Fireside Chat Transcript:
Tyler Cowen and Tom Kalil
From the Day One Project's "State of Innovation" Launch Event
January 23, 2020
Dan Correa: We have now heard a fireside chat on the importance of inclusion, particularly within the tech sector and computer science. We’ve heard proposals on manufacturing, advanced manufacturing, as well as barriers to commercialization. We thought we would close out with a session that would be designed to be catalytic and inspirational. Two individuals who approach the topic of progress, science of science policy and institutions that are relevant to innovation and economic growth from different perspectives, and we’re excited to be welcoming Tom Kalil, who is the Chief Innovation Officer of Schmidt Futures, and Tyler Cowen of George Mason, to the stage to talk about the state of innovation and progress studies. Please join me in welcoming these two individuals for our final fireside chat.
Tyler Cowen: I’m very happy to be here with Tom Kalil today. Let’s start with meta-science. So, let’s say somebody makes you the policy czar at the National Science Foundation. What would you change?
Tom Kalil: Well, they have a very small program in the science of science policy, but I think that’s an area that needs a lot more investment. So, the idea is that we should apply the experimental method to how we allocate science funding. I’ll give you one example. Pierre Azoulay compared the scientific output of individual researchers that were supported using a traditional mechanism at the Howard Hughes Medical Institute, and discovered that betting on people, not projects, was actually more productive, and that’s an idea that was actually embraced by the National Institutes of Health with programs like the NIH Directors Pioneer Awards. We need to be doing more of, that is, testing new hypotheses about how we could get more bang for the buck for investment in research and development.
Tyler Cowen: Not all at once, but in terms of procedures, what would you change at the National Science Foundation’s peer review panels, what would you do?
Tom Kalil: I think one area that the National Science Foundation underinvests in are those areas that have the potential to be platform technologies. So, what are the areas that are not going to just lead to some isolated product innovation, but are going to be areas that have a large transformative impact on our economy and our society?
Tyler Cowen: Okay. Now it’s a week later, we’ve got a news article that you’ve moved to the National Institutes of Health , and you are at the end of your first week, what are we talking about changing?
Tom Kalil: Well, let me talk about one specific area that I think we’re really underinvesting in, and it’s, I wouldn’t say it’s the fault of the NIH, but it is an area where we ought to be doing more. So, traditionally, the way we’ve dealt with diseases is, particularly the chronic diseases of old age, one at a time, so we have an effort in cancer and diabetes and Alzheimer’s, there are a growing number of researchers who have evidence that we should be making a direct attack on aging, as opposed to the individual diseases associated with aging. So, that’s an area where I really think we’re underinvesting relative to the importance.
Tyler Cowen: So, we have you on an airplane, we’ve arrived in London, the House of Lords is now up in York and is trying to start a new science agency….. is he likely to succeed, and how do you assess the current situation in the United Kingdom and they ask you for advice, what do you tell them?
Tom Kalil: Well, I really think he’s onto something with respect to the importance of talent. If I look at the organizations that have had an outsized impact, it’s something that they’ve really focused on, so for example, a number of people have referenced the DARPA model. The reason that agency has been successful is they go out and they recruit people who are the peers of the best scientists and engineers in the country. So, I think his general idea, which is finding amazing people, I think is really important. A lot of the programs where you have brought people from outside, it’s important that you pair them up with people who have worked in the government for a while. I think you need a buddy system between new talent and people who understand how the place works.
Tyler Cowen: Are we in the United States underinvesting in weirdos? And what should we do about that?
Tom Kalil: Well, yes, we’re going to have a National Weirdo Initiative. So, I do believe we need to encourage a higher level of ambition. I don’t know whether that is the same as being weird or not, but one of the things that I worked on for both President Clinton and President Obama was, in the same way that President Kennedy said let’s put astronauts on the moon, and we decided to not only sequence the human genome but drive the cost of that from $100 million for the first one to a thousand dollars today, what are the similarly ambitious goals that we should be setting in the 21st century? So, for example, we worked with the agencies to launch something called the BRAIN Initiative, which was -- could we transform our understanding about how the brain encodes and processes information. The Department of Energy launched an initiative called SunShot, where the goal is to make solar as cheap as coal. In the private sector, Elon Musk has said, “I want to die on Mars, but not on impact”, so his modest goal is making humanity a multi-planetary species.
