Michael Kratsios on the New Golden Age of American Science | EP #276

I was a kid in the candy store reading the Golden Age report. What you're describing there is a complete fundamental AI-native, AI-agent-up reimagining of the entire scientific process.

And I think it's something that is possible. My sense is, um, this is the golden age of America. AI is a technology that is going to impact every agency, whether you're flying drones, whether you're doing AI-powered medical diagnostics, whether you're like in at the SEC and working on financial services. AI is going to impact every single one of you.

Do you have sort of a longer-term compelling vision of what you think America could be like.

We as a government need to be opinionated about what the most important things are for the future of our nation. I mean, we're going, putting man back on the moon in '28. We're gonna be able to build the first elements of a lunar base by '30. We're gonna put a nuclear reactor in space by '28. I mean, that's crazy.

Yeah. Okay, my favorite idea. It falls into the crazy idea. I can't believe Michael actually wrote this down. All right.

Now that's a moonshot, ladies and gentlemen.

Welcome to Moonshots, everybody. Today I had the pleasure of interviewing a friend, Michael Kratsios. He's the 13th director of the White House Office of Science and Technology Policy and the science advisor to President Trump. Michael is the principal architect behind 3 landmark initiatives that are shaping America's acceleration during the singularity. The first is America's AI Action Plan, The Administration's Roadmap for Winning the Global AI Race. The second is Genesis Mission, a Manhattan Project-style effort to accelerate breakthrough discoveries. And then most recently, Science: A New Golden Age, his blueprint for rewarding bold, unconventional ideas and dramatically increasing the rate of scientific discovery. So this is the spot.

This is it.

Yeah.

You go up and give a briefing.

Yeah. Ladies and gentlemen, I've called you here today to let you know that we have now officially approved a trillion-dollar science budget. Congratulations! We're going to be solving every problem on the planet within the next 4 years of this administration.

Great job.

Yeah.

Thank you so much.

Well done. One day we'll make that announcement. This interview takes place at the White House, and I'm asking these questions on behalf of myself and my moonshot mates. All right, let's jump in. Enjoy. So, Michael, we are arguably living during the most extraordinary time ever in human history, where science and technology is hyper-exponential. And you're in the thick of it. You're in the middle of it, right? Ray Kurzweil predicts we're going to see as much progress in the next decade as we've seen in the last century. And that's like going from the Ford Model T to Starship in the next 10 years. On top of that, we're on the edge of AGI, maybe the next 3 years, ASI. How does the government process ever keep up with that?

You know, we're trying our best. It's something that I think generally governments have struggled with for a long, long time. And I think what we have to do is make sure that areas where we are seeing this tremendous growth, that we allow the regulatory system around them to exist in such a fashion that it doesn't get in the way of this progress. You know, I typically think of technologies kind of in two buckets. There are technologies that are either born free or they're born in captivity. So born free technologies are things like what the internet was in the 1990s. And the best thing that the government can do in those situations is to step back.

Yeah.

Get outta the way. Don't jump into it.

I think it happened so fast they didn't have a chance to get in the way.

Yeah. Well, there was interesting, there was a bill passed in 1996 was essentially like Bill Clinton was behind it. It was kind of bipartisan. And I think it kind of allowed some of this stuff to kind of take hold. But, and I think extends for a lot of, a lot of technologies. You know, be, be careful before you start, before you start regulating. Um, an example of that is, is in, uh, is in AI. And I think I— this always comes up when I think about AI regs. You know, the EU AI Act was passed and finalized by the EU Commission before ChatGPT was even invented.

Yeah.

So there's no way that what they have today actually applies to the LLM, LLM, uh, uh, of today. But I think the, the second types of technologies are the ones that we pay particular attention to, because those are the ones that, you know, we actually have to to take action on, and those are technologies that are born in captivity. Think of commercial drone operations. Think of AI-powered medical diagnostics. These are technologies that cannot be commercialized. They will not be, their benefits will not be realized by the American people unless the government affirmatively does something. And those are places where you have to be really careful because if you wait too long or you aren't thinking about them, it actually holds up progress. So first we kind of think of these two buckets of work and make sure that for technologies born in captivity, we're thoughtfully approaching, you know, how you can change the regulatory structure to allow that to ultimately be safely deployed to Americans.

I'm thinking about the timeframe you've got. If we're going to see this kind of extraordinary progress to AGI and ASI in 3 years, do people here in the White House understand the speed of that change? And, you know, and the agencies? I mean, it's dramatically— it's not a little bit faster, it's dramatically faster.

You know, we're trying our best to bring people along to the new pace and velocity of change when it comes to technology. And I think the best best sort of manifestation of that is our Genesis Mission. This is where the president stood up with Secretary of Energy and with me and said like, look, the most important thing for the nation is to make sure that we're applying this unbelievable technology called artificial intelligence to scientific discovery. And it's not just at one agency, it's across all of government. And I think those are the types of actions kind of from the White House level that are the only way you can kind of like really push this down into agencies. But I will say, I've, you know, in most cases, Government is not the, is not sort of the leading force in the cutting edge of where technology is. And, and that's going to be obvious, I think.

Yeah. We've seen recently ministers, uh, or AI ministers be appointed in various nations. We've seen, uh, out of, uh, the Emirates, 50% of the government operations being driven to AI. We've seen in Malaysia, I think it was the president, is like, has an AI representation, be able to speak in all the languages. Yeah. Do you, do you see that potentially happening here in the US in some fashion on the government side?

I don't think so. I, I, what I've always thought about AI policy, and this was sort of kind of our view of the world beginning in the first Trump administration where President Trump signed the first, you know, executive order on artificial intelligence in history in, in 2019. So this was years before ChatGPT and years before it was kind of on the front page of every newspaper. I think our general view and, and how we think about generally AI regulation is that, AI is a technology that is going to impact every agency, whether you're flying drones, whether you're doing AI-powered medical diagnostics, whether you're like in at the SEC and working on financial services. AI is going to impact every single one of you. The idea that you could sort of like centralize that effort in one person and be able to get the right and best policy answer across all those domains, I think is a tall, tall order.

No, I completely agree with you. I guess the question is, will we see AI enter the government in terms of, you know, advising on policymaking or advising in cabinet positions where there's a, you know, sort of an AI instantiation of that, that member of government to be able to, you know, counsel at the speed that we're seeing.

You know, I think I'm not sure where the future holds, but at least in the short term, what I would hope is that all our agencies can actually even start using, using AI. I mean, I will say—

We have a lot of catch-up.

We're sitting in, in, in the White House complex today. And I will say currently, We, you know, large language models are not allowed for use on our, on our, uh, on our system here because of Presidential Records Act. But hopefully we'll change that soon. But, uh, but that, that's an example of, I think, I think the pace at which sometimes, sometimes government tech operates.

This episode is sponsored by Google for Startups. Think about this for a second. You now have access to the same generative AI models that cost hundreds of millions of dollars to train. Google Startup Technical Guide for Generative Media gives you a complete blueprint for deploying Google DeepMind's models in production. Images, video, audio, all of it. Real architecture, real results. Find the link in the show notes below. All right, I, I'm excited to dive deep into the Golden Age Report and Genesis Mission, but before that, uh, I do wanna talk a little bit about AI.

Yeah.

