Derek Thompson on AI, abundance, and human progress
53:47 Share

scientific productivity is somewhere between flat and down. Why aren't we interested in fixing that? Think about the importance of productivity for every other sector. What if manufacturing never had productivity growth? What if agriculture never had productivity growth? We'd still be living in the 1700s. If we could speed up the productivity rate of scientific research, the inventions we could pull forward in time are unthinkable.

Hi, I'm Reid Hoffman. And I'm Aria Finger. We want to know how, together, we can use technology like AI to help us shape the best possible future. With support from Stripe, we ask technologists, ambitious builders, and deep thinkers to help us sketch out the brightest version of the future, and we learn what it'll take to get there. This is possible.

In a world that feels like it's changing faster than ever, we need smart thinkers to help us make sense of it all and build the best possible future for everyone. That's why we're thrilled to have Derek Thompson with us today.

Derek is a longtime staff writer at The Atlantic and the host of the podcast Plain English, where he dives deep into news and culture with clear viewpoints and takeaways. You might also know Derek from his bestselling book, Hitmakers, about what becomes popular, or from On Work, his exploration of the history of labor and the future of work.

Now, he has teamed up with another prolific writer, podcaster, and thinker, Ezra Klein, to co-author Abundance on the Future of Progress and How to Unlock It. They argue that the shortages America is experiencing, from housing to energy, are actually chosen scarcities, and that to address this, we need to shift toward a politics of abundance, focusing on building, problem-solving, and actually getting things done.

That is, how can we and our government embrace an attitude of investment and risk-taking that unlocks more clean energy, more affordable housing, more scientific breakthroughs, and more economic prosperity for all, which honestly sounds particularly essential right now. Here's our conversation with Derek Thompson.

Derek, thank you so much for being here today. I am an avid listener of your podcast, as well as, of course, have read Abundance. So we're going to try to try if we can to hit on some new topics. But again, I could just sort of listen to your solutions forever. So I'm going to start with a hypothetical for you, because this is a late night fight that I have with some of my co-workers. And I'm going to start with a hypothetical for you.

If you had to choose to be born in 1980, 2020, or 2040, what would you choose and why? I would choose 1980 because I was born in 1986 and I have a high degree of certainty that I like my life. Any other...

choice I'd be making is opening a door that has a certain distribution of uncertainty that I can't calculate at the moment. Being born in 2020, among my many concerns would be one, the competition of AGI with jobs at the bottom of the corporate ladder, especially for 22-year-olds that just got out of college.

Number two, the ability of artificial intelligence to compete with human friends. I think that we're getting very good at building Einsteins of EQ faster than we're building Einsteins of IQ with artificial intelligence. And so I think that a lot of young people are coming into a world where human friendships are going to be potentially endangered, although I'm very interested in making sure that isn't the case.

And then by the time you get to 2040, while I am absolutely an optimist, generally speaking on science and technology, we're talking about a time horizon where these kids are, they're turning 20 in 2060. I have no idea what the world's going to look like in 2060. Maybe it is so much better than 1980 that it's almost embarrassing the same way that 1980 was embarrassingly better than 1900. I mean, you got penicillin, you got cars, you got television.

But I like my life. I'm happy with my life. And so I am not going to roll the dice on some alternate universe in which I'm born 60 years later. All right, Reid, I'm going to pull an audible just because I have no idea what you're going to say. What would you say to the same question? I was going to do a follow up question with Derek. Kind of why don't you follow the general inductive of every 60 years and take the gamble on the inductive? And because I'm an entrepreneur and a technologist and an investor, I'm going to do a follow up question with Derek.

I'd go out as far as I could. It doesn't mean that there couldn't be a nuclear war. It doesn't mean that there couldn't be something. AGI would not be the thing I would think would be the damaging case. But I think there are, you know, an asteroid could hit the world and that would change things.

But generally speaking, because when you look through the, you know, the three or four millennia of recorded human history, it speeds up. But like 50, 100 years, it's always a lot better. You know, my thing would be gamble and go to the future. That doesn't mean I don't like my life. I like my life a lot. But which I think is a good answer. But I would definitely go forward. And I'm curious why that inductive leap, Derek, doesn't make you take that leap forward.

I agree with the inductive logic that over

Pretty much any 80-year period going back to the first Industrial Revolution, things more or less got better in most categories. I think it's important to look at how things didn't get better over generational periods. For example, if you look at urban working environments for British men in the 1820s and 1830s, they were significantly worse than they were 60 years earlier because the early urbanization of those places was just horrific for public health. Before, the real breakthroughs

in public health in the later part of the 19th century, it was really quite miserable to be a British laborer during the first part of the Industrial Revolution. So I'm a little bit concerned about the fact that while over 80-year periods, things tend to get better, over 40-year periods, there can be a little bit of a wobbly lineup.

I also, I take climate change and artificial intelligence seriously as technological and phenomenological events. I think that they both involve an enormous range of outcomes.

Could we solve climate change in the next 60 years? Absolutely. We could build fusion technology and carbon capture at scale. It's totally possible. It's also possible that we don't, and that's a world that almost certainly has more climate chaos. And in the realm of artificial intelligence, which I'm totally fascinated by and want to be hopeful for and can imagine all sorts of ways in which it turns out better for humanity, I also think that this is a really fast-moving technology, and I'm not confident about what it's going to look like over the next, say, 10 or 20 years. So I'm positive.

possibly hypocritically making an argument simultaneously for being an inductive optimist while being neurotic about the fact that I really like my life and don't want to gamble it on this sort of cosmic roll of the dice.