Tyler Cowen: So, Elon’s a weirdo, right? Tesla, already is succeeding, the rocket’s come back to earth, that’s amazing, he’s at least two for two… and there’s no one like Elon, you almost can’t aspire to be like him. Does that show something is wrong…is something wrong culturally with the United States of America, if someone as unusual as Elon has needed to be three for three with these amazing initiatives?
Tom Kalil: Yes, so I think, again, we need to shift more to the idea of being willing to bet on the individual’s success. So, I think that’s one of the things that gives someone the autonomy to pursue an idea
Tyler Cowen: Institutionally, let’s say you’re connected with a well-known foundation and you want to seek to change things to bet more on individuals. What is it you actually concretely do, do? Purely hypothetical.
Tom Kalil: Well, I could use some examples from the very real philanthropic organization that I work for. So, Dr. Holdren spoke about the importance of different disciplines, we’re identifying people who are brilliant in one discipline but have interest in exploring others. So, they have a Ph. D. in one discipline in the experimental sciences but may be interested in understanding computational sciences. We’re identifying brilliant young computer scientists who are interested in spending two years to explore different tech for good applications. So, that is one way in which we’re making bets.
Tyler Cowen: Purely hypothetically, let’s say you are working as a major science advisor for a two term president in this country and the president asks you, what do we do with these individuals? Other than your czar-like performance with the NSF, the NIH, what else would you tell the president?
Tom Kalil: Part of it is just shifting how we allocate the money. So, if we look at the way a proposal is evaluated, for example, with NIH, the researcher has this section called specific aims, so they’re saying, this is what I’m going to do. Right now, there is a very delicate balance because, on the one hand, you have to have preliminary results, but on the other hand, there has to be a little bit of newness because otherwise, the reviewers would say, oh, this person has already done it. We should have more programs like the Pioneer Award, where we’re saying Tyler has been innovative and creative in the past, and therefore, we think the odds are high that if we fund him for another five years, something good will come. There’s a great book on the rationale for doing this called Why Greatness Cannot Be Planned, and the metaphor is that the future of scientific and technological advances is like stepping stones across a lake with lots of mist. So the problem is you don’t actually know what set of stepping stones are going to get you across the lake, and number two, the feedback that you get is deceptive, so for example, the thing that wound up being really important for computing was vacuum tubes and if you were looking at vacuum tubes, you would not have been convinced that was important.
Tyler Cowen: So, you mentioned DARPA and ARPA before. As good as it used to be? And if not, what went wrong?
Tom Kalil: No, I think it is as good as it used to be. So I think it goes through cycles, so for example, they made some very brilliant investments in “biology as technology”
Tyler Cowen: When do we need a new agency altogether, so, there’s talk of this in the UK, that civil service regulations are so strict, they want to start anew. Should we consider a new version of the NIH, start from scratch, new version of the NSF, start from scratch, yes, no?
Tom Kalil: Yes, I think we should be willing to do that. I think what is really clear to me is that the United States is at least somewhat serious about applying science and technology to basic science, national security, energy, space, and health. But other problems, like economic and social mobility – we’ve not talked about applying science and technology to those problems. So, the Department of Labor research budget is zero. OMB prepares a chart every year that has R&D by agency and DOL is zero. But I would argue that stagnant or declining real wages, and the potential for that to get worse is a national problem, and I think science and technology would have ideas about how to increase the wages of non-college educated workers. So, I’m more focused not on what’s the right org chart for the federal government, but what are some areas where there’s systemic gaps.
Tyler Cowen: If you look at America’s universities, it seems they’re less popular than before. It may or may not be justified. They’re certainly more expensive, they’re more controversial, administrative cost is rising. What is there to be fixed, and how can we improve it?
Tom Kalil: I don’t know that I have the answer to deal with cost disease that’s associated with universities. I think that one organizational challenge that all universities have is that universities have their colleges and departments and disciplines, and the world has problems and there is not a very good interface between those things. So one of the things that I would like to empower, particularly at the undergraduate and graduate level, to allow students to major in major in a discipline with a minor in a problem. So, what is everything a student would need to know and to be able to do to make a contribution some problem?