You and I both know that AI is the engine fundamentally. It's gonna uplift every American, every aspect of Americans' lives. And I think we both feel that very deeply, right? The challenge, of course, right now is that three-quarters of Americans fear AI. Those are the numbers right now. 71% of Americans oppose data centers near their homes, which is a larger percentage than, than object to nuclear power plants in their neighborhood, which is insane. So I guess, you know, one of my missions is helping people see the optimistic future, you know, reduce fear. So I guess the question is, how does the administration get out in front of this? You know, why is there so much resistance? How do we demystify AI? What's the conversation going on around those concerns?

Yeah, um, AI has a massive PR problem.

Yeah.

And I think the AI companies realize that. Um, I think everyone that's sort of adjacent to, orthogonal to the industry realizes that. And it's And it's a problem. I mean, you and I know the, and do deeply believe in our core of what great things, you know, AI can do, but, and bring to everyday American lives. But, you know, I, this, you know, I think back first off of like, how did we get here? And I think back actually to the first AI safety summit that was held by the UK government in Bletchley Park a few years ago under Richard.

Way back 2 years ago.

Way back, yeah. This is, I don't know, was it 2, 3 years ago? I don't know, it's all sort of blurry. I think it was maybe the year after ChatGPT came out. And I think what was so fascinating about that when I attended it, almost the entire concept of it was wrapped around fear associated with AI. What could go wrong? Yeah. And we collectively as the smartest people in the world and the like most important government leaders must come together to make sure that these harms don't impact the world around us. And I think that's an example that the narrative for so long coming from government prior to President Trump has been so fixated on the negative impacts of this technology. Like of course people are skeptical. Yeah, yeah. If the only thing they're hearing from politicians and from, and from people in industry is that, oh, there's going to be a bunch of job losses and everything is dangerous and maybe there's going to be bio risk. Like, I think those are the things that, that people, people get worked up, and it shouldn't surprise us that the PR is bad. So I think for us, you ask kind of a positive take on it, I think we try to look for areas where Americans actually connect with AI in a positive sense. And I think by far what we have seen, it's in the healthcare domain.

Sure.

And that's where play— if we can actually show the impact in changing the way that individual individuals themselves get healthcare, their families get healthcare. I mean, those are the places where I think you can, you can start, you can start to push back a little bit on that narrative.

Is there any effort to try and change the narrative from a, uh, I, I guess from a centralized sense, right? It, it kills me that the narrative in China is the flip of that, where 80% are pro-AI, uh, and I just want to get out and like, like shout from the rooftops, like, you have to understand what's going on. Where does that responsibility lie? Probably in the AI labs, sure. But does the administration have a place in that as well?

I think the government can serve as a convener to bring to the surface all the great stories that are going on around this industry. And I think sometimes we think about AI in a little too narrow of a sense. AI is having this dramatic impact in manufacturing, in hiring across the country, in this—

100%.

—huge buildouts that are happening for everything that's sort of supporting the AI industry. Those are great stories. I mean, I just heard from, from Jensen was here and he was talking about kind of the supply chain that supports NVIDIA. He said that the order that he's put into Corning for the chips that he's going to be building in the future is the largest sort of order in the history of Corning. And they're going to be producing more than they ever have in history just because of his order.

Yeah, it's lifting up the entire GDP of the nation in an extraordinary fashion. And those are the stories should be told.

I mean, factories are being built, people are being hired, stuff is being produced. And I think those are the things that I think can connect with Americans.

Uh, one of the biggest fears is AI-related job loss. Yeah. And AI-related loss of incoming jobs. And, you know, I, I report on this on Moonshots every week or so in as part of the conversations we're having. And it's confusing because there's a lot of people who are putting forward data. Yes, there's job loss. Yes, there's these layoffs and others who are feeling like, no, we're going to see more job creation like we've seen with every technology so far. Do you have a sense of this? Do you have a sense of what you believe is—

I mean, personally, I believe kind of in, in the long term, I'm very optimistic about the impact that, that AI is going to have, have on jobs. I think we've been, we've been thinking about this problem, or I have since the first Trump administration, where kind of the most of the narrative around AI was around automation. And there was fears about all this automation-related job loss. And if you even fast forward through that period, That was never, never realized. And only employment has increased. But, but I do think it's something we have to think about. And I think we've put in some programs that start to hit at this problem. One of the things that I think that even you mentioned, your problem is the data around what is actually happening is not very good. And I think we're trying to launch an initiative at our Department of Labor that was, that was actually called for in the AI Action Plan. We want to start collecting better data on the impact that AI is having on the labor force. So you can collect data from players that typically don't submit their data to the Department And through there, you can, you can actually start making the right decisions on where to do better reskilling, retraining, and where to sort of do more targeted labor assistance programs.

I remember one of the conversations we had in Miami over lunch at the FII Summit was the idea of sort of a parachute program where you incentivize companies if they are doing an AI-related layoff to give the employees who are letting go AI upskilling so that they get trained into that environment. Do you think something like that might materialize?

You know, I think, I think every, every company will think about it differently. I do think opportunities like that are interesting. You know, I, I'm still trying to wrap my head around what it, what an AI-related job loss even really means. Yeah. I think like a lot of folks these days, like when they're doing a layoff that they would've done anyway, just like to assign it or blame it to AI because it, it plays better in, in the press.

And their stock price goes up if they are, you know, producing more revenue with fewer people. Precisely. Yeah. Yeah. Yeah. You know, uh, so on the notion of fear, because it's one of the things I'm always trying to quell and address, right? Because fear is an awful place to face the future from, especially at the speed of change. And people don't understand AI enough to understand its implications on their lives. Uh, we just launched this year something called the Future Vision X Prize. So it's the world's largest film competition, uh, for creators to create a film that shows a hopeful, positive vision of the future where technology and AI is working together? Because part of the challenge for me is that, you know, films like Terminator, Ex Machina, Black Mirror, majority of all the sci-fi films out there are dystopian. And if that's what we're teaching the average American, like, this is what happens when you have robots and AI. So I guess my question along those lines are, Uh, do you have sort of a longer-term compelling vision of what you think America could be like, uh, on the back of this AI revolution, beyond just better, you know, uh, better healthcare from AI, which is, which is low-hanging fruit? Agreed, of course. Yeah. Is that those— does that future visioning happen here?

I mean, we, we think about it more in the terms of national missions, which I think we'll kind of talk a little bit about in Golden Age. And I think what my My general take is, you know, we as a government need to be opinionated about what the most important things are for the future of our nation. And there was an era beginning with the Manhattan Project, going through Apollo, that we had big ideas, we had bold things, and national pride around them. And now— exactly. And it, and it motivated young people to go into science, and, and we were looking at a North Star of something that a lot of people thought couldn't be done. But we did it. And I think there are places where we can do more of that. An example of that is all of the space-related efforts that you, I think, talked to Administrator Isaacson and Jared about, about very recently. I mean, we're going, putting man back on the moon in '28. We're gonna be able to build the first elements of a lunar base by '30. We're gonna put a nuclear reactor in space by '28. I mean, that's crazy. Yeah. If you told someone we're gonna put a nuclear reactor in space that can sort of, that has enough propulsion power to send folks to, to, to Mars in the next year and a half, that's nuts. Yeah. But we're gonna do it because we're Americans and we can accomplish it. Accomplish that. That's so special. And I think that, plus some of the other national missions, can kind of—

why don't you name it? Just, we'll get to them later, but name a few of those, like, big bold Manhattan Projects, Apollo programs.

