Like that's it. It takes these institutions to do that kind of work.

Dan Koh: And so can you talk about this in the context of a Trump presidency both on whether you're seeing the kind of collective innovation in this particular area but then also just at a higher level. Right I know very little about how public funds really spur innovation like you do. So I'd love for you to talk to us about that. But I think just in general there's a sense or a narrative from the Trump administration that government's bureaucratic it's not innovative. We should just cut a bunch of funding and and bloat or whatever and just give it directly to the private sector and kind of step back. Yeah I'd love for you if I'm going to work I would love for you to give your overview on that and how we should be thinking about that in reality.

Arati Prabhakar: Yeah there's just a deep disconnect in what this administration is doing and what it's saying. That's true in many areas but specifically on science and technology and innovation this administration came in and basically just started gutting this capacity we've built over 80 years to have a federal support for science and technology that then supports the entire country's innovation capacity. They've done that through trying to remove and now they are removing people from the people who run federal R&D agencies versus the National Institutes of Health the National Science Foundation NOAH. That takes care of our weather and many more things. They've they've pulled back funding. They've gone after universities in a broad brush by trying to undercut indirect costs which sounds like a wonky accounting thing. But this is the backbone of how our universities have become such a huge actor and such an important actor in American innovation. And then they've gone after specific universities that we've seen over and over again. Columbia's was the start of that that those actions are they are they are harmful today because of individuals whose work is getting stopped. They're harmful today because of clinical trials that are getting stopped in the middle. But the deeper damage that's being done is generational in nature. So when I look what I really worry about is one university after another has stopped or slowed graduate admissions and in in research fields because because of the great uncertainty about what these cuts mean that means we're going to have we're going to have a generational gap even if we were to get back on track immediately. We're starting we're already damaging our prospects for the future so that that's that's what they're actually doing. And then what they're saying is that we need to be dominant in science and technology. And yet you simply you know we as a country simply can't be dominant or lead without this federal foundation for research. I've spent half of my life in the private sector and half in public service and I I see what each side does. But if the government doesn't do what it has done for 80 years the private sector can't and won't fill the gap. It won't fund the long term risky research. It won't fund this foundation to educate students.

Dan Koh: Take us through an example that is most kind of prominent in your mind of that.

Arati Prabhakar: Sure. Yeah. Yeah. And it's it's it's not always holding hands and singing Kumbaya together. It happens in stages. I'll tell you a couple of very different stories. One of the really moving experiences in my life I worked at DARPA early in my professional life in Defense Department Defense Advanced Research Projects Agency that was started in right after Sputnik was launched. First time human beings have put an artificial satellite on orbit. Unfortunately the human beings who did that were Soviets at the height of the Cold War. So you know the United States freaked out. And then we did some really constructive things. We really doubled down on federal support of R&D. We pushed our kids to get STEM educations. And and you know it really has had I don't think was a good moment in the end for the Soviets because we reacted so smartly to that. One of the things we did was create an advanced research projects agency in the Defense Department to prevent that kind of technological surprise. Well the way you prevent surprises you go create surprises of your own. And so I worked at DARPA early in my career left for two decades and came back and got to lead it. And and that was an organization that was phenomenally effective and functional and did big things when I was there. Initially I came back two decades later and I will tell you it was even better at doing its job. So that was really really enlightening to see. I've never worked anywhere in private or public sectors that got better over time but they are aggressive about weeding out bureaucracy and doing their work. Very mission driven organization. Early in my career when I was there I worked on semiconductors in electronics. I started the semiconductor office while I was there. Many years later I heard a soldier who had served in Iraq and Afghanistan talk about the fact that when he went on patrols what he put in his backpack was the combat gear and the navigation gear that he had. He said I carry that instead of food and water because it was more essential to my survival. And I thought when he talked about that I remembered the work we were doing on advanced semiconductors that made it possible to have comms and nav gear. That was this big that you can carry in your backpack rather than you know the size of a truck that you had to have in a convoy. And when that soldier told me that story it felt all of that work I had done all the dollars I invested on behalf of the American public felt worthwhile because we had done this thing that changed his ability to do this impossibly. And I'm sure not to mention the fact that his ability to communicate with his fellow soldiers changed the mission completely completely. So that was very worthwhile. And that is the exact same technology base that we and I didn't build the products that went in his backpack. What I did was help invest in the research that made it possible for a bunch of defense contractors and commercial companies to build products that ended up in his backpack.