Makes total sense. Let's start with something a little bit more fun on the AI stuff. Obviously, we'll get into some depth on it. So I'm curious to hear your thoughts on co-authors, humans and AI, and specifically, just because it's entertaining and we're talking to you, what can AI do today that Ezra Klein can't and vice versa? I love that question. Ezra is very brilliant.

He cannot, however, write me a 10-page graduate-level report on any topic I ask him a question about in five minutes. In fact, he probably can't do it in five days. Ezra's very busy. ChiatubiT has...

infinite metabolism for my curiosity. And I use its deep research function a lot. Not to write, because I find the writing quite wooden. And in fact, I like the fact that ChatGPT's writing is wooden, because I don't want to be tempted to slurp over AI content into...

my pros. I want to keep that wall quite high and quite sturdy, but it's an amazing co-researcher. It's an indefatigable co-researcher. It literally does not get tired, can answer all of my questions, has answered a few questions in the last few days on Smoot-Hawley and the history of the American tariffs. So that's an advantage to AI. One thing that AI can't do is point out the flaws in my questions. So when I ask ChatDBT a question,

It's not very good at saying, "Derek, you've asked me the wrong question." And one thing that's important for book writing is knowing when it's time to finally write the chapter, knowing when you've done enough reporting. Have you read the books? Have you talked to the experts? Have you talked to the experts who disagree with the experts that you've talked to? What Shashabt isn't going to say is, "You need to do more reporting."

You need to talk to more people. You don't understand this enough. All it will do is try to fix the material that I have.

So that's a good segue into the future of work. What can AIs do right now and what can't they do and what do we need humans for? And I think, you know, as it relates to AI, I also try to be incredibly optimistic, but the fear that AI will take people's jobs is certainly at the top of a lot of people's lists about what they're worried about. And, you know, in the past, the historical realities of tech innovations has usually been that they create more than they destroy, although often there's this transition point.

So what would you say to someone worried that more resources and government efficiencies will be good for the lives of Silicon Valley engineers, but not so much for the average American? What do you think it'll take to get to the shorter work week and the better labor rights and sort of the better quality for all backgrounds?

To extend the logic of my last answer, it requires us asking entirely different questions than the ones that we're asking right now. Right now, I think the questions that are being most profitably asked and answered in AI progress are, how do we refine this technology? How do we train it on the right data set? How do we have enough energy in order to run the training and to run the inference? Those are really important technological questions, but technology doesn't have values. Technology doesn't have virtues. You can have the exact same

steam engine technology for pumping coal mines in seven different countries with seven different public health systems and seven different income redistribution systems. And you'll have very different outcomes when it comes to the health of those workers and the remuneration of those workers. So the technological question doesn't answer the policy question. And you're asking a policy question, I think. So I would like to see the folks who are in charge of

AI policy actually asking these questions. All right, in a world where the typical software programmer can become 30% more efficient, what's going to happen next? It seems right now with the policies that we have, it's very likely that what's going to happen is those incredibly smart and high novelty computer programmers are going to adopt new technology and become more productive and just produce more

code working five days a week or even working six days a week. There's sometimes like a sort of Jevons paradox situation where as you realize you can become more productive, you're actually going to work even more because you realize how easy it is to use your per unit productivity to make more money. In order for there to be

a reduction of the workweek, for example, you almost certainly need either state and national policy or the ability of unions and labor to demand four-hour workweeks contingent on hitting certain productivity thresholds.

And those questions, I think, are entirely outside the realm of what OpenAI or Anthropic are looking at for their business. But these are questions that need to be taken up by our political leaders because they are political questions. So just to move to one of the topics of early this year, I've been watching your tweet stream on tariffs. It's kind of like, how does one hit oneself in the head very hard with a brick? So tariffs seem to be like the anti-abundance strategy.

Can you talk a little bit more about like what's going on with these tariffs? What would need to be fixed? Is there a possible silver lining here?

It's much worse than hitting ourselves in the head with a brick. We should be looking for friends to join us in this effort. And instead, we're shooting ourselves in the stomach and the bullet is passing through the spine and then hitting them in the stomach as well. Sorry to be gruesome. No, totally agree. But that's what's happening. And why is it happening? It's happening because Trump, with the help of Musk, has been very diligent about turning the executive branch and by extension, all of the federal government into an extension of his personality.

And Trump has never been a free trade guy. And by extension, he's never been what I would call an abundance person. He is profoundly scarcity minded. He's profoundly zero sum. He is not shy about this. He's been against trade for 40, 50 years.

In The Art of the Deal, he says explicitly that the way he does business is he A, makes grandiose announcements, B, uses those announcements to pull counterparties to the table, C, tries to use his leverage to squeeze whatever he can out of those counterparties, D, rinse and repeat. It's exactly what he's doing to the American university system. It's exactly what he's doing to the white shoe law firms. Is it really a surprise that that's exactly what he's doing?