Yeah, I think the other one, it doesn't have quite a quick— it's sort of the, the Genesis mission, which is our AI for science mission. We essentially want to double the productivity of the entire scientific enterprise in the United States over the next decade. And that's going to have an incredible amount of sort of like fallout results there. Um, the third one that the president, um, directed through an executive order is to create a scientifically a relevant quantum computer by the end of his term. Um, and this is finally saying like, we appreciate and we, um, I think it's amazing that this incredible basic research has been done in, in quantum for so long, but we're going to like put a stake in the ground and we're going to say we're going to build this machine and have it do something. And have it do something. Yes, yes.

Do you have a, do you have a sense of, uh, in the quantum world, because quantum supremacy and all of these, all these terms have been as loose as AGI and ASI, What do you hope the first functional, capable quantum computer is able to do? What industries are you impacting most with?

I think to me, sorry to keep going back to health, I think pharmaceuticals is where, where I'm most excited about. I do honestly believe that the types of calculations you can run on those for the particular molecules and things that you want to apply towards, toward drugs is going to be, it's going to be pretty transformational. And kind of the last one, which is outside of our term, which I think is important for us to keep thinking about is fusion energy. People have been saying it's 10 years away.

They've been saying it's 50 years away.

50 years away, forever. It's always so many years away. But we've— Department of Energy put out its latest iteration of kind of our national strategy a few months ago. And I think we're— I don't think any time in history have we had more private sector investment in this particular energy source than today.

Yeah, my last count, there were 37 venture-backed fusion companies. Which is, which is crazy. It's nuts. Yeah, it's nuts. So I want to take a moment because it's timely to talk about the conversation that's been sort of dominating X and the AI sphere in DC, which is the open source versus closed source activity. Right. So Jensen comes out with the Open, Open Secure AI Alliance. Where is the policy today? How important is it for American companies to develop top-tier, uh, open weight models?

It's very important. As administration, we believe that the U.S. must lead the world both in closed and open source models, and that the best thing for the country is that we have a vibrant ecosystem on both sides of that coin that's able to support any type of customer that wants to work on AI or use AI. Um, you know, right now I think our open source ecosystem is, is one where we could, we could be doing better. Um, it's, uh, I think we I think we have a couple of very well-known startups that are extraordinarily well-funded that are pursuing getting close to the frontier. Yeah, Mira Murati's release was amazing. Yes, terrific. I think we're waiting on Reflection for their model later this year from what I understand. But I think that the ecosystem and the tooling around open source continues to get much, much better. And the US is continuing to be the default harnesses and tools around open source. But to me, I think, and we expressed this on the first page of our AI action plan from last year, The US has to lead on open source. And right now, look, I'll be honest, the Chinese have very well-performing open source models. And if you're an American entrepreneur and you're cash-strapped and you're trying to bootstrap your company and get started, I can't blame you for using the cheapest model out there. And at this moment, it's Chinese. But I think over time, I think we'll be able to cultivate a pretty vibrant ecosystem here.

Yeah, that's what America does. Yeah. You know, the Financial Times reported that Beijing is out in the world exporting open source, you know, basically as an instrument of influence and going to the Global South and providing them access to, you know, infrastructure and capabilities. One of your stated goals in the AI Action Plan, which I love, is the world should build on America's AI and tech stack. Back. Yeah. And I think fundamentally, because I mean, a lot of people are going to use the frontier AI, but the majority of the world, the vast majority of the world, is going to use the, you know, on-prem open source models. Uh, how does, how does America think about that? Maybe this is not necessarily your realm, but I'd love, I love your thoughts on that.

No, we think about this a lot. To operationalize the— well, you mentioned the action plan. We launched something called the American AI Exports Program, and the vision behind mind it was, look, we have the best AI stack in the world. We have the very best chips in companies like NVIDIA and AMD and many other new entrants. We have the best models that we all know about. We have the best applications. And if you're a customer around the world, if you're a government or if you're a, someone in the private sector that's looking to build AI, there is nothing better in the world than the American stack. And we launched this program at Commerce, but it's now become a whole of government effort where we put out an RFP and we said, great American companies, come back to us with what an American stack looks like. And submit those proposals to us. We're now looking at those proposals and ultimately we're gonna create essentially like turnkey American AI stack options for the world. Then we'll back those with the government financing organizations that can make them economically viable for a lot of countries around the world. So organizations like the Export-Import Bank and the Development Finance Corporation can provide financing to make these more economical. So a complete package. A complete package. And I think, and I think that's something that a lot of governments and a lot of folks around the world want. For them to be able to, to have to, to choose 7 different vendors to set up their stack, that's probably not what they're looking for. The other appeal that I think the US, um, has right now is everyone wants our chips. We have the best chips by far. And this—

and, and yes, I want your chips. He wants our chips. Yeah.

And, and sort of between the export controls we have in place and other factors that are going on, particularly on, on EUV, you know, our, our lead over where the best Chinese chip is is, is continues to increase year over year. So I think from a, from a, from a kind of basis standpoint, kind of the foundation of the stack, we'll continue to have the best product. Now I will say, I mean, I think this product, this AI Sport program, I think kind of was, was born because of a lot of the frustration I had in the first administration with, with Huawei and our inability as the United States to counter some of the actions that the Chinese were taking on telecom. Because, you know, they had a good enough telecom stack that was run by Huawei, and they had it super subsidized by, by the PRC, and they went out and kind of proliferated it pretty quickly. I think we're at a moment where we want to avoid that. And, you know, to us, I think we have a pretty vibrant, you know, open source ecosystem ourselves. But I think our sort of open source solutions plus our code source, in addition to the chips we have and the great American applications that everyone wants views, I think will make our, our, uh, our stack more competitive.

Do you think the restrictions on NVIDIA's chips to China in retrospect might have been a mistake? I mean, one of the things that we've seen in the past, I saw, we saw this in the launch vehicle industry, in the, uh, satellite defense industry, that when we start restricting our exports, it simply cultivates the competition to build their own capabilities and they come back and compete with us. Yeah.

Um, I don't think so. I think that was probably one of the, the, the savviest, uh, decisions that we made as administration. And I, and, and I think it was probably one of the most impactful in being able to throttle their ability to, to make competitive, uh, models to ours. Um, my personal belief is that it's been a, um, priority of, of Xi Jinping to have competitive, to have a competitive semiconductors, semiconductor industry for a long time. It started, um, certainly in Trump 1, and that's why the, the EV lift photography, um, export controls were so critical. I would say that is probably one of the most impactful export controls in the history of the United States. If that hadn't been done in, in 2019, and because of that, their ability to create their own sort of competitor to, to our leading-edge chip has, um, has, has, has really succeeded. This episode is brought to you by Blitzy, autonomous software development with infinite code context. Blitzy uses thousands of specialized AI agents that think for hours to understand enterprise-scale codebases with millions of lines of code. Engineers start every development sprint with the Blitzy platform, bringing in their development requirements. The Blitzy platform provides a plan, then generates and precompiles code for each task. Blitzy delivers 80% or more of the development work autonomously while providing a guide for the final 20% of human development work required to complete the sprint. Enterprises are achieving a 5x engineering velocity increase when incorporating Blitzy as their pre-IDE development tool, pairing it with their coding copilot of choice to bring an AI-native SDLC into their org. Ready to 5x your engineering velocity? Visit blitzy.com to schedule a demo and start building with Blitzy today.

