Kumar Garg: Science for the people

January 26, 2022 1:01:38
Kaltura Video

Former assistant director of the White House Office of Science and Technology Policy Kumar Garg and director of the Science, Technology, and Public Policy program Shobita Parthasarathy discuss deploying science, technology, and data for the public good. January, 2022.



0:00:25.8 Jason Owen-Smith: Good afternoon everyone. On behalf of Dean Michael Barr and the faculty and students of the Ford School of Public Policy, as well as the Science Technology and Public Policy Program, it's a great pleasure to welcome you all to this policy talks event. My name is Jason Owen-Smith, I'm a professor of Sociology and Public Policy, the Executive Director of the Institute for Research on Innovation and Science, and a faculty affiliate of the Science, Technology and Public Policy Program, which is known as STPP.

0:00:56.1 JO: STPP and IRIS are two of the host of today's events. STPP is an interdisciplinary university-wide program dedicated to training students, conducting cutting edge research and informing the public and policy makers on issues at the intersection of technology, science, equity, society and public policy. If you'd like to learn more about it, you can do so at our website, stpp.fordschool.umich.edu.

0:01:25.2 JO: IRIS is a data repository that anchors a national consortium of universities who share data to understand, explain and improve the public value of research and higher education. Before I introduce this event, I wanna make a couple of quick announcements. First, the next event in our lecture series is a panel discussion titled, "Cultivating socially responsible engineers, the role of universities in public policy."

0:01:47.7 JO: It will be on Monday, March 21, 2022 from 4:00 to 5:30 PM. We currently plan to hold the event in person, but it'll also be live streamed. And for those of you who might be interested in the STPP graduate certificate program, the next application deadline is March 1st of this year, and we'll be holding an information session about it on Tuesday, February 15th at 4:00 PM. You can access registration details for both the panel event and the information session on our website.

0:02:17.1 JO: And now for today's event. Our featured guest today is Kumar Garg, Senior Managing Director and head of partnerships at Schmidt Futures. In this role, he works to help all major Schmidt Futures programs find successful leverage, as well as helping to run the technology and society portfolio. Garg previously helped shape science and technology policy for President Barack Obama's administration, serving in a variety of roles in the White House Office of Science and Technology Policy or OSTP.

0:02:44.0 JO: This included spearheading efforts to bolster STEM education, including the Educate To Innovate campaign, major State of the Union initiatives to train STEM teachers and bring computer science to all students, and the creation of iconic events such as the White House Science Fair.

0:03:01.9 JO: He was also involved in policy development, implementation and communication on a wide range of science and technology issues, including advanced manufacturing, behavioral sciences, biotechnology, broadband, digital media, entrepreneurship, the Maker Movement. The Maker Movement. [chuckle] Space, nanotechnology, and prizes. He holds a Bachelor's degree from Dartmouth College and a law degree from Yale Law School.

0:03:31.6 JO: In conversation with him will be Shobita Parthasarathy, Professor of Public Policy and Director of the STPP program at the Ford School of Public Policy. She studies the governance of emerging science and technology in cross-national perspective, and has written widely on intellectual property, biotechnology, artificial intelligence, and more broadly, equity and innovation policy.

0:03:52.9 JO: The event's hosted by the Science Technology and Public Policy Program and the Ford School of Public Policy and co-sponsored by the Michigan Institute for Data Science, MIDAS, the Institute for Research on Innovation and Science, IRIS, and the U of M Office of Research. Following the discussion, our speaker will answer some questions that were received in advance from registrants. Thank you to everybody who submitted. Shobita, I'll turn it over to you.

0:04:19.9 Shobita Parthasarathy: Wonderful, thank you so much, Jason. Kumar, I'm so excited to have you here today. There are so many directions that our conversation could go in, and hopefully maybe we have an hour, maybe we'll be able to touch on all of them. But first of all, just thank you for joining us today. I think it's gonna be really fun.

0:04:42.4 Kumar Garg: No, I'm super excited about this. I'm a big believer in public service, so I'm hoping at least some of the folks who are listening get inspired to wanna do their own tour of duty in government.