But then there's also a bunch of rationales that have accreted around the centrality of Trump's personality. Some people, the All In podcast is like, he's doing this to reduce America's borrowing rate on the debt, so we need a depression in order to bring down the 10-year yield. You've got folks like Peter Navarro, his trade representative, saying, "We're doing this to replace the income tax with trade revenue," which is completely ridiculous. You've got other people like Paul Merlucky saying the opposite, the tariff is just a threat.

in order to get other people to become free trade absolutists and then will become free trade absolutists in return, note that these things are totally incompatible. You cannot raise income

with a tax that is designed to disappear. The rationales are in conflict with each other. We are in the absolute twilight zone. It doesn't make any sense. I think the story of this era will be how to lose friends and alienate partners, which is terrible for the economy, terrible for the American people, and terrible for the world in general.

Well, I'm going to bring us to one great thing that Trump did, the COVID vaccine. I mean, Operation Warp Speed and getting that vaccine to us, the right is the party of not loving vaccines right now. But truthfully, that was a great thing. And that showed many people that government, public-private partnership, good things can happen when you move quickly in the arena of science.

And I think a lot of people see that and they say, OK, we can get a COVID vaccine, but why can't we get a high speed rail? Why can't we end homelessness? Like, why can't we do all these other things? And obviously that's, you know, the center of the book. So I'd like to switch to a little bit of optimism. What are the other examples besides the COVID vaccine that you think showcase abundance either in the U.S. or externally in other countries?

Well, starting with the US, I think Operation Warp Speed is an extension of a beautiful legacy of American history, which is Americans rising to their best when things are at their worst.

And you saw this after 1957 and Sputnik scared the administration into launching ARPA, Advanced Research Projects Agency, later renamed DARPA, which among other things got the self-driving car revolution kicked off, invented the internet, and did advances in semiconductor manufacturing, a lot of wins at DARPA.

And even before that, in World War II, we tell the story in the book of penicillin, which I love because I didn't realize this until I went deep into the history. People think of penicillin, if they're science nerds, as being this canonical example of scientists making important discoveries by accident.

It's true that in 1928, Alexander Fleming essentially just comes back from vacation as he's studying bacteria, realizes that one of the bacteria colonies in a petri dish has been totally zapped by some foreign substance that's maybe blown in through a window, and then analyzes that substance, realizes it's a mold from the genus Penicillium.

essentially, by accident, invents penicillin. That's the story most people know. It's a story that belongs in many science textbooks. But if you fast forward 13 years to 1941, penicillin has proved so difficult to scale and test that only five people in world history have ever been tested with penicillin, and two of them have died.

So imagine being a journalist in 1942, asked to write about this new technology called penicillin. Your first paragraph is like, it's been tested in phase one clinical trials and 40% of the population is dead. It doesn't sound very promising.

But it's brought to America when Vannevar Bush and several other people are launching OSRD, the Office of Scientific Research and Development, which is this all-of-government effort to update American science and technology. It burns the Manhattan Project. It helps us invent radar. And it very importantly helps us really invent technology.

I mean, not just find the molecule itself, but grow it at scale, test it, and then manufacture it at massive scale and ship it out all over the world.

By the end of the war, 1945, the rates of bacterial death among British and American soldiers declined by a factor of 18. That's very good. And penicillin becomes maybe the most important medical breakthrough of the 20th century. I tell this story because it's important, I think, to see how deliberate government policy can be.

pull a technology that would otherwise exist 20 years from now into the present. That's what OSRD did in World War II. That's what Operation Warp Speed did in 2020. Synthetic mRNA, if we went through the typical vaccine development process, would have probably taken about 10 years to go from discovering the molecular recipe to actually injecting it in people's arms in CVSs and Walgreens across the country. Instead, it didn't take 10 years. It took 10 months

because we had this incredibly deliberate emergency effort to subsidize, use something called pull funding, which means dangling a reward if various companies build a vaccine that's efficacious. We found a way to streamline the manufacturing process, streamline the distribution process. We reached out to companies like Corning to

developed special glass vials to carry the Moderna vaccines in. And then finally, we made this shot, which the government had essentially bought out from private industry. We made it available to people at the cost of $0.00. This is an extraordinary, extraordinary accomplishment.

And it's just fundamentally bizarre that it's been politically orphaned. Republicans should be able to run an Operation Warp Speed and say, let's do this for everything. Let's do an Operation Warp Speed for pancreatic cancer, for dementia. Let's just call it Trump Warp Speed or something. Just put Trump's name in it and then say, we're going to do this for everything.

every other malady. Democrats could have run on this and said, hey, credit where it's due. Donald Trump and his Secretary of Health and Human Services really did a mitzvah for the world with the development of mRNA vaccines.

We should learn from this. Neither party has taken this lesson to heart. It's insane. It's like landing a man on the moon and then having both parties say, hey, we didn't do that. You know, one party is like the moon landing was faked, actually. And the other party is like, oh, the moon. I hate it. It's very, very strange what's happened to this policy. So let's blend that a little bit with your, you know, kind of optimism and concerns around AI. So one of the things that you think the government should be doing to deploy AI to accelerate

Like if you were to be able to craft policies, what would be some of the first gestures? I want to start by saying this is an area where I very much still consider myself a student rather than any kind of expert. I still have a very wide open door in this subject. But I'll say one thing on inputs and one thing on outputs. Maybe I'll organize it like that in my head. Sure.