Let's talk about robots one second. You know, I love robots. I've got my robots on order. We've got a number of great US-based humanoid robot companies, but it's— you can count them on your two hands compared to China with 150-plus humanoid robot companies. And it feels like they've been really aggressively supporting them, funding them, using them in national events, creating centers for robotic advancement and so and so forth. Uh, when do we start doing that?

Um, I think we have to do more of it. I, I, I believe that sort of one sort of manifestation of AI that is gonna, is gonna be critical of future success of the country is, is robotics. Um, as you probably saw, we took some pretty, um, dramatic action around, uh, around humanoid robots just, just this week, where we essentially limited the importation of, of any non-US humanoid robot that hasn't already been shipped going forward. And I think that shows how important it is to us for the US to lead in this domain. Homegrown industry. Homegrown industry. And we have to start sort of building the supply chain to support this. I mean, we believe as administration that these robots are critically important for the future of the country, for advanced manufacturing, for so many other things. And we have to sort of build that supply chain muscle to be able to have the supply chain security in the future.

Do you see capital going to these companies to accelerate that capability?

I do, I do. And I think the best example of that is we took a similar action around UAS or drones. Yes. In, I think in December of last year. And if you look at the numbers of the investment that went into essentially sort of drone supply chain since that action has been pretty, pretty dramatic. And I think that, that is kind of what we hope to see here. And I think this industry needs a bit of a push because because in like, just like so many other places, I think that the Chinese are certainly subsidizing and dumping on, on, on robotics.

So I want to return for a second before we, we head to, uh, New Golden Age and Genesis Mission, uh, to the speed of change. So a mutual friend, Elon, friend of the pod, you know, uh, when I interviewed him a few months ago, he's like, we're going to see double-digit growth in the GDP in the next 18-24 months, triple-digit growth in 5 years. And then he was on The Economist I don't know if you saw the clip saying we're going to be by 2036 basically post-capitalist. We're going to have, uh, anything or anything you could possibly want will be delivered by AI and robotics. Now, honestly, I think Elon is the most brilliant engineer on the planet. I think most everybody agrees on that. And his predictions are always directionally correct. The timing may be off a little bit, but the speed of change that he's projecting sort of breaks every system. And I'm just wondering, you know, A, could you imagine that speed of change? And, you know, are— is the appropriate members of cabinet thinking about what the economy looks like in that situation?

You know, it's hard for me even to kind of wrap my head around kind of that velocity of change. But, you know, maybe I don't quite ascribe to that particular speed, but I do think things are changing. And I think our cabinet and the leadership in the White House recognizes that and is getting in front of it. You know, I think as you probably saw, Secretary Bessant was very involved in some of the post-Methos activities that happened in the USG. And I think that's a great example of saying like, look, like if we have a super capable cyber model that in the hands of the wrong actor could pose a risk to our systemically important financial institutions, We have to approach this seriously and quickly, and we addressed it. And I think that's the kind of action that I think you see a lot of the leaders in the administration taking when it comes to these rapid changes.

All right, let's turn to Genesis and Golden Age. In your paper, you outline a series of challenges that need to be solved in the age of AI. And I'd like to hit on each one, one at a time. Yeah. Because they're really important. And just again, as I was saying before we started, I was a kid in the candy store reading the Golden Age Report. It was like you went much further than I expected, sort of naming ideas, and we'll get to those. So the first point you make is that scientific productivity has been declining despite larger budgets. Uh, uh, uh, Eroom's Law, right? Moore's Law spelled backwards. That's sort of like the discovery per unit dollar has, has dropped. Why? What's going on here? We got better tools.

Yeah, I, I— to me, I think we have, um, been, uh, unable to— or, um, or, or just don't, don't— unable to, to change the way that we conduct science. And I think this goes back to kind of one of the, one of the main sort of reasons why, why we wrote this report. The president wrote me a letter after I was confirmed and essentially kind of like charged us with how do we revitalize the science, the science enterprise. And we went back and we kind of thought about it and kind of the data that you talked about, kind of this declining productivity was kind of one of the, one of the first things that we looked at. And we asked ourselves, you know, like, why, why, like, what, why is it like we have better technology than we ever have? Our budgets are more than we ever have. And I think it's most probably relevant in the, in the biomedical field where the NIH budget now has ballooned to almost like $45 billion. Yet the cost of drugs is more expensive than ever. Yeah. Kind of list goes on. And I think one, one conclusion we had was like, we just are not experimenting enough in the way that we conduct science. We're doing the same thing over and over again and just putting more money towards it and believing that the outcomes are gonna get better. And I think one of the sort of the main sort of core theses of the, of the whole Golden Age report is we have to be more experimental and more ambitious about testing out new ideas. And why I found it so particularly relevant and shocking in the world of science is you You would think that in science, people would— those people would be the most interested and the most excited to try different ways of doing science. Yet, funny enough, that community doesn't want anything to change. Yeah, they want the system to be exactly the way it was 30 years ago.

Yeah, the way I say it, if you're an expert in something and there's a revolutionary breakthrough, you're no longer the expert at it. So there's a disincentive for, for doing that. Um, could it be regulatory bloat? Could it be, you know, legal bloat? Uh, could it be just the paperwork that's developed over time? It is.

Yeah, I think it's all the above. So some things that we talk about in the report are, um, are around, uh, research burdens. So I think there was a, a very well-known study that the National Academies put out a few years ago that essentially said something like 45% of the time that a researcher spends is to do the administrative work associated with their grant. And that is one of the most depressing statistics I can think of. These scientists are the crown jewels of our country. We want them to be doing their work 100% of the time. Same thing in healthcare.

Yes. Physicians are spending all their time filling out paperwork.

Yeah. And I think that the regulatory stuff does, does kind of matter too. And I think one example that's very close to my heart and has been one of my pet projects for a long time is how do we bring back supersonic flight to America? And an example of that is, it's kind of a, kind of a regulatory issue where in the US there has essentially been a speed limit for flight flights over land. So essentially, there was— if you're flying over Mach 1, it's just not allowed anymore because of the sonic boom and the concerns, right? And the reality is, I think research has shown, and, and I think Boom Supersonic, a company, um, showed that they're able to fly over Mach 1 without creating a sonic boom. So our regulations, the way it stands, disincentivize them from ever trying to fly over Mach 1 because they'll never be able to. So, um, again, due to the president and executive order that he signed, you know, we are now changing that rule to create a noise limit rather than a speed limit. Yeah, I If you can keep it, if you can keep it quiet, fly, just fly.

I love that. I mean, the breakthrough, you know, the policy changes on both supersonic flight and on eVTOLs, uh, just to, I mean, aviation industry had been stuck for 50 years. And so unleashing in that way. One of the second challenges you note here is young talent waiting too long before they're given a chance to implement their bold ideas, that our funding, publishing, and credit system was built for a world of human-paced discovery, and I could not agree with you more.