Tyler Cowen: This is a bit like the idea to develop progress studies focused on the integration of theory and practice and that that is rewarded more at some institutional level. What should foundations do to be a part of that?
Tom Kalil: I think foundations have the ability to take more risks than they do.
Tyler Cowen: Why don’t they?
Tom Kalil: I think that failing is uncomfortable and people generally don’t like doing it. So, you have a choice. Why would people buy IBM in the 1970s? So, one of the reasons that DARPA has been successful is the program managers are shooting for the moon, and they can’t get a program approved if they’re not being ambitious enough. The program managers say, it's the only agency where, if we don’t develop the Internet we get a B.
Tyler Cowen: So, let’s say you went to a foundation but they have a lot of money and they ask you for advice, and you have your czar hat on. What do you actually concretely tell them to do?
Tom Kalil: So, let me describe one area that I think would have a big impact which is very inexpensive, and that is agenda-setting, and what I mean by that is to identify initiatives or areas that have the following characteristics. Number one, that if we achieve the goal it would a big deal, so for example, if I said, I have an idea for curing Alzheimer’s, I wouldn’t need to explain why that would be important. Second, something has changed about the world that leads us to believe that, while that might have been impossible before, it is now within reach. And that you’re also providing a roadmap for how to achieve that goal. So, the reason I think agenda-setting is important is that there are so many rich people that have not figured out what they want to do with their money. So, if you look at the 2,000 families with half a billion dollars or more in wealth their current level of philanthropy is only 1.2 percent. If you assume that they’re doing at least as well as the S&P 500, they could be dramatically increasing their philanthropy without any net reduction in their wealth. So, I think that this process of saying, being able to answer in different areas, where are we today, where do we want to be over the short, medium and long term, and what are some steps that we could take to achieve those goals.
Tyler Cowen: So, let’s say they agree with your advice. Over the next five years, what is the main pipeline problem they face? What’s the binding constraint they’re going to run up against stopping them from being more effective?
Tom Kalil: I don’t think they’re anywhere close to that.
Tyler Cowen: Pick your number, but what’s the first pipeline problem they’re going to hit?
Tom Kalil: Can I tell you what I want them to do? I think that there’s no shortage of areas where we have a market failure, that is, there’s a difference between the private return and the social return, and the social return is higher, and the private return is not high enough to induce private investment. So, I think one of the things that is interesting is that philanthropists have at least four components. They can engage in grant making, they can invest in commercial firms but from an impact perspective. They can engage in advocacy, policy change, and they can engage in coalition-building, so they can say, what is the coalition that we need to build in order to solve this problem. So, the UK government has estimated that the cost to the global economy of failing to solve the problem of antimicrobial resistance is a hundred trillion dollars. The UN has estimated the number of deaths from antimicrobial resistance to go from half a million to ten million. Companies that have successfully developed new FDA-approved antibiotics are going bankrupt, so if you are an investor, you would not be like, oh, goody, line up some more investments for me to make in new antibiotics. So, that’s an example of where we need not only more investment in R&D, but we need research to figure out what are the set of incentives that would motivate the private sector to act. So, that’s just one of a lot of areas where greater focus by philanthropists and foundations is needed.
Tyler Cowen: Based on a considerable and growing literature, it seems to me that progress in science might be slowing down. Productivity figures are sluggish, rising at one percent a year when it used to be two percent, crop yields are growing at slower rates and the Green Revolution is further in the past, Moore's Law seems to be slowing down. There seem to be more negative signs than positive. So, are we swimming against the tide here in trying to rescue American science and innovativeness? What do you think?
Tom Kalil: I think there is a lot of mileage that we’re going to get out of combinatorial innovation, which is that we now have a very large number of scientific and technological building blocks. So, not only are we continuing to improve existing building blocks and create new ones, but people are coming up with new ways to combine those building blocks. So, I agree that the number of engineers, Moore’s law has been slowing down, the number of engineers required to stay on Moore’s law is going up, and I actually used to work for Gordon Moore. We ended Dennard scaling a while ago, but people are coming up with new ideas. So, for example, you’re coming up with application-specific architectures, like the TPU or FPGAs, so I do think that there are ways out of the crisis, out of this challenge, but I agree that many areas, once you go from being able to put one transistor on a chip to ten billion, continuing that rate of progress is difficult.