So I think it's very important to do a lot of data center siting and construction in the U.S. and among friendly countries. And that means, frankly, having a different attitude toward energy permitting. It's really depressing to me that the president and the White House are trying to gut the Inflation Reduction Act because it's

In our book, one thing that we say about the energy demands of AI is that hopefully AI's energy demands can be the horse that pulls the cart of permitting reform for clean energy. That doesn't seem to be happening right now. And I consider that an enormous shame because I don't want AI to just accelerate car

carbon emissions purely. I would love to see this built out with clean tech. And I know that some of the companies that are behind the AI revolution feel the same way. So that's on inputs. I think energy is an enormously important place where government can play a big role. On outputs, I would love to see government challenges for certain AI breakthroughs in order to pull AI development in that direction. So let's say, for example, that I'm suffering from some rare disease. Well, maybe we need to develop

a new medicine to fix that disease. Or maybe the same way that GLP-1s turned out to be an excellent medication for obesity, and also, by the way, inflammation, there might be drugs already approved that are very good for treating my rare malady. And I would love to see AI developed in a way that made it easier to sort of search the entire

corpus of FDA-approved drug molecular outcomes so that doctors could essentially say, ah, it looks like this drug approved in 2005 has precisely the medical or molecular outcome that would be beneficial to someone like Derek who's dealing with this protein problem related to his kidney that's making him sick. I would love to accelerate drug repurposing through AI, which I think we're in the early innings of doing.

I would also love to see if there's ways that we can make AI better at coming up with novel breakthroughs. And I think that might require a different kind of training. If the government created lucrative prizes for those sort of outcomes, I think we might be able to pull artificial intelligence toward that direction.

But right now you have a little bit of a wild west going on in AI. And I'm sort of enough of a free market person that I think there's value in having a lot of different theses tried out that aren't fully molested by state demands. But I also think there's a world in which the state can be more purposeful about saying, hey, it is clearly in America's national interest that we pull the technology in this direction and we're willing to offer a multi-billion dollar prize to that end.

On this podcast, we like to focus on what's possible with AI because we know it's the key to the next era of growth. A truth well understood by Stripe, makers of Stripe Billing, the go-to monetization solution for AI companies. Stripe knows that when launching a new product, your revenue model can be just as important as the product itself.

In fact, every single one of the Forbes top 50 AI companies that has a product on the market today uses Stripe to monetize it. See what Stripe can do for your business at stripe.com.

OK, so we're going to play a very quick abundance inspired game because we have too many things that we want to hear from you on, Derek and read. I'm going to pull you into. And so I'm going to ask you guys if we have too much, too little or just right on either of the following topics. Oh, fun. Cool. All right. So number one, Derek, the number of books the average American adult is reading cover to cover each year. Too much, too little, just right.

That's easy. Clearly, clearly too little. I mean, American reading for pleasure has declined by like 50% this century. It's really declined most for seniors. There's just a lot of Fox Newsification of the typical senior's attention, it seems. But you also have a ton of reports showing that literacy scores for young people are falling as well. So I think students need to read more books rather than just read passages that they can answer SAT questions about. And I definitely want to see older people read more books too. Read any disagreements.

Total agreement, not just because I would also recommend strongly that people both read Abundance and Super Agency. All right. I know the answer to this question, too, but I would love a sentence on why, because some people might think it's antithetical. The number of new American universities founded in the last century? Way too little. The U.S. has not created nearly enough high-quality research institutions in the last 100 years. When you go back to the period of American history between, say, the 1860s and 1910s,

We invented the research university. Johns Hopkins gift created Johns Hopkins University. John Rockefeller, his gift created University of Chicago. Carnegie gave to the institution that became Carnegie Mellon. Stanford, Leland Stanford gave to institution that became Stanford. We created so many

exceptional research institutions in a span of just 30 years. And we really fell out of the habit. Other than expanding the UC system, which is really wonderful in the first half of the 20th century, we've really, really fallen off the pace. So I'm a huge fan of not just expanding the number of seats at places like Harvard, Yale, et cetera, but also creating entirely new research institutions. Yeah, the abundance answer is both. It's more seats at the current universities and

But also, in addition, what I would add to Derek's answer is diversity of different shots on kinds of innovation opportunities.

of university institutions. So not just things like Minerva and other things, but like new kinds of universities doing different kinds of things. Right. Rethink what a university could be. It doesn't have to be the same thing. All right, Reid, I'm going to go to you first. Too many, too little, or just right Walmart locations in the United States? I think too few. I think the productivity of these kinds of centers, when it creates that kind of thing, it enables a lot of other productivity. So I would say more is good.

My certainty here is incredibly low. So I'm going to say just right, not because I think America would be better off if there were fewer supercenters, but rather because I think it would be interesting if we had another wave of innovation in physical retail.

You had Walmart really take off in what, the 1970s and grow a lot in the late 1980s, 1990s. Amazon founded, I believe, in the 1990s and really takes off in 2000s and 2010s, especially in physical retail with its acquisition of Whole Foods.

It's now been a few decades since the founding of America's biggest retail giant. And it'd be interesting to see someone else get a crack at that because Amazon turned out to be really good at a lot of things that Walmart wasn't initially very good at. So I'll say just the same because stagnation in Walmart locations with an overall growing retail sector would, I think, imply some innovation in other parts of the market. Some innovation coming in. No, absolutely. Absolutely.