Yeah, the challenge I think a lot of young people face is, is they're kind of stuck in this zone of the way science was done long, long time ago, before the internet even existed in some ways. And everyone is kind of stuck through this process of needing to do research and then trying to get the research published in a journal, and then you have to do that a number of times before you can get tenure and so on. And I think this structure tends to, I think, not be the the way that is conducive to, to discovery. It's the way to, like, how you succeed in a, in sort of a structured system, right? Yeah. And if that— and, and I think that— I think there's, there's huge opportunities to kind of reform that. To me, I think NIH is always an example of this, and, and I think this is very bipartisan. I think, um, a statistic I heard today which was shocking is that, that the median age of an, of an intramural NIH scientist— so this is a scientist at NIH who's doing research in NIH, not getting a grant out is 71. Oh my God, when I heard, I couldn't even believe it. That is crazy.

And, you know, the average age of a Nobel Laureate's prize-winning work is in their mid—

in their 20s. Yeah, precisely. And I think we've created these systems where we're somehow okay with that, and then we look in the mirror and then, then, and somehow we tell ourselves, well, everything's doing great if we just give it a little more money. And it's like, that's not how you solve these problems, and that leads to the bloat.

So you identified a bunch of great mechanisms for driving progress in the Golden Age. I'm going to hit on 4 of them that I think I'm excited about. The first one is long duration grants. Yeah. 5-year grants funded on day 1. Yeah. Yeah. Talk about that one, please.

So to me, I think grant duration is something that people don't talk about enough. Again, over time we've come to this sort of like zone of comfort where most government grants are roughly in the 18-month period. And that's just because the way it works with academic calendars and just how easy or quickly we can as a, as a government be able to review these grants and adjudicate them and give them out. But that's not the pace of scientific discovery. There are certain scientific endeavors that are quick and short, and it's a small experiment that only takes 3 or 4 months to look at and you can get done. And there are other ones, like you talk about, that actually require long— a longer period of time, that you want a brilliant scientist to be able to explore an idea which is going to take a little bit of time to sort out and focus on that versus another grant application. Exactly. Because most of these guys, after they've gotten their grant, they're already working on their second before they even finish the work on the first, because that's— they have to keep the pace going. So to us, we believe that you have to— and this is again, how do we support scientists? At the core of the entire, you know, New Golden Age report is everything we do is in service of the scientists. And this is a perfect example of it. It's like there are certain scientists that want shorter duration grants, and there's some ideas that need 5 years to play out. And we as a government need to make our money available to be able to be in the service of those scientists are going to be making the great proofs.

Uh, the next idea you put forward, which I love, is fast grants. A few pages reviewed under a month, you know, sized for the proof of concept. Yeah.

And I think we saw this and a lot of excitement around this. Um, and we've seen this historically around, uh, times of crisis. And I think this really came to the fore in, uh, during COVID and Tyler Cowen and others came together and, and pooled some capital and actually did this on the side themselves. We're making grant decision in a matter of like hours. For people who are submitting to, to work on problems related to COVID. But there's no reason it should be restricted only in times of crisis. There's lots of ideas that if we're able to, to answer them, then that may unlock other things that we'd want to do later on.

And I can imagine, I mean, a lot of these scientists are probably using ChatGPT or Gemini or what it might be to write their grants, and I'm expecting that grant reviews will also use AI, uh, as a mechanism. I mean, time should be massively compressed.

I think so. And I think a lot of thought is being taken at a lot of our funding agencies about how we can kind of accelerate that while sort of at all times keep kind of that meritocracy and that merit-based review gold standard.

Next one that I loved was experimentation with golden tickets. So what's a golden ticket?

Yeah, I think another issue of like, back to your question of like why we stagnated. I mean, one reason a lot of people say is like we're just not risk-taking anymore. Anymore. It was sort of like, um, the least common denominator ideas are the ones typically get funded. And people who have kind of out-there ideas, crazier ideas, things that may not work, but if they do work, it's pretty incredible— they're just not incentivized to even submit that application because they're not going to get— not going to get the award. So the question is like, how do you get around that? How do you get people— how do you incentivize people to be a little more out there? And one is this golden ticket idea. And the concept here is that on a merit review panel, you may have 3 or 4 people, each person on the panel gets 1, 2, 3 golden tickets. And with a golden ticket, you unilaterally can make the decision to fund a particular grant independent of what the rest of the committee thinks of. And because of that, you get two things happen. One, you incentivize people to have a little crazier ideas because you have a chance that if one person believes in you, you, you'll, you'll go forward. The other is, I believe, and I think the— this may even be the, the even better outcome, is you incentivize incentivize better people to be part of the review panels.

Because they, they have some—

because you get a golden ticket. So rather than having these people that are sort of like, you know, I don't know, mid-tier people who just do this, you know, because I, for, I don't know what reasons, you kind of can bring even better people in to, to kind of do the reviews.

I, I, I do love that. Um, you know, it's interesting. I was talking to a professor at Harvard, I won't name him, who has been extremely successful. And he was saying that he was getting dinged in his reviews because he had gotten too many grants and was too successful, and they needed to give other people a chance. Yeah. Um, the whole peer review process is, uh, is something that is unfortunately extraordinarily broken. You know, I define a break— the day before something is truly a breakthrough, it's a crazy idea, right? So where are we experimenting with crazy ideas?

It's a challenge. It is. And I think we have to do— and I think, you know, I think government, um, correctly, I I think needs to be a good steward of taxpayer dollars. And we have to be very thoughtful in the way that we develop our programs and we evaluate them. Yeah. And I think one piece of, of Golden Age is this concept of meta-science, the science of science, our ability to evaluate the ways that we're conducting science, the ways that we're doing funding and evaluating whether or not they're working and then course correcting when they're not. Like you and I in this podcast may really believe that doing fast-track grants is a really great idea. But we need to run the experiment. Like, let's do some fast-track grants and let's see what are the types that work, what are the ones that aren't, and then update. And the next, the next iteration of them is even better. And we as a government just don't do that. We never do meta-science. So one of the things we call in the, in the, in, in Golden Age is to launch meta-science units. And those have been announced at, at places like NIH and NSF already.

Nice. Okay. My favorite. Idea. Yeah, uh, it falls into the crazy idea. I can't believe Michael actually wrote this down, uh, is, uh, you describe, uh, the use of prediction models, crowdsourced intelligence agents, and decentralized autonomous organizations to fund scientific research directly. I'm going to read a paraphrase from your report because I think it's, it's very powerful. Imagine a scientific marketplace where funders post bounties for breakthrough AI agents identify promising leads, hire autonomous labs, and verify cryptographically signed results. Agents ex— exchange data, hypotheses, compute, and funding through microtransactions while smart contracts release payments as milestones are met. Prediction markets could guide grantmakers, bounty markets could direct resource towards unsolved problems, and reputation systems could identify reliable agents. Mm-hmm. The system would operate continuously at machine speed speed, replacing slow institutional coordination with market incentives, while human experts remain essential for judgment, ambiguous results, and deciding which scientific questions and breakthroughs matter most. What you're describing there is a complete fundamental AI-native, AI-agent-up reimagining of the entire scientific process. Yeah, I look, I—

this opportunity gave— this report gave us an opportunity to kind of dream big of where we can end up. And I think it's something that is possible, and I think it really is. And I think what, what we try to get at there in the report, or I do, is, is I think incentives aren't always that easily aligned in the current system we have today. And over time, we have technical solutions to be able to bring those incentives together and to do the information sharing at a pace and at a speed which will allow all things you just mentioned to, to come true.

Uh, I mean, having, uh, DAOs those involved having, you know, agents run the cycle, you know, arguably millions of times faster than a human would, if not even faster. Sounds like a mechanism that could make Elon's 2036 prediction happen. Actually come true.