Tyler Cowen: As you know, Peter Thiel has argued that today, a lot of contemporary culture tends to be about dystopianism and how the future will be terrible or fearful in some way, but the science fiction earlier in the 20th century, it was optimistic, and in some ways, it even underestimated the rate of progress we would make. Is there a cultural problem that we’re not pro-science enough? There’s no Gilligan’s Island where kids turn it on and they see the character of the Professor or they see Dr. Spock and they think, I want to be that way. That’s a long time in the past. What’s the Jetsons of 2020?
Tom Kalil: Yeah, so, it’s a really interesting, and there was a conversation between the president of ASU, Michael Crow, and Neal Stephenson, in which Crow basically said, you science fiction writers need to step up and do your part and write about a future where big things get done and where you would actually want to live. I mean, we don’t want to live in the world of Brave New World or 1984, and so, I do think it is important for our storytellers, for the people who do movies, to imagine a world where, yes, there are challenges, but it’s not some dystopian future.
Tyler Cowen: What’s your favorite book or movie in science fiction? You can’t say Star Wars, right?
Tom Kalil: No, I mean, the reason that a lot of people do dystopian science fiction is it’s just a lot easier, right? So, I think Minority Report is a great example of a future that was compelling, in the sense that they really did a lot of world building, but it wasn’t necessarily a future where you would want to live.
Tyler Cowen: And which book or movie do you want to live in?
Tom Kalil: It’s a good question. I’ve always been a big fan of I, Robot series, for example.
Tyler Cowen: That was meta science. A few questions about science. As you know, yesterday, China declared a complete quarantine of Wuhan, its seventh largest city, and I think the most important central rail depot in China, 900,000 people pass through that train station every day. How is science going to help here?
Tom Kalil: Sure. Well, this is something that Dr. Holdren and I got to work on during the Ebola outbreak. So, prior to the Ebola outbreak in 2014, I had been briefed on a DARPA program that was trying to figure out whether there was a way to dramatically reduce the time to go from bug to drug. So, the idea was you would identify patients who had been exposed to a pathogen and survived, you would isolate the most effective antibodies, and then you would create a synthetic oligonucleotide construct that would reprogram the muscle cells in your body to produce that antibody, and the end to end time required to do that would be dramatically less than the traditional vaccine development process. So, I’m not saying that’s the thing that’s going to work, but that is the type of thing that we should be thinking about. It’s not terribly comforting to have either an emerging infectious disease or an engineered pathogen and the scientists say, well, just give us ten years and we’ll have something for you, so we have to invest in platform approaches and we have to invest in approaches where the end to end process is a lot faster, and we have to try things like a universal flu vaccine. So there, the idea is, are there parts of the virus that are evolutionarily conserved, so if you built the vaccine based on that, you would have broad spectrum immunity from a new flu.
Tyler Cowen: Believe it or not, I know some people who are space skeptics. They say, well, when Nevada is crowded, I’ll start to think about Mars or the moon. How do you talk them out of that as quickly as possible?
Tom Kalil: I think it’s more an aesthetic issue than a rational issue. I just think that there are people who, if you say, it’s important for us to have hard, long-term things to work on, they will resonate with that, and other people don’t. So, I think it’s more like trying to change someone’s taste in music than it is a discussion about the adjusted net present value of the activity.
Tyler Cowen: What’s the next economically important thing to come from space. So, the network of satellites clearly, highly important. We have that, we could improve it, but it’s mostly there. What’s the next economic importance of up there?
Tom Kalil: Beyond communications and satellite imagery?
Tyler Cowen: Yes, which we have.
Tom Kalil: I’m not convinced that there is one. I mean, so, again, to me, the question is, do you think it is part of a thing for human civilization to do is to expand beyond earth, right? I think that boils down to the question of whether or not you think, that is, you are intrinsically motivated by that goal or it leaves you cold.