So this is something maybe everyone has known forever, but I just learned about is government funding housing in Singapore is wild to me. Do you think we should have more or less the same government funding housing in the U.S.?

I'm a huge fan of social housing. I also think that if you take social housing policy seriously, you see that it's basically impossible to build housing with the social housing policy strictures that we have in places like California. And this is a part of the book. So, you know, it's been funny or interesting with some folks on the left, you know, criticizing the book for not exercising

explicitly advancing enough left-wing policies like social housing. But fundamentally, social housing is sensitive to many of the same physical world regulations that private housing construction is sensitive to. So Singapore founded their housing model in a much stricter and less liberal era, and so it made it easier

for them to build what they intended to build. And while I am a huge fan of democratic liberalism, I also think that we should make it easier to build as well before we plow a trillion dollars into a public housing system.

that just ends up building quote-unquote affordable housing for something like $500,000 per unit, which would just be a massive deadweight loss for U.S. spending. Absolutely. Reid? Well, I think abundance and efficiency in housing costs. I do think that one of the things you look at that have a massive increase in costs, which is

housing, education, healthcare, et cetera, should be targets of an abundance-like model in order to make it much more plentiful. It's part of what makes the quality of life better. Now, I do think the Singaporean housing model, I don't know which parts of it would be

doable in the U.S., but it's been amazing for Singapore. And the notion of the melting plot blend, which I think has greatly increased social cohesion in Singapore, and then also the question around

making that more available as a general societal right is, I think, good things. Now, which things we could do like it, I think it's an interesting question given the cultures are so different. Fair enough. All right. Last question. Derek, I'll go to you first. Do you think we need more or less border security? I'm going to be honest. I don't know the level of border security that we have right now. My general impression is that the border security situation, say, in 2020,

2022, 2023, had gotten to a point where not only were individual states finding it impossible to absorb the number of immigrants that were waiting in the asylum lines, but also was creating a terrible anti-advertisement for the Democratic Party, which is bad for immigrants

immigration in the long run, because if you're seen as the pro-immigration party that can't handle the border, then it really complicates the case for liberalizing immigration law in the long run. So I'm going to say the annoying thing, which is that having been on book tour for a while, I have just no sense of what the status quo of border security is. But I think that construction

conservative critics of the Biden administration were absolutely justified in critiquing the border security situation. I'm generally one of these people whose position on immigration is, A, we need a ton more highest-tailed immigration. B, I'm for liberalizing immigration generally across the board. C, it has to be an orderly process because in the event that it's seen as a disordered process, you're undercutting your strong case for both

high-scale, and general immigration. Reid? I mean, generally speaking, I think in many things, more is better. I think higher border security is good. I think it's, again, I would tend to go more innovation-oriented. I mean, part of the things I found nutty about Build a Wall is like, okay, let's do the two-century-old thing versus like drones or sensors or, you know, this kind of thing. But I do think that it's important to do. I think it's important for the coherence of society to have a good sense of

we have good social agreement on what immigration looks like. Now, we are a country built on immigration, and I think high-skilled immigration is our superpower. It's part of what makes our entrepreneurial community. It's part of what makes our scientific community. So the high-skilled immigration thing is, I think, you know, we should be trying to 10x because it benefits all Americans. But, you know, I think that, you know, more intelligent, call it technologically forward border security, I think is a good thing.

And that concludes our very first, and I guess possibly our last, game of too little, too much, or just right. And Reid, let's get you back in the co-host seat. Take it away. One of the things in your book that I had wanted to come to, because you may or may not have known that I co-founded Manas AI with Siddhartha Mukherjee. So we are tackling cancer, which you say is one of the most frustrating areas of declining scientific productivity. So we are trying to use

AI to increase the scientific productivity, but also the drug development part of it, which is actually particularly challenging. And that's usually because most regulatory agencies don't figure out how to create abundance. They figure out how to create minimization of risk, which also tends to be somewhat anti-abundance. And what you want is minimization of risk with a growth of abundance and how do you get the balance of it?

So go through a little bit about what do we need to do to incentivize the kind of scientific progress, the pipeline of young scientists, out-of-the-box ideas, because the current administration is basically trying to destroy American science. Yeah.

If Siddhartha were on this call and asked me, what do I think we should do to the NIH in order to increase the productivity of cancer research? I would reject the question entirely and say, can you just talk about that for an hour? Because there's literally nothing that I can tell Siddhartha Mukherjee about cancer research that that man doesn't already know. So let me answer the question. Then let me figure out if you absorbed any nuggets of wisdom from your co-founder, because he is an absolute genius.

I think the NIH is an absolutely brilliant organization, and I think what's happening to it right now is an absolute tragedy. That said, just because it is a tragedy to destroy the NIH does not mean that the NIH pre-destruction did not need to be reformed. The NIH is, in its current evolution, essentially an 80-year-old program that, like any 80-year-old bureaucracy, is showing signs of its age.