Well, look, my hope is that this inspires some folks. I think one thing that I've observed, and I think you have too, is the level of philanthropy scientific capital is greater today than it's ever been in human history. If you even just look at the, at the, the OpenAI Foundation itself, it's almost, you know, a quarter trillion dollars in today's valuation. Um, so to me, I think we have an opportunity for, for really smart people to sort of push the envelope to try some of this stuff.

And you've got folks like Yuri Milner, Eric Schmidt, Marc Benioff, all, you know, funding science directly. They're doing, they're doing really incredible work.

And I I think when we think about, you know, and, and I think kind of one of the underlying or main premises of the whole Golden Age report is that the science ecosystem has changed. In 1951, Endless Frontier was written by, by Vannevar Bush that kicked all this off. You know, 70% of, um, R&D was done by the federal government and 30% was done by the private sector. And that has like flipped entirely today. The majority is done in the private sector. And, uh, and only about 30% is funded by, by the federal government.

Yeah. The companies can take a 10-year horizon if they need to.

Yeah. And I think what we see, and you have this flip with the, with private sector being more involved, you also have philanthropy playing a bigger role. So now if you look at all the sort of pieces on the chessboard, people, there's, you can bring all those people together to drive scientific discovery in a way which you could never imagine in a system designed in 1960. And when we think about about sort of, take for example, an announcement we made today about 4-year PhDs, this idea that we want to get more PhDs out faster and actually have their experience during their PhD program prepare them for a job in industry, not only in academia. So the idea that you'll have a private sector company that is paired with an academic institution to help a student pursue a 4-year PhD, I mean, that's amazing. And that's something that, that is reflective of today's reality, not something you'd imagine in sort of 60.

All right. The last mechanism I want to hit on is the use of incentives. Mm. So your report leans into prize challenges, advanced market commitments, pay for results, not proposals. Right. And you cite the $10 million Ansari X Prize in there. Thank you. I appreciate it. So I've spent my, you know, 30 years of my life focused on incentives, and it's sort of my home turf. Most agencies haven't experienced against this area, haven't learned how to use it. So XPRIZE right now, we've launched $600 million of prizes and we've driven about $30 billion of R&D as a result of those prizes. Uh, how could XPRIZE help your agencies?

I, to me, I think it goes back to partnership. You know, if there, if there are big national, um, scientific endeavors that we need to pursue or problems we need to solve, I think, I think there's opportunities to try to pull money together and make that, make the prize even bigger and, and draw people to, to solve these big challenges.

Do you imagine that the government would put out, uh, like billion-dollar incentive prizes, or are these going to be small? Do you imagine on your grand challenges— you know, I'm talking to Elon about this— I said, let's, let's launch 10 $1 billion prizes that focus every grad student, every company, like, these are the important things we need to achieve.

Um, you know, I think that the government spends about $200 billion a year in, in R&D. If you think about it, about $45 million, as I said, is kind of sits at NIH with biomedical. I think having, having $1 billion may be a lot for a single prize for the, for the government budget to swallow. But, but I think, you know, in the range of $100 million is certainly doable if the, if the, if the project is big enough for sure.

Speaking of large prizes, our largest XPRIZE right now, it's the $101 million XPRIZE Healthspan. It's a prize We have 830 teams competing and the goal is reverse your functional age by 20 years, give you cognitive abilities. You're 20 years younger, muscular and immune system capabilities. You're 20 years younger. I have, you know, I fundamentally believe and there are a number of incredible scientists like David Sinclair and George Church and others who agree that if you wanted to impact the US economy the most, boost in a positive fashion, you would tackle aging. Yeah. That if you could enable people to have— and it's really healthspan versus aging. You know, today in the United States, the average life expectancy is about 78, 79. The average health expectancy is 63. So the last 16 years of your life, you're in poor health. If you could move that needle up, give people an extra an extra decade of health, an extra decade of productivity, it would transform everything. Yeah. Is that— is— so I found longevity lacking from your report. Just, you know, you hit a lot of other great things, but—

Well, now that you mention it, we probably should have included it. I think it's something that's critically important. And I think it ties to just general health of Americans. And I, you know, we should find ways to work together on this and figure out what programs the NIH and other places can kind of accelerate that.

Yeah. A lot of folks believe aging is a disease that can be at least slowed, if not cured. We had Dario say that he could imagine doubling human lifespan in the next 5 to 10 years. And Demis talking about curing all diseases within the next decade. And those are impactful.

That is huge. I mean, I remember when I went to— the first time Demis sort of told me, you know, cure all diseases in 10 years, I wasn't sure if he was kidding or serious. Are you serious? He's dead serious. Yeah. And I think he thinks it's possible with where AI is going. Voice agents are just software, but software deserves a real development platform. I'm Nick Leonard, CEO and co-founder of VoiceRun. VoiceRun is your runtime and development platform for voice agents. On VoiceRun, agents are built and configured in code. No limiting no-code platforms. For developers, that means total control. And for enterprises, that means extensibility that meets your compliance complexity. We've built VoiceRun CLI-first, meaning we've kept your Claude code, codecs, and even your OpenCLAW in mind when we built it. Your assistant of choice can build and deploy voice agents, can test and simulate scenarios, and can analyze and evaluate at scale. In other words, we've closed the loop on voice agent development. We don't build demos destined to fail in production. VoiceRun is where the best voice agents happen. Visit us at voicerun.com.

So, you know, when I look at the 6 national technology missions that you named— AI, quantum, commercial fusion, lunar exploration, robotics, and next-generation semiconductors— all amazing. I find biotech missing from that. Yeah. Is there room for that to come on?

There, there's certainly room for it. And, and one car you may have been explicitly listed. I will say I think a lot of the biotech work has been nested under the Genesis Mission. So the first national initiative around AI for science. I think we announced about $5 billion worth of Genesis Mission grants just last week at the Genesis Mission Summit. And so many of them relate to biotech. So I think it's critically important. Yeah.

That was a question my dear friend Alex, AWG, wanted me to ask. Um, I, you know, my, my entire Moonshot mates, uh, Dave Erickson and Alex and Salim, are jealous I'm here alone with you.

So we got to have them out next time.

I, I, I wave hello to you guys. Um, interesting, uh, some states and cities are far more pro-tech than others, right? You've got states and cities, and we just created a map— I'll show you after this— that, uh, that Max put together looking at this across the the US, sort of— I call them singularity zones, right? Where they're pro-autonomous vehicles, pro-drones, pro-nuclear, pro-data centers, and so forth. Yeah. Uh, do you imagine that you could see some type of regions volunteering to be innovation hubs where the regulations are sufficiently relaxed so that they can accelerate research? I mean, we see that, we see China doing this. I see other places in the world, uh, Chile, did this as well. Yeah. Just to, just to accelerate the work and, and actually have the entrepreneurs who are interested in that region move there to build their, their system.