0:04:52.1 SP: Absolutely. Actually, I'd love to start there maybe, and maybe later we'll talk about your own trajectory, which I think is incredibly fascinating. But you work for Schmidt Futures, a philanthropic initiative that is focused on the intersection related between science, technology and society, and I'm wondering for your organization, what does that mean in practice? And then maybe more generally, what do you think the unique role is for philanthropies in today's science and technology landscape?

0:05:27.4 KG: It's a great question. Because I cut my teeth in public policy, I sort of think of largely one should approach science and tech innovation from a public policy lens, which is trying to think more broadly about what are the things that drive advancements in science and tech, and then what are ways to apply those advancements towards larger national parties and social outcomes.

0:05:53.9 KG: And so one of the things we did when I was in the Obama administration, we did it twice, was actually develop the first ever national innovation strategy, and it's actually ended up being quite useful just because I actually think all these different things can just feel like a bunch of words.

0:06:14.1 KG: So one day you're talking about basic R&D funding, next day you're talking about broadband, next day you're talking about immigration policy, next day you're talking about patent policy. Well, how does this all add up? I think the core idea that animates at least my personal work is, if you look historically, more than half of GDP, like the actual wealth of a country is generated through advancements in science and technology and its application.

0:06:42.5 KG: So it's like our wealth-driving engine. And then before we have any larger debates about distribution of that wealth, what is that engine that drives that wealth? And when that engine is working, it drives a huge amount of prosperity both in the US and the world. And one of the things is, is this just like an automatic fact that we have that engine? Actually it's not. The United States has actually made over the past 150 years major investments.

0:07:09.3 KG: One of the things that President Obama used to say was in the midst of the Civil War, President Lincoln actually invested in, created the first National Academy of Sciences. And during, in the run up to World War II you have major advancements in building the infrastructure for basic R&D that drives a lot of innovation forward.

0:07:29.4 KG: So I think one question that animates the thinking is, as we sort of think about the threats and opportunities for the 21st century, how are we investing as a nation, as a world, in the actual underlying science and technology infrastructure that can drive that forward? So for example we're obviously living through a world in which every day you'll see a cover about mRNA technologies and how they created the life-saving vaccines.

0:08:00.0 KG: But I think that sometimes this gets treated as like maybe we found a needle in a haystack. But actually this is the result of decades of work that the... That was in... Of both investments in a range of different areas. And often something that we invested in, whether it was on Ebola or on developing a therapeutic and vaccine on AIDS, ended up mattering a lot for this crisis.

0:08:27.7 KG: And so I think one question that both should animate government and can animate philanthropy as well is trying to think about what are... What does that median future look like? And what should be the sets of investments we should be making now? So done well, when we were in OSTP, we tried to think a lot about what's coming and what are ways that we should be working on that.

0:08:55.7 KG: I remember I spent a lot of... Now, there's been a range of advancements on... There was a big question around what should be the sets of investments we should make when the Ebola crisis was happening. I feel proud of the fact that some of the folks at OSTP made the strong case that there was a set of experimental investments that were happening at DARPA on mRNA.

0:09:25.6 KG: The big question there was, "Well, what if this doesn't end up helping on Ebola?" And they made the case, "Well, these are platform technologies, they could end up mattering both for Ebola and for something else." And those investments ended up mattering a lot. So government when it does this right, it starts to think about that investment now that can happen in the future.

0:09:48.1 KG: So for example, I think that we should be thinking now about pandemic prevention, which is, there's a huge set of international organizations that are fighting to invest in, how do we think about the 100 likeliest candidate for the next pandemic and how do we map them now? This is something that Congress and others will have to invest against.

0:10:07.8 KG: And so I think philanthropy can do a couple different things related to that. I think one thing it can do is just issue-mapping, so there's a lot of things that I think feel obvious once people start working on them, but before you fully articulate it as a live policy topic can feel a little bit obscure. And so I think agenda-setting can play a huge role. So.

0:10:33.1 KG: I remember when I was in the administration, we did a major initiative on the brain and mapping the brain. A lot of that was actually a set of workshops that the Cogley Foundation sponsored that brought together leading neuroscientists to say, "There's actually advancements in neuro-technology that are happening that might allow us now to be able to study tens of thousands of neurons versus just a single neuron or just doing brain-mapping," and that middleware technology was gonna really matter.