Dan Koh: And how does that work. There's a certain amount of money that comes from the government. How much of it is done at DARPA itself and how much in that case.

Arati Prabhakar: In that case all the money all the R&D that's done through DARPA happens out in universities and companies and national labs. The people who DARPA's only 200 people to about 200 government employees. But their job is to is to identify these seemingly crazy possibilities out ahead and then rally people and fund them to demonstrate that they're possible. And the idea is that if you're in video or you're one of these companies you may not have the capital to do the kind of R&D that will invidious got all the capital they need. But but often it's startups often it's university projects that then become startups. So it can be all of those and you need all of them because they all have different skills they all have different missions. But what a Draper program does is rally them to do something that they really weren't going to be able to do by themselves. So that was what I told you was the military side of that story. But you know every time I pick up my iPhone I think about I think about the chip that's talking to the cell tower because that's gallium arsenide technology that I first came to DARPA's to work on in 1986. I think about the lithography that made those chips with these atomic scale transistors on it. That's the lithography technology that I that everyone thought was impossible when I started a program in that area and around 1990. And you know it's so meaningful to me to see and I want to just be clear it takes nothing away from the genius of Steve Jobs right. Who put all of that cool stuff all those components together in a in a product that's so addictive. Like I'm itching a little bit because I don't have my phone in my hand. So so kudos to him for doing that. But if we want the Steve Jobs of the future we have to keep investing in this technology base that makes it possible for those amazing commercial products to happen. And so that's there's a couple of examples.

Dan Koh: I think it's just really interesting for the listeners to know some of the other examples.

Arati Prabhakar: Yeah I'll tell you another one. I'll tell you one That's probably the most meaningful to me. I returned to DARPA in 2012 as the director and I was like a kid in a candy shop. I was catching up with all the program managers to find out what what the bleeding edge thing was that they were working on. And one of my program managers in 2012 he was a geneticist an M.D. Ph.D. and an Air Force colonel because that's what that's what you got at RPA. And he said there's going to be another pandemic. This was obvious to everyone in the infectious disease community because all the conditions are ripe global travel and more people eating meat around the world and meat markets that are unregulated. So he said we know there's going to be another pandemic and one The many problems that we face is that it takes years or decades or never to develop an effective vaccine. 2012 he said But there's this research in many and it could be the basis for a rapid response vaccine platform. And I just met a company called Moderna. It was a tiny startup in Boston. And Moderna when it started was focused on using money therapies for cancer.

Dan Koh: Can you explain like I'm five what messenger RNA means?