Tyler Cowen: You’ve been closely connected to the Mapping the Brain initiative, and I read almost every day, but certainly every week, about significant advances in neuroscience, mapping the brain, but I’m also a practically-minded economist, so just as some guy who lives in the suburbs, I want someday to go to the brain store, let’s call it the brain store, and what am I going buy from the brain store that has come out of the Mapping the Brain initiative? What should I be looking forward to at the brain store? Not science, not research. Something I want to buy.
Tom Kalil: Sure, sure, sure. Well, presumably, if you had a brain disease, you would want to have better diagnosis, therapies, cures, and prevention, right? So that’s one thing. I think something else could be new learning technologies that are informed by a greater understanding of how we learn, that’s something else that could come out of it. And I think neuromorphic computing, so if you look at the supercomputers that we’re building today, the supercomputers are going require their own dedicated power plant. Somehow, the human brain does everything that it does and it only uses 20 watts, so clearly, Mother Nature has figured out something about low-power computing that engineers don’t yet understand.
Tyler Cowen: Chess players typically take non-addictive drugs. They think it helps their performance. Is this all going make us smarter as a group of humans overall?
Tom Kalil: I have no idea.
Tyler Cowen: Okay. We now turn to a segment I do in many of my conversations, overrated versus underrated. You give your opinion.
Tom Kalil: This can only get me in trouble.
Tyler Cowen: Prediction markets, overrated or underrated?
Tom Kalil: Underrated. Yeah, I mean, I think that we should be doing more to allow them to operate without getting in trouble from the CFTC.
Tyler Cowen: If there were a prediction market in the future of smart cities, what would it say?
Tom Kalil: Probably overrated.
Tyler Cowen: Why, what’s the problem?
Tom Kalil: I think a lot of it has been too much technology in search of a problem, as opposed to really starting with the concrete problems we want to solve.
Tyler Cowen: The venture capital model as a means of funding science. Overrated or underrated?
Tom Kalil: It doesn’t fund science.
Tyler Cowen: Why not?
Tom Kalil: A venture capitalist is chastised for something called funding something that they refer to in the industry as a science project. But, so I think you mean more innovation than you do funding fundamental science.
Tyler Cowen: Look at Google itself. The search algorithm was based on a mathematical breakthrough, which is science.
Tom Kalil: Which was funding by the National Science Foundation Digital Library Initiative.
Tyler Cowen: Sure, but more needed to be done, right?
Tom Kalil: Yes, absolutely. No, I mean, look, I think that it works well for a class of innovations. So, it has tended to do better in those areas that are capital efficient, that is, that there’s a relatively short period of time between when the venture capitalist makes the investment and when they’re going find out whether the entrepreneur has product market fit. I think it does less well in other areas that are more capital intensive and take a lot longer.
Tyler Cowen: Prizes rather than grants as a way of rewarding scientists, overrated or underrated?
Tom Kalil: Well, I’m not a neutral observer here. So, I think that the thing which is underrated is increasing our ability to make financial commitments that are contingent on success rather than failure. So, right now, the federal government makes trillions of dollars in loan guarantees, which are financial commitments contingent on failure. So, the government is saying, Tyler, if you go bankrupt, then Uncle Sam will assume your debt obligation and we’re barely scratching the surface on our ability to make financial commitments that are contingent on success, so that includes not only incentive prizes but you talked about Elon Musk. The development of the Falcon 9 was supported by an agreement between NASA and SpaceX, which was a series of milestone payments that NASA paid for intermediate progress towards the development of the Falcon 9 rocket, and NASA set the goal, which is develop a rocket that can go to the International Space Station and deliver and retrieve cargo and gave SpaceX a lot of autonomy about how to meet that goal.
Tyler Cowen: It’s a big topic, but overall, Chinese science, are we too scared of it, or underrating it? What’s your view?
Tom Kalil: I think that right now, we have a lot of talk and no action, number one, and number two, the action that we do have is all in like a defensive crouch, as opposed to saying, they have double digit increases in investment in research and development. Why are we allowing our investment to decline as a fraction of GDP? So, that’s the conversation that I would like to be having, not solely focused on concerns about Chinese graduate students.
Tyler Cowen: The food in the White House cafeteria, is it better or worse than the food in the Google cafeteria? (laughter from Tom and audience)
Tyler Cowen: Okay. The final topic of talent. Talent is so often a binding constraint in organizations, and foundations, and research laboratories, and universities. What have you learned about looking for and evaluating talent that is not obvious and perhaps many of the people in this room do not know? How do you start thinking about that question?