This is not my opinion. This is something close to a cliche among scientists that ask the NIH for grant money. To be very efficient with my critique of the NIH, which again, I think is the crown jewel of biomedical research around the world. It's just a crown jewel that has like in a bureaucracy accumulated problems. One, like most bureaucracies, it has become very good at incremental conservatism. And science, I believe, moves forward one step

shocking discovery at a time. And so if you have a peer review process that is very good at rewarding incremental conservatism, you are almost by definition declining to fund the most high risk, high reward research that I think is most likely to move science forward. Number two, throughout history, young scientists have often been the most paradigm shifting thinkers. The classic example here is quantum mechanics. Quantum mechanics was not

a breakthrough among a bunch of 70 year olds who thought about the photoelectric effect for a while and were like, "Oh, I finally figured it out. It's quanta." It was like a bunch of children, Schrodinger and Heisenberg and Planck. These were basically children at the time and they revolutionized science. And it was the older scientists at the time, like Einstein, who had trouble accepting their conclusions. That's a great proof of the idea that if you want to fund

paradigm shifting science, you should have an explicit strategy for funding young scientists. And instead, the share of scientists under 39 receiving NIH grants has declined by something, I don't remember the stat in the book, but it's something like 50% in the last 30 years. I mean, it's just stunning the degree to which science is getting older. There's nothing wrong with older scientists. It's just that there's a different kind of thinking that comes from being frankly young and stupid enough to think that you can change the world with a new idea.

Sometimes it's really nice to have that. And Silicon Valley, frankly, has demonstrated the profitability of allowing young people to believe that they have paradigm shifting ideas. And then finally, and this is one that I feel most passionate about sometimes, it's not just about bias against high-risk, high-reward research, and it's not just about bias toward older researchers. It's about a bias toward paperwork. The average scientist in America now spends up to 40% of his or her time filling out

grants and filling out paperwork about the grants that they filled out. As I say in the book, imagine if the entire American Academy caught a virus that gave them chronic fatigue disorder for half of the year. And so they simply could not work. No one in science could work between January and

and June of every year. Imagine if this was a Trump plan, right? Imagine if Jay Bhattacharya said, "Hey, there's a new rule for NIH funded scientists. You're not allowed to do anything until July." We'd be like, "That's effing dumb. Why would you do that?"

But this same principle of essentially zeroing out half of scientists' time for non-productivity is what bureaucratic paperwork is already doing. It's already taking half of their time. So this is a virus we've given ourselves. It's a policy that we've designed for ourselves, even though I think no one can explicitly defend it. So I think that NIH should take a page from science and

Try to master the science of science. Run more experiments. Give young scientists a bucket of money for 15 years and say, go away and don't call us back for the next decade and a half. See what you cook up. We're not going to ruin your life with a bunch of paperwork requirements.

Fund new research or fund new ways of rewarding grants. Maybe fund some by lottery, fund some by what's called a golden ticket, which means the three of us are on a review panel and read, Naria, you hate this idea that someone proposes and I think it's brilliant. So I use my golden ticket to fund that grant, even though you guys think it's stupid. This is going to be better for sort of fat tail research.

It is totally fanciful to think that anybody has a formula for scientific discovery other than running experiments. If there was a formula for discovering fusion technology, then everyone would know the formula and we would have fusion technology. By definition, there is no formula for discovering the unknown. And in the face of total uncertainty, you should be

wide open to experimentation. Instead, what we have at the NIH is a system that has built up over 80 years to concretize certain theories of scientific discovery that I think might not be the best ways to push science forward. Think about the importance of productivity for every other sector. What if manufacturing never had a productivity growth? What if agriculture never had productivity growth? We'd still be living in the 1700s.

Well, there's lots of research suggesting that scientific productivity is somewhere between flat and down. Why aren't we interested in fixing that? If we could speed up the productivity rate of scientific research, the inventions we could pull forward in time are unthinkable. So I think it is a huge, huge national policy blind spot that we're not talking about this more. But now I want to know what Mukherjee thinks.

Well, I would like to know what Mukherjee thinks too, because I agree with you about the fact that he is just an out-and-out, very broad-spectrum genius. You know, I myself have thought that this whole bureaucratization of science, especially once you've identified, you know,

You know, the talent is foolish for exactly the reason of like fatigue syndrome disease. Years ago, I used to do this deliberate grant making to try to break scientists free of the writing time. I'm actually very hopeful this will be one of the accelerations that AI will give us because AI, you know, like if it just made that 10x more effective and just doing the bureaucracy, you win a bunch of that back.

I helped a colleague of mine, David Sanford, start this thing called the Hypothesis Fund, which is go out and identify through a venture scouts model, the youngest scientists, just give them money to go do bold things. Because the key thing is going and doing bold things. And one of the challenges you have in essentially government and scalable capital allocation is the incentives tend to be that any error is used to kill you. And so you try to go to no error, which means no risk.

which means you can't be bold. And so you had to do various things to kind of make the bold things work. And we need to actually have that because the only way you get to the scientific invention, the only way you get to the abundance is through what you could call experimentation. I think the thing that I love about what you and Ezra have done with abundance is it's exactly right, which is how do we build as much goodness in the future as possible?

Technology being a huge driver of it, but also, of course, social systems and incentive for young talent and identifying that talent and having the 20-year-old quantum mechanics folks go, and now we're going to be bold. So, you know, plus one to all of that.