Arati Prabhakar: Yeah it's a it's a messenger RNA it's a fragment of nucleic acid sequences that the that for a variety of reasons people much earlier researchers Katie Couric Katie Couric Andrew Weissman had had actually speculated postulated that it could be a very effective way to deliver a particularly clean and effective vaccine rather than all a very complex slow to develop ways that we currently have for developing vaccines. But their idea was considered fringe that it wasn't it really was sidelined I would say because it was so far from the mainstream of vaccine thinking. Dan Wootton Dorf the program manager at DARPA whom I mentioned saw that and many other related technologies and said Well look we should give it a try. So this is crazy. It's so outside the way we think about making vaccines but it's in the realm of the feasible. And RPA exists to try things right. And it's okay It's okay to fail. It's not okay to try for a small advance. That's what real failure looks like. But if you're reaching for a big advance it's okay to take some risk. And if you fail you know you go on to the next thing but it's such a big deal if you succeed. So it was exactly the kind of DARPA type of project. So Dan set out to demonstrate that and more in a vaccine could actually work. He had found Moderna which was working on M RNA for cancer. And you know if I had been there venture capitalists I too would have said I don't want to work on infectious disease. No one wanted to work on infectious disease. That was a terrible way to you Couldn't rely on that for a market. Unfortunately you can always rely on cancer. It's the tragically reliable market and it's a different development It's a different development path because their particular You're trying to induce an immune response right. That's what a vaccine is all about. So we were with a big vision and some capital to invest to you know to fund the research. That's how we persuaded Moderna to start to You said we'll give you this funding $25 million as I recall was the size of the first for 2012. Dan had already started working with them just before COVID. This is way before COVID but we knew we knew the world was full of infectious disease and that the time would come when we really needed to move quickly. And in the military by the way we send our military service members to all parts of the world where they are exposed to infectious disease. And so it's a specific military problem to protect soldiers and war fighters. And then of course the larger national security concern would be what if there's a pandemic which turns out to be worth worrying about. So this is 2012. Dan had started before I came. We deeply accelerated that program. It was Madrona. It was many other branches of research to get this vaccine into the body and to have it lasts long enough to do its work you have to encapsulate it in a little And a nanoparticle a lipid nanoparticle there's a whole field of research around that. The long story is that when we started it was people thought it was dumb. They thought the word naive got used a lot. It was really a crazy idea. And then the turning point came when we got as far as Phase one clinical trials. It was for the first ever mRNA vaccine. It was not for COVID COVID didn't exist but there's always an infectious disease to practice on. Unfortunately this I think was a COVID vaccine for chicken gonad a mosquito borne infectious disease. And we only got as far as the safety trial the phase one clinical trial. But once we got into a human being and in an approved FDA trial and in addition to proving safety we also saw immune response in that human being. That's when the infectious disease community went from That's crazy to wait a minute That might not be crazy. And so that was about 2017. That's when NIH started working with Moderna. Fast forward to the pandemic and in 2020 when as soon as Barney Graham and his team at NIH identified the particular spike protein on the COVID vaccine that was what we needed to make a vaccine to replicate. They did I mean a lot of other things had to happen. They had to do that miraculous branch of work. But 42 days after they identified that spike protein Moderna was able to ship doses of COVID 19 vaccines off for the first clinical trials. And that's that was a big part of how we got the You know the fastest most effective vaccine development in the history of.

Dan Koh: And where do you think it would have been if Moderna wasn't in the picture at that time. Pfizer obviously was filling a lot of the void but Yeah. What would have.

Arati Prabhakar: Yeah So this is really what happens when you build a technology base. You work with individual companies and university researchers but you're building an entire field. And when you look at the technology that Pfizer drew on working with Biontech which was their their development partner in Europe all of that technology base that is the community that DARPA funded not by itself but as but I will tell you it's just that whole technology base would not exist at the level of robustness that allowed for this rapid development.

Dan Koh: And we don't have to deep dove into these but internet, GPS, other innovations, can you just give us an overview of these these examples you just gave just on a high level.

Arati Prabhakar: Yeah People always thank me for DARPA inventing the Internet and I'm like you're welcome I was nine years old when that happened. But yeah I mean that was that started with some researchers at DARPA thinking about this crazy idea that you might connect computers remember computers or this thing down the hall in a refrigerated room and you used punch cards right. But they had this idea that if you could share computing capacity you could start opening up the whole information revolution. And lo and behold literally from DARPA projects first the ARPANET an experimental network to connect computers and then the Internet protocols that are still the the communications protocols on which the entire Internet has been built. Those were developed under DARPA's stewardship and then brought into the world because that was we had a bunch of researchers that were communicating with each other and doing email but it was just sort of tucked away in this research community and that was brought out into the world through the National Science Foundation work that Al Gore did with the Reagan administrator administration to to get it out to the world so that companies could start building on it. And that's where the entire Internet revolution all the companies that built the backbone of that all the companies who built services and then social media on top of the Internet polite sparks like that that go a long long way.

Dan Koh: And so I want to talk to you next about this administration. You've made some allusions to it but how have you seen their priorities set around this kind of funding of innovation. Are they still and by the way is Congress does Congress fund ARPA. Is it sure It's like any other government agency that's appropriated authorized and appropriated by Congress. But how do you think their perspective on public R&D actually is and what what concerns you or even excites you about the next few years under this administration. With that perspective.

Arati Prabhakar: yeah I'm quite concerned. And you asked about priorities and I can't discern priorities. What I see is really mindless actions that are broad brush that it's hard to it's hard to attribute any purpose to them other than just raw destruction. The actions are the removal of key staff and the removal of employees at federal R&D agencies. The the pullback of contracts or the freezing of contracts the the the freezing or the pullback of expert advisory panels for the review process that goes on. There's something of a witch hunt for contracts and research grants that use terms that might in some way offend the sensibilities of this crowd. Words like women Words like Pronouns I mean I don't really it's it's it's if you had written a novel about this you wouldn't believe it. It's so un-American. What is going on. It includes an attack on universities first with an attempt to cut indirect rates which sounds like a wonky thing but is pretty fundamental to the operation of the university component of our innovation.