I, I would love that. I think we, we as an administration, I personally have been very, very focused on this as a way to drive innovation. The very first piece of paper that I worked on that, that I got the president to sign was in 2017 to launch the UAS IPP, which was the Innovation Pilot Program. And integration pilot program. And what was so special about that program was essentially, this was 2017, commercial drone delivery wasn't really a thing yet. It was kind of a dream that people wanted, but how do we accelerate that? How do we get people to actually start running the tests of like, what is it like to try to do deliveries? Where does it work? Where does it doesn't? When does it bother people? What like systems you can place them on the drones, all that. And the EO essentially called for the creation of this pilot program at Department of Transportation. And FAA then went out and picked 10 10 state, local, or tribal governments that paired with certain UAS operators to run these tests. What you essentially had, you had communities stand up and say, "Hi, I want to have this innovative technology. I'm going to carve out," and the FAA gave them sort of like regulatory clearance to be able to run these operations. And we've done that again in this administration with our EIPP, which is our eVTOL program. We're going to start seeing testing of eVTOL vehicles all over the country. So to me, I think this is really important. And if you remember back in the campaign of 2016, that, sorry, of 2026, sorry, 2024, the president also put out a number of videos around some efforts that he was gonna be doing in Trump 2. And one of 'em was actually around this. I think there was a whole narrative around creating these sort of innovation zones and cities where we can test these great new technologies.

Yeah, I think on our evaluation state by state, Texas came out number 1 in everything. Yeah. Just very growth.

It's very much an American,— is very core to us as a country, this idea of federalism in a way where, you know, states can, can choose to make decisions and people will vote with their feet and companies will vote with their feet and go to places where they're most comfortable. I think back to sort of this Elon conversation, I think we saw where a lot of the, a lot of the self-driving vehicles that were, that were, that were being tested in California up and left and went to Arizona when, when the rules changed. And, and I think that's a good thing. I think it's competition among the states. Regulatory arbitrage. Garage.

It's, it's a very powerful incentive for states.

And, and I think some ways I, you know, and I think this dovetails a little with some of the issues on, on data centers, for example, where I think, I think there, there could be a future world where you're actually going to be having communities competing for data centers. And we should be thinking more, at least the data center operator should be thinking a lot more about what incentives they can provide to the communities that they're entering in, such that communities are so interested in having them that they're competing amongst themselves. Themselves. And the president has taken a lot of leadership on this, um, putting forth what's called the Ratepayer Protection Pledge, where essentially directed all these, all these big tech companies and AI companies and data center builders to build, bring, or buy all their own power. And you're not gonna go into any community unless you cover all the costs associated with it. And if anything, you're gonna lower the costs of electricity in the community.

And that's been, and that's what the data shows. Yep. The, the locations with AI data centers are getting on average lower, uh, electrical cost. We have to address the water issue as well, which, uh, you know, we've talked about a lot of times on the pod. You know, you know, golf courses have 30 times more water use than data centers, and like almond farming, farming is like 50 or 60 times more. That was particularly, uh, interesting. Jarring. Yes. Um, let's talk about education. Uh, it's, uh, it's quarter report. You, you hit it on chapter 4, and as a dad of two 15-year-olds, I am pissed at the educational system. Today. It is, at least in the US, still tied to the Industrial Revolution. It's not tied for what's coming. And we're seeing, you know, huge resistance of high schools and even colleges to, to use this. And it should be just the flip. We should be— it should be AI first across the board. Um, you know, how do you see this being radically reimagined? Because I think it does need— you know, the, the public health, the public educational system, uh, is not serving the— our kids' future for the world that is racing at us, to use Elon's term, like a supersonic tsunami.

To me, on education, I think, um, it's something that we think a lot about for a lot a lot of tech, tech domains. We always have to put, in my opinion, the parents first. Sure. And I think we should be providing parents options with the ways that they believe they can best educate their children. You know, there's amazing opportunities for people on one extreme to be participating in things like Alpha School where, you know, you do, what is it, 3 hours or 4 hours of AI in the morning and then the rest of the day is on sort of social skills and, and other stuff. Um, and for some sets of families that works amazingly and people get really hands-on with, with this technology. I think there are other parents that prefer to have sort of like no technology in the classroom and they want their kids to learn the classics and they can end up, you know, learning about these other technologies kind of in, in other domains. Um, I think what, but, but reality we face today is that, that most Americans get neither of those things. Mm-hmm. They kind, kind of get a, a, a broken middle that doesn't provide provide any semblance of a positive future for those children. There's so many districts around the country where you see that, you know, majority of students graduating from high school can't even do basic math. So I think we have a lot to improve in this country. And I think on my side, when I think about kind of how do you integrate STEM into these organizations, that's where my sort of portfolio usually goes. To me, one of the most sort of depressing things about sort of American education today is just the declining number of American students that want to enter the STEM field. Yields. Yeah. And, and that's not good for our country's health, for our national security, for the future of our economic growth. Like, we need people to pursuing these STEM degrees. Um, and I think there's a lot that we can do from sort of a government standpoint to inspire people to go back into this domain. And hopefully all the work that Jared's doing through space and so many others—

I mean, the Apollo program, you know, is, is 100% responsible for everything I've done in my life. Yeah, it was Star Trek and Apollo. Uh, I mean, so, but asking the school systems to change, uh, over the course of the next few years, which is I think the time horizon we need to be talking about, is extraordinarily difficult with teacher unions and, and public education boards and so forth. So, you know, I wonder, is there an opportunity for sort of an, uh, an AI educational overlay that becomes available where parents parents and kids who want that can do that in the afternoons or weekends.

I, I think that's possible. I mean, from what I understand, that the Alpha School guys, for example, obviously they started with like running the schools themselves in certain locations. And, you know, it's not necessarily that cheap, so there's only certain people that can, that can participate in it. But I think the goal is to be able to essentially like open source the software or make it economical for anyone to, to kind of run those programs. So in my sense is a lot of that stuff is happening, and, and hopefully there'll be lots of options for parents who want to, want do that. Okay. All right.

Another fun topic, which I love, is, uh, the autonomous self-driven labs. So, um, the piece of your vision, uh, is— and let me, let me quote this— AI proposes the hypothesis, robots run the experiment, the model reads the results and designs the next one 24/7, no humans in the loop. I mean, you're basically running closed closed-loop science at machine speeds. Yeah, yeah. Um, so this is something you and I talk about that Lila Sciences is doing. They're building out a million square feet. Uh, you're looking at having this retrofit into the federal labs.

We believe that we should be building these and, and being able to use federal dollars to test experiments on these labs. Um, to me, I think, and maybe you even know better, kind of a long pole in the tent on some of this stuff is actually, is actually building the hardware that can, that can run sort of a, a wide range of experiments as you would want. So I think we'll probably start in more narrow domains and then expand as the, as the actual hardware comes itself. But, but I think that's a future that we want to try to achieve. I mean, I think that, that, that what has held a lot of scientists back is just how long it takes to run the experiment. And if you have a vision and you want to test it, wouldn't it be great if you could just go online, go online, put hypothesis in and hit go and then start thinking about the next one and not needing to wait?

And I love also the vision you had where, you know, a high school kid could potentially do that. A college kid someplace could get access on national lab equipment to run the experiment.

I remember in high school, you know, there were a few kids who'd always found some way to like meet a professor and somehow get to the university and after school run experiment or get in the lab. And I think now this just democratizes it and it gives it a huge opportunity for anyone that has, has ideas to be able to test them without, um, without the overhead and the burden of, of a lot of the institutions that hold the keys to the infrastructure.

So if this vision gets implemented, one of the questions I have is, it is fundamentally doing the— what a traditional research university does, right? So I've, I've spent in my 10 years in, in, uh, in college and graduate work a lot of time in the lab designing experiments, pipetting, plating dishes, and so forth. And the idea that it could be done. So, uh, do you imagine that this, this autonomous lab capability is effectively going to also retrofit into the university system?