I love it. I do want to note that, Derek, if you haven't checked out what David Fagenbaum is doing at University of Pennsylvania, he is actually using AI to figure out what drugs that are on the market right now, for one thing, can help autism or help all of the spectrum of things. And he's super being helped by AI. So I love that it's already like seeping into what people are doing. But sort of switch gears from sort of science and the amazing things that we can do to something that just seems simple, like

I will tell anyone who will listen that there's only two things you need in life. You need to move your body and you need deep relationships. And I am so obsessed with the loneliness epidemic. And you've talked a lot about it. You wrote, you know, the antisocial century and how this wasn't just COVID. This was accelerating before COVID. A lot of people blame social media, the downsides, kids being home, adults being home. And you talk a lot about it.

How do we need to rethink society or design choices related to technology to sort of halt this epidemic? I think the first thing to recognize is that our lack of purposefulness is part of what got us here. We adopted these technologies without any real sense of how they were going to advance our values.

Technology doesn't have values. Technology works within systems that imbue it with values, right? There's nothing about the phone itself, nothing about Instagram itself that demands that its users spend less time with people.

But because the technology makes a certain kind of communication and a certain kind of entertainment more frictionless and convenient, it ends up displacing a lot of human-to-human connection in the absence of someone purposely pushing against it. That's fundamentally what I think we're seeing. I don't think these technologies, I don't consider them evil. I consider them, like so much technology, powerful.

And powerful technologies can have

powerfully surprising side effects. One mental model that I've talked about a bit, even though I don't think it's really deeply discussed in my essay, The Antisocial Century, which is just, by the way, an essay that billows out my observation that Americans are spending less time with other people than ever, and that this has a lot of implications for the way we live, for our economy, for our politics, and for the future. One way I think about it is something like this. So with the agricultural revolution, we used to be organisms whose bodies were designed around scarcity.

And then suddenly we invented an abundance of calories at our disposal. And over time, many people, especially in rich countries, gained weight because of all the calories that were available to us. And so you could say that, well, of course, caloric abundance is a pure good. It's a pure good that comes alongside this risk. And it took us a while to really make peace with or recognize the risk. And so maybe you could say we're dis-evolved

for a world. Humans are dis-evolved for a world of caloric abundance.

I think you could argue that humans are dis-evolved for a world of entertainment abundance and for communicative abundance, which is not to say that abundance, the material realms that Ezra and I talk about in the book are important, you know, housing and energy and scientific discoveries. I think these are public goods you want to make more of, but ours is not like an aimless abundance. So let's have more of everything. There's risks to the ample supply of some goods and services. And even more importantly, something I see is like,

There's this TikTok trend of young people, 20-somethings, celebrating when their plans are canceled. This is very interesting to me because it goes directly at this question of like, is this really a loneliness epidemic or is it something much stranger? Loneliness, many sociologists would say, is a biological response to the gap between your felt and desired social connection. So you feel like you're not around a lot of people and so you start feeling lonely and like, I want to get off the couch. I want to hang out with people.

What we're seeing in this world is actually stranger than mere loneliness. You're seeing young people who hang out with their friends less than any previous generation celebrate when their friends cancel plans with them because they feel so tired all the time and don't have maybe like the dopaminergic drive to leave their house and suffer the indignities of driving in a car and parking in order to hang out with their friends. And so like, I find that beyond lonely.

I find it really phenomenologically interesting that this is how we live today, that we have access to these technologies of incredibly ample entertainment, and we're using them in many ways to replace human-to-human interaction. That is a really, really interesting thing to me. I just listened to Jonathan Haidt on Ezra, and I decided that what I'm against is YouTube

and I'm pro movies. So I'm going to now have my kids watch movies all the time with morals and they have to be engaged and we're forcing the kids to watch it together. And so I'll just ask, are you like, okay, let's delay the smartphone until 16? What do you think? My feeling is...

I want my daughter, like myself, to have a healthy relationship with the technology rather than an aversive relationship to the technology. I think there's a danger both in sort of self-regulation and regulation of our children

in making something seem like the enemy, especially when it's going to be a part of their lives. It's the same thing with food. It's the same thing with exercise. It's the same thing with desserts. Look, I grew up in a household where chocolate was banned because my mom didn't like it. And as a result, whenever my sister and I were at other people's houses, we were just like gobbling up chocolate. And so that's a situation where you make an enemy of something and then kids, they all have a different kind of relationship with it. So it's important for me

to help myself and help my kids reconcile themselves with the inevitability of technology and build a healthy relationship with it rather than consider any of it like totally off limits.

Fair enough. Probably good. So as we've talked a little bit about our kind of living in the wrecking ball administration, you know, whether it's wrecking science, wrecking international friendships and partnerships or wrecking the economy. But one of the things is that sometimes the wrecking ball is a good thing for clearing out things.

What are like some number of policies, three to five, that you think hinder abundance, however well-intentioned, that if you were swinging a wrecking ball against would be the right ones to get rid of? And then what would be some policies that would be pro-abundance?

Policies to add Operation Warp Speed for, if not everything, at least five more things. I'd want to convene a panel of scientists to think about, okay, what technologies are close enough that the pull funding mechanisms of Operation Warp Speed could be incredibly useful in bringing this technology to light and especially creating prizes around it that would make the final application of that technology cheaper?