Dan Koh: So what does that mean exactly is it about the way we account for the things that a university has to do to do a research project that's funded by the National Institutes of Health or any other agency?

Arati Prabhakar: The the way the accounting is done is there's a on top of directly paying for the time of the researchers There's an amount of money that's also allocated to make sure that the building stands up and the janitorial staff can clean things and that the lab can be safe and that you have the compute to run the data that you need to go with the lab. And that's just a lump sum that gets negotiated with each university. And by law it's something that cannot be arbitrarily changed. But this administration has tried to arbitrarily change it and cut it dramatically which the effect of which would simply be to put universities out of the business of doing this this key role that they play in our foundation of research and education. So all of those actions taken together really are having this effect of gutting and even the actions that are getting rolled back are creating so much uncertainty that they are they are damaging our innovation capacity because they're undercutting federal R&D. So that's that's what I see happening. And then the language of the stated purpose of from this administration is that they want to dominate in science and technology which has no resemblance to the actions that they are taking.

Dan Koh: So is it your estimation that these actions in this posture is making us less competitive with some of our global competitors like China.

Arati Prabhakar: Absolutely.

Dan Koh: And so what is the alternative vision. What if you were still at ARPA or you were still if you were advising President Trump what would be your advice to him.

Arati Prabhakar: So I can't get my head around that one. But okay. Well look I mean here I think it's really important to be clear. The reason we fought for federal R&D was not because we were doing it perfectly. We fought for it because it is utterly indispensable. And if it goes away the private sector is not going to do that work. But the work that I spent so much of my time on at the White House running the Office of Science and Technology Policy in President Biden's administration and then before that when I ran DARPA before that when I ran Nest. But the the work is to take this phenomenal capacity for science and technology and innovation that the most powerful capacity the world has ever seen but a capacity it was built in the last century. The work is to aim at at the great aspirations that we have as a country today. And some of those aspirations it's about refreshing you know the aspirations that have always driven public investment in R&D national security economic growth and better health. And so those are those are forever aspirations for any society. And today they mean for example that we want to make sure that there's economic growth and opportunity not just on our coasts but in every part of America. And so very much as part of President Biden's vision for creating opportunity across the country and revitalizing manufacturing across the country a great example of something that happened on the R&D side was that the National Science Foundation was able to start its first new directorate in decades and it was a director called Technology Innovation and Partnerships. And essentially what that directorate did was say we're going to keep being really strong at basic research that NSF funds at universities across the country. But we also know we need to add investment to make sure that those research investments turn into economic opportunity. And if you look across the country there are many parts of our country that have not had good economic opportunities don't have manufacturing or other businesses. And the kids who grow up there are have to leave to find those opportunities. And yet those places often do have a great university that's doing pretty interesting research. And NSF boosted many many they funded a whole host of these regional innovation engines that were about building alliances between a local university and labor unions and and whatever industry was there. And the community that knows what they they can be really good at but nurturing that so that we have not just great research around the country but we actually start turning that into opportunities. For kids in every region. That's one example of what it means to to to invest in R&D for economic development in 20 you know in the 2020s. Right. This is the modern era. That's something that we've got to work on. So those are the traditional missions and the great aspirations. But you know people have known about those for 80 years. There are few things we've got on our plates now as a country that didn't exist in the last century. Climate change existed if we weren't recognizing it and dealing with it. But that's something that requires deep innovation to have an economy and flourishing and resilience even through a time of tremendous climate change that we're now starting to experience. We're living in a time where the fruits of prior research have created very powerful technologies like A.I. And one of our great challenges as a society today is to make sure that those that we use those technologies in ways that reinforce our values rather than eroding privacy and destroying the trust that we have in society and bolstering discrimination. So even dealing with the results of R&D becomes one of the innovation challenges in a really interesting way. So that's how I think about what we need to accomplish. And then we have these amazing mechanisms. We have the National Institute of Health Institutes of Health we have NSF. We've got DARPA to do its piece. We have NEST to do its piece. We have Noah which is so important for weather and the environment and our climate though those are all places that that need to be able to be strengthened and bolstered and then aimed at these great challenges for the future. That's the work that we were making progress on. And now that's that's being undercut often destroyed. And when the moment comes not if but when the moment comes that we can build this country again those those are the great aspirations that will build the R&D system to fulfill.