Tom Kalil: So, one of the things I look for is intrinsic motivation, so rather than saying, okay, what are the carrots and sticks that I’m going to use to try to get more performance out of this person, they are highly motivated to do the right thing, and the only thing I have to do is help them remove obstacles from them making progress. So, I think that that is really, really important.
Tyler Cowen: I sometimes say personality is revealed on weekends. You ask a person, what are your open browser tabs, and see what they say, and that will tell you something about who they really are. What are the most difficult mistakes that other people make when looking for talent, keeping in mind that in areas of minorities, women, many other places we don’t do a great job in mobilizing the talent we have, what are the non-obvious biases we should have a better understanding of or awareness of?
Tom Kalil: Well, I don’t think it’s a mistake but I just think it’s lazy, which is, I think it’s very easy to go to high prestige institutions, so this is something that I was certainly guilty of, which is that you can rely on the signaling effect from where someone went to school, for example.
Tyler Cowen: What is your best non-obvious interview question that you use when you talk with candidates? How do you engage them?
Tom Kalil: Well, a question that I ask lots of people is, and my team is going to cringe because they’ve heard this like a hundred and fifty times, but you have a meeting with the president and he says, Tyler, if you give me a good idea, then I will call anyone on the planet and there can be more than one person on the line, there can be a conference call, what’s your idea, and in order to make your idea happen, who would I call and what would I ask them to do? So, I’m interested in what someone would do if they were not constrained by the resources and relationships currently under their control.
Tyler Cowen: Do you think today that talent or money is a more significant constraint on scientific progress?
Tom Kalil: Talent.
Tyler Cowen: Talent. Why?
Tom Kalil: From the point of view of, I guess I’m answering that question from the point of view of the government, that is, I think we would make progress more rapidly on a bunch of key issues if we had the right person in the right job.
Tyler Cowen: Let’s close this all up and tie it together in looking forward. If you were to sum up what you are doing as Chief Innovation Officer at Schmidt Futures, to realize this whole vision you’ve laid out for us, how would you paint just a short picture of that, what you’re up to and how you think about it?
Tom Kalil: One thing is, what are the areas where we need greater interdisciplinarity, and I’ll give you one example that we’re focused on, which is that, in the same way that we have hybrid disciplines like computational biology, one of the things that Schmidt Futures is convinced is we need to support the emergence of a discipline that some people are calling learning engineering. This is something that my colleague Kumar Garg is working on. So, we want people who are joint between computer science and the science of learning. So, we think that this could inform the development of new digital learning platforms, so something that Marginal Revolution University could take advantage of in the future, and it could also transform the way we do research on the science of learning, so that’s one example. But with respect to science policy, I think that there are two important things that we could do to advance the ball, and one of them is talent, how do we identify people who are interested in serving, and the second is ideas, how can we tee up some more ideas for either the second term of the current administration or the first term of an incoming administration.
Tyler Cowen: Now, to close, let me turn your own question against you. In your version of the question, it’s four people you get on the phone. In my version of the question, it’s forty people you get into a room and it’s the room you have here and then the online audience watching or listening to us, and what exactly is it that you’re asking them to do to help realize the vision you all at Schmidt Futures are laying out?
Tom Kalil: Yes, so, as I said, if I could identify one thing to work on, it would be on the talent side. So, in my experience, there’s a couple things that have been really helpful in doing that. One is, not everyone is going to be moved by a general call to public service. So, one of the things that happened during the Obama administration post healthcare.gov, is we made a big push on recruiting people with more technical talent, and the thing that was effective was not just, 'oh, come to the federal government', it was, if you come to the government, you can help reduce by 50% the time that is required for veterans to get their benefits. So, it was identifying something that was ambitious but achievable that they could accomplish during a tour of duty. So, what I’m interested in having this community and everyone who’s watching, is to come up with both those types of stretch goals that would motivate someone to come into government, and then, to identify the handful of people that would hit it out of the park in terms of achieving that goal.
Tyler Cowen: Tom Kalil, thank you very much, and thank you all for listening.