In the book, we talk about clean cement. We talk about the monococcal vaccines, I believe, for the low-income world. I think there's probably other medicines and other clean energy breakthroughs that'd be sensitive to poll funding. I'll describe that bucket. We'll call that bucket Operation Warp Speed for everything. I would want to create at the NIH a science...

a "prizes layer," a layer of the NIH that was created for the purpose of running more projects with the current amount of money they have or more, rather than the current strategy, which is, "We don't like the word mRNA, so never put that in your research. We don't like certain words, so don't research them." There's a kind of, it's not totalitarian, but there's a kind of autocratic insistence

that scientists don't do certain things that I think is totally opposite to what you want science to be. So rather than cut and dictate, I would want to open up a new layer of the NIH that is devoted to studying scientific projects that are run out of the NIH. Those are two policies that I would want to see added. In the housing realm, I think it'd be interesting to see what happens if you make certain pots of

of federal money contingent on states accelerating permitting or taking away zoning laws in high demand cities, I'd love to see the federal government use its might to encourage states to accelerate housing and energy production. So those are policies to add. And policies to subtract would be, I'd like to see an enormous increase in exemptions for NEPA. - Pi, can you explain?

NEPA stands for the National Environmental Policy Act. It's a United States environmental law that requires federal agencies to assess the environmental effects of their proposed actions and decisions.

This is one place where it seems like there's mutterings out of the Trump administration that they might do some of this. Nothing is all bad. And one can simultaneously pursue a tariff policy that makes absolutely no sense and a NEPA deregulatory policy that in certain cases does make sense. So I would like to see the number of exclusions for NEPA really escalate. Very similar to the way that the California Senate is trying to push through or is debating right now a rule that would exclude urban

construction, downtown construction, I think especially in infill areas and also near transit, exempt that from CEQA, which is the California equivalent to NEPA. I think that's a really clever policy. I would love to see various ways of accelerating permitting. The Biden administration passed the IRA and then did not get behind Manchin-Barrasso, which was the permitting bill, which is absolutely like the yin to the yang for clean energy. You can't cite if you also can't permit the long distance transmission lines

And so I would love to see something like that be passed into law. I would like to see more governors and mayors take a page from Josh Shapiro, who a couple of years ago when the I-95 bridge fell down, declared an emergency declaration and used it to wipe away environmental review, permitting processes, procurement processes, and basically rebuilt the bridge that would, under normal circumstances, have taken 24 months and 12 days.

And I know emergency declarations considering the tariffs are a fraught subject right now. But what if Wes Moore, what if Jared Polis, what if Karen Bass, what if more Democratic politicians were less afraid about recognizing the proverbial fallen bridges in their cities and states

as fallen bridges, as emergencies, and announced that they wanted to dramatically accelerate the process of working on these problems. So that's more at the sort of like institutional layer than the policy layer. But I think that would be an interesting idea. I'm just all about it. Like getting politicians to, you know, take some risks and then do something amazing.

All right, rapid fire. Is there a movie, song, or book that fills you with optimism for the future? Ezra had a good line on our tour the other day where he took a question about optimism and redirected it toward curiosity, which is to say there's a kind of optimism that is synonymous with curiosity. You believe that by answering questions, you can learn more about the world and make it better. And the book that fills me with the deepest sense of curiosity that I've read

recently was When We Cease to Understand the World by Benjamin Labateau, which is absolutely brilliant, very, very strange, and just made me deeply curious about scientific discovery and the madness contained within. It's very strange and very wonderful. Awesome. What is a question that you wish people would ask you more often? What are you thinking about?

I have a very loud internal monologue. So at any given time, there's someone in my head screaming at me about something. Typically it's like nice screaming, but that question can always be answered. What are you thinking about? Love it. Awesome. A variant of that, which is where do you see progress or momentum outside of your industry that inspires you?

I think what's happening at the intersection of AI and medicine is incredibly interesting, and I'm really, really rooting for it to go well. And our final question, can you leave us with a final thought on what you think is possible to achieve if everything breaks humanity's way in the next 15 years? And what's the first step to get there? I'm really, really interested in the advances in blood tests and proteomics converging to create incredibly powerful diagnostics for patients

all sorts of things. I'm talking about something close to an all-purpose blood test that's like, why don't I feel good? But that also serves as an all-purpose blood test to what are my risk factors for a thousand different maladies whose protein proxies or surrogate endpoints we've worked out in various AI models. The ability to essentially take an X-ray of your cellular genetic

and biochemical health would be absolutely sensational and requires a lot more advances in proteomics and genomics and a lot of other nomics and blood tests. But that's a technology that I feel like is both sci-fi and also possible. Derek, thank you so much for being here today. Really appreciate it. My pleasure. This was fun. Awesome. Possible is produced by Wonder Media Network. It's hosted by Aria Finger and me, Reid Hoffman. Our showrunner is Sean Young.

Possible is produced by Katie Sanders, Edie Allard, Sarah Schleid, Vanessa Handy, Aaliyah Yates, Paloma Moreno-Jimenez, and Malia Agudelo. Jenny Kaplan is our executive producer and editor.

Special thanks to Surya Yalamanchili, Saeeda Sepiyeva, Thanasi Delos, Ian Alice, Greg Beato, Parth Patil, and Ben Relis. And a big thanks to Eva Cairns and Alex Premiani.