I think universities are going to be probably one of the first folks to build these labs. I think they want to, they want to provide the tools to the greatest Ameri— they want to provide the best tools to the great scientists that work at their institutions. And I think they would would be doing a massive disservice to their academic community if they don't have these as resources. To me, I think if you're an administrator at a university, you want to provide the greatest resources possible to your scientists. And I think these are going to be what everyone's going to be demanding.

All right. I'm going to wrap this up with this question, and I want to go deep on it. In your report, Genesis talks about doubling activity, which I guess historically might have been thought of as ambitious. You know, in the moonshots world that I'm in, it's like, okay, 2x is like anti, like, you know, how do we 10x it? Yeah. Is it 2x because it's politically acceptable? Is it, is it 10x? Is it 2x because you don't think 10x can happen? Been. Yeah. Um, where do you see the limits here?

You know, when we were thinking about the 2x number, that's actually a number from, from last year when we were actually building, um, what ultimately was building the program that the president signed as a Genesis mission in, in December. And, um, and it's funny you say that because I've been sort of noodling on that actually myself for quite a few months now. And as we were, we were finalizing the report, I think we had even a conversation internally about whether or not we should rethink that, but we sort of already stated that publicly, so wanted to keep running with it. Um, my sense is, um, you know, government is, is one of the, the hardest institutions to, to change and to move.

Yeah, it's linear at best. It's linear at best.

And I think, you know, Elon learned it firsthand and, and did, and did God's work here to, to help, help the country. Um, to me, I think we should probably be aiming for 10. And to be honest, with the transformations that have happened in, in, in AI even over the last 6 months, we should definitely be, definitely be pushing for 10. Um, and, and I think even if you, if you think back to sort of early this year, you know, Opus wasn't, wasn't even out yet. You know, Anthropic revenue was at, what is it, 10?

Was it 10 billion? Now it's at post 70. Exactly. Fastest growing company on the planet.

Yeah, it did. It's unbelievable. And that's just in the last, in the last 7 months. Um, we also had sort of the, the Mythos moment and sort of the, the Fable release. Um, you know, OpenAI's been releasing their their new 6th model very soon. I mean, the, the pace of innovation on AI is just, is just insane. And, uh, and, and I think we, uh, we got it, we got it, we gotta aim big, we gotta do 10x.

I'm looking forward to that correction report. Uh, uh, Alex Wieser-Gross and I wrote a paper called Solve Everything. It's at solveeverything.org. We look at how do you structure the situation such that AI is able to, you know, it's already, to use Alex's word, cooking math, right? Math as a discipline is being wholesale solved over and over again. We're seeing every week reporting on new, new breakthroughs, new proofs being, being either dismantled or proven. But on the heels of that comes physics, chemistry, biology, material sciences. And at least we imagine an inflection point, right, where, you know, GPT-6 and whatever follows for, uh, for Anthropic and whatever Grok becomes, especially now that Elon's, you know, uploaded all of SpaceX's engineering data, which was a, you know, a baller move. Um, we see an acceleration in the rate at which science is fundamentally, uh,, you know, sort of— I don't want to say other words than accelerated— that, that it, it's stunning to people.

Um, that's what I'm excited about. That is why we did the Genesis mission. You know, we believe that every scientific domain, whether you're in chemistry or physics or math or biology, you're going to see this dramatic transformation and acceleration of discovery and, and hypothesis testing. And that's going to be because of AI. And for us, and, you know, to unlock that— and back to question about— back to your note about, about uploading the, the data to Groq. Our whole insight was that the US national labs have a tremendous amount of scientific data across a wide variety of domains that has not been made AI ready, that have not been uploaded to, to models, and have not been part of the scientific process. And we have this huge opportunity before us to actually bring all that great data from 70 years of scientific discovery in the US into an AI model to be able to accelerate network.

When that— by the way, when that data gets, uh, brought in, let's say wholesale health data from NIH and such, and breakthroughs occur, does, does the value of that breakthrough in order to the American people? Who gets to get the financial benefit from the breakthroughs that come out of federal data?

So I think those are, those are public goods. That's, that's data that, that is a public good, and anyone can use it. Think of it kind of as like, uh, as weather data. One of one of the big sort of actions that the federal government made many years ago was to make all of our weather data from NOAA freely available. So now all these apps are built on top of it. You know, your weather app or everything else essentially runs on NOAA data. And I think we believe that, you know, taxpayers have been funding unbelievable science at DOE for decades and let's make the most of it.

Okay, one more side question. Milei, uh, Argentina comes forward and says, we're going to— no taxes on AI companies. We're going to enable AI personhood for agents here. A pretty aggressive stance. Yeah. Yeah. I'm just curious, what was the, what was the internal conversation when that came out? Uh, like, like, wow, good on him. We should, we should join in.

Uh, I think that was a very interesting take on where, where we're going. I think, uh, I, I think, I don't think we're quite at at that place right now.

We're not going to give agent person, uh, AI personhood in the U.S. yet? I don't think so.

I think for us what we want to focus on is making sure that the benefits of, of, of AI are actually realized with American people. And we have a lot of work to do to make sure that that actually happens. Um, and, and, and the president has been very focused since, since day one to make sure that we continue to lead the world in this technology. Um, and there's a lot the government can do to do that, but most importantly, we just have to kind of, you know, let our horses run. You know, we have the best companies in the world and we need to create a regulatory environment that allows them to just keep making the best, best breakthroughs in the world.

Amazing. Michael, thank you so much, buddy. Thank you.

This was so fun. Thanks for your work.

Really brilliant. Welcome to the health section of Moonshots, brought to you by Fountain Life. You know, my mission is to help you use the latest technologies, including AI, to not just do your work at home, teach your kids, but to help you live a long and healthy healthy life. I'm here today with an extraordinary physician, the Chief Medical Officer of Fountain Life, Dr. Dawn Musalem. Dawn, let's talk about cancer. Uh, you know, I know, uh, from the member database that we have at Fountain, our members who come in who think they're healthy, it turns out 3.3% of them have a cancer in their body they don't know about. That's right.

You know, the majority of cancers that we screen for, those aren't the ones that are necessarily taking lives when found at a late stage. We know that when cancer is found early, the chances for cure are much higher. We know it's much easier to treat a cancer when found early versus when found late. What we're finding in our members is over 3.3% were found to have these cancers that were otherwise— wouldn't have been found or detected.

Yeah, you know, it's interesting. People, you don't feel the cancer until stage 3 or stage 4. And if you don't know what's going on inside your body, it's like driving your car with your eyes eyes closed. And you can know. And so when members come through Found, how do they detect cancers?

So we're doing full-body MRI, and we also do early cancer detection screening. This is very, very important. And these are not typical tools used in the conventional care setting when it comes to prevention. This is a hard thing because currently these are not studies that insurance would yet be covering. But the goal is to collect these numbers, do the research, and work hard to democratize wellness.

Yeah, so at the end of the day, you can know what's going on inside your body. It's your obligation to know. So check out Fountain Life. You can go to fountainlife.com/peter to get access to the latest technology to help you detect cancer at the very beginning, at stage 1, when it is curable, before it gets to stage 3 or stage 4 and you're a world of hurt.