Dan Koh: It's interesting because it feels like R&D falls into this general theme from the Trump administration of short term political optics of a win in in sacrifice of long term innovation and gain for the country. And what I mean by that is like you know you cut USAID right. They say okay we've saved 20 billion I'm making up the number but roughly some billions presumably. And what people don't realize is that you cut that aid all across the world the standing of the United States and our our our view and interpretation from other countries as a leading voice as leaders in the world as people who care about not just ourselves but our global standing is all better for the United States if we do that. But that takes time to to reset. It takes time for China Russia to fill in the void even though they've already done that rather quickly. Similarly if they're if you're going to cut funding for climate if you're going to cut funding for innovation it may not be tomorrow that all of a sudden a cancer drug or some other thing doesn't develop as fast. But it does slow that development maybe beyond the time in which the president is the president. But the longer term impacts of this it seems like you're saying is is vast and incredibly concerning.

Arati Prabhakar: Yeah absolutely. And again I think some of the immediate impacts people are seeing some of the immediate impacts their data sets that we rely on for just our everyday lives everything from the weather to understanding how infectious by the way the reporters get are getting their data. It originates from know us. Absolutely. So like your local or local weather person who everyone probably recognizes more than the actual anchor on TV ultimately that data is is coming from Yeah Noah. Yeah absolutely. And so where does the weather forecast come from. These are computational predictions with human interpretation but they are all based on raw data that comes from sensors that are on ships and satellites and buoys all around the world that Noah implements operates collects that data curates that that noisy messy data runs the models and then provides that widely to the community. And then you know people tweak and adjust and do their version of the forecast versus someone else's version. But none of that exists without this this foundation. This is what Noah does. And that's what's being undercut the funding for those physical systems to maintain them to operate them and then the smart people to analyze and turn that into something useful.

Dan Koh: So there's a lot of people out there who are feeling incredibly distressed disaffected. There's also a lot of people who may have voted for President Biden who voted for President Trump this time. And I think there's a lot of conversations happening in communities about how they engage their neighbors who fit into any of those buckets. So my final question to you is you know what is your advice to listeners to reengage disaffected voters whether it's Or whether they're just staying home to just get them wanting to be part of the civic process again.

Arati Prabhakar: I think we're in a time now where the actions of this administration are starting to manifest in lots of people's lives whether it's you know hours that you can't reach someone you need for Social Security on the phone whether it's seeing ICE agents in your shops and seeing people knowing that people are being put away without due process really just appalling transgressions that are happening. So I think it's starting to manifest. And I and unfortunately for dark reasons people are paying attention. What to me what's so important about this is I want to make sure that we don't lose the thing that's so important about this country this this historic experiment that that we the people can build a future that's better than the future that that a king or a dictator can build. That's the essential idea of America. We've proven that that hypothesis is true for two and a half centuries but we're being tested again. We don't get to live unprecedented times. Right. We just don't get to. But what our times call us to stand up for this experiment and to make sure that it continues to succeed and thrive. And I I just I feel so blessed to have lived in these decades in America. When I got to see over and over and over again the things that this country can do because of its creativity the freedoms that we have but because the government goes to work every day to create these possibilities so that industry can bring these new products to life that change our lives. I've seen it over and over again and I want that future for my kids and I don't want to let go of that. And that's what that's what keeps me fighting. And I hope as people start engaging more deeply and thinking about the choices that they need to make because it's going to depend on 340 million Americans across this country what the you know how this story turns out. But I hope that people will keep in mind the big things that this country has done and is still doing after our 85 car.

Dan Koh: Thank you for coming on The People’s Cabinet

Arati Prabhakar: Great to be with you Dan. 

Dan Koh: We hope you enjoyed this episode of The People's Cabinet. If you did please like, subscribe, ring that bell, and put it in the comments below Who else you'd like to see sworn in. And stay tuned for new episodes every Tuesday. Let's go.