0:11:01.3 KG: So one, I think question is, is like, what are the sets of areas where additional work could be really important? So one of the areas that we've worked on is something called "convergent research", and what focused on is this idea of a focused research organization, which is, in various biological and other fields, there are often things that are deep bottlenecks that matter, and what are ways that you could actually identify a team that could work for five to seven years in a pretty dedicated way specifically on resolving one of those bottlenecks?

0:11:38.6 KG: And so that's something that's actually quite hard right now, government doesn't... We have academic centers that we fund, but often that's actually pretty distributed research, or we have individual PI research. But large scale, high-risk research on bottlenecks ends up mattering a ton, even when you don't know if it's gonna work out. So I think that's one area, both kind of agenda-setting and issue-mapping is one area.

0:12:07.6 KG: I think a second one that can be quite useful is just thinking about what are areas where actually developing a to-do list of what actually needs to happen. So for example, one of the things that the scientific community is really worried about is antibiotic resistance. And the hard part about that is no one knows what to do, which is, we know that 60,000 people already are dying a year from it.

0:12:39.2 KG: And we know that if things get worse, you might need... We might get to a point where it's too dangerous to have general surgery, because we don't know if you might be able to recover, and that might cause that number to go to 2 million a year. But the problem is there's no easy public policy answers, because the problem with antibiotics is they're not a particularly good business. You take them for two weeks and then you're better.

0:13:07.2 KG: How do you actually design market mechanisms to invest in new biologics that have that potential, and then store them and not use them, kinda requires a very different business model than we have today. And that might require a bunch of work and experiments. And so that's the other piece that I think philanthropy can play, which is both designing that to-do list and potentially de-risking it by doing some of the early pilots.

0:13:32.5 KG: When I was in government, that ended up mattering a lot. Government can be a really powerful scaler, but often, you can get the idea going. So Operation Warp Speed actually had its roots in advanced market commitment on a vaccine that the Gates Foundation and five other countries did 10 years prior, and a bunch of academic work from Michael Kramer and Susan Athey and others.

0:13:58.3 KG: So it's like this dialogue that's always happening between the research community, the scientific community, policy makers. And I think done well, philanthropy can help de-risk and get more shots on goal. But I think government ends up being the essential player in all of this, and so I'm a deep believer that it's not a substitute, but a partnership.

0:14:23.9 SP: It's so interesting. One thing that I wanted to also draw out in your conversation, which links to some of the things that I know you've been involved in, is that it isn't just about the substance that is placing bets on particular areas of areas of innovation, but also to some degree, placing bets on styles of innovation.

0:14:45.4 SP: So you've talked about convergent research, but I know you've also really been involved in discussions around open innovation, and that you recently published a piece on the role that open innovation has and also could potentially play in health innovation in particular. I was wondering if you could talk a little bit about that. What is open innovation from your perspective? What do you see it as particularly good at addressing?

0:15:14.0 KG: Open innovation is just a fancy word for having a wide lens for who could have the answer to a problem. So there's a... I got interested in this because Tom Kalil, who I worked with in OSTP, had been very... He had written previously about the role that prizes and challenges can play in bringing non-traditional performers to a solution.

0:15:44.7 KG: One of the interesting things is that... So we started working on something that became challenges.gov, which is actually the first federal platform, so that federal agencies can actually sponsor prizes. And so 10 years later, we've had hundreds of millions of dollars in prizes and a ton of new solutions. And often what you find is that we have too narrow of a view of who might have a tangible solution to a problem.

0:16:09.6 KG: So a classic one is, there's a number of organizations that partnered with Harvard around a diabetes challenge. Different therapeutics in diabetes and solutions. But they opened it up to beyond just the medical community, and what they found was a number of the ideas actually came from patients, which once you actually think about it, kind of makes total sense.

0:16:34.6 KG: There's a number of ways that somebody who's actually a patient might actually be able to come with interesting solutions that might be relevant. But that's the big mistake. Think about who applies to NIH grants and NSF grants. It's pretty limited by, you have to have a PhD. You have to have an academic institution. So even as it might seem obvious that solutions can exist and come from other quarters, we don't actually design our systems to make that possible.

0:17:02.6 KG: So I think there's a viewpoint diversification logic to it. And the question is, is like how do you then build systems to allow more people to raise their hand? I think open innovation is never like a silver bullet, because you actually have to create systems for people to do that. But what I found is that it can actually create new channels even for the traditional models to sort of play off against.

0:17:35.3 KG: And so, for example, one of the things that we continue to sponsor now is challenges on platforms like Kaggle. So Kaggle is a large data science platform. You can post a challenge and then hundreds, if not thousands of teams will form to develop new solutions. And then some of the solutions end up being quite different than what people expect and you can actually get new collaborations out of it.

0:18:01.8 KG: So I've been a huge fan of this model as one that can widen who gets involved. And then I think there's an emphasis around just democratization of science and technology. One of the big public debates we're having is around trust. I think there are different ways to engender trust. One is like, "You should just trust me." But I think one of the things you actually get, I always think is, you can build empathy for any process if you were a part of it.

0:18:37.4 KG: I think it's incumbent on folks in the scientific and technical community to think about ways to not just think of the broader public as people to communicate to, but actually to build into their design and solution set. So one example I remember that was quite interesting was during Ebola, one of the design problems was the suit that healthcare workers used when they were on-site. And the reason is, is that it would take like two hours to take off, and someone else had to help you take it off. So the risk of contamination, that was very high. There's a very high mortality risk if you get exposed.

0:19:23.6 KG: So the actual solution came from folks in the design community, like folks who actually design garments, which is they actually developed like three pretty viable designs for a suit you could take off, just by putting the zippers in the right spot, which is like a totally different solution. So I think those kind of things can both be inspiring, but can also make people feel connected to the work in really important and powerful ways.

0:19:49.3 KG: And we need... We need lots of people to actually see themselves as this call to action. I think one of the things that I'm intrigued by is this generation of students are gonna be the COVID generation. I don't want them to have a deficit mindset towards that. I want them to have a mindset that they actually saw up close how much this is a problem that's gonna be with us, and that science and technology can actually be part of the solution.

0:20:22.2 KG: And so we're already starting to see that, but I think we wanna encourage it. Because I think that's how we're gonna develop new solutions versus saying we're having the same mass debate that we had 100 years ago.

0:20:32.7 SP: So I can't tell you how much you're speaking my language right now. [chuckle] I think in two respects. I think your point about the potential benefits that democratizing innovation can provide in terms of serving people, but also, and this often doesn't get raised, which is why I'm so excited that you did, which is that it's also tied to questions of public trust and engagement and empathy.

0:20:58.9 SP: And I actually just published an article in Issues in Science and Technology that lays a lot of this out and talks about how do we institutionalize this kind of practice in policy institutions, in both policies, but also practices at research funding agencies, etcetera. And so in some ways this question is kind of selfish, but it's something that I've struggled with in thinking through this.

0:21:25.5 SP: But it's also something that, in my defense, it wasn't just me. In fact, a version of this question came in from one of my students. We won't talk about whether that student has been overly influenced by me or not. But I think that the challenge in all of this is that you provided these examples of the diabetes patients or the healthcare workers working in Ebola. But that might happen, so they might come up with this wonderful solution. And then unfortunately, I think that we don't...

0:21:56.0 SP: We may have more open, more democratized and some inclusive grassroots innovation processes, but the problem is that we don't, it seems, have a ton of creativity when it comes to taking that innovation and ensuring that it actually serves the public in an equitable way. So we only see the private sector or the marketplace as that conduit, which often means then that it isn't the diabetes patient or the healthcare worker that's getting the intellectual property or who can control who gets access.

0:22:34.5 SP: So the prices might end up being higher and then the diabetes patient can't get access to the innovation that they in fact came up with. We know of those kinds of cases. And of course, on some level of course, the situation around COVID and the global access is sort of one of the manifestations of this.

0:22:54.3 SP: And so I'm wondering, how do you think about that? How do you think about kind of opening up the solvers, the field of people, but then ensuring that the benefits are also, to some degree, democratized? From your perspective, what are the kinds of things that innovation policy makers or philanthropies for that matter can do in that space?

0:23:18.9 KG: Yeah, I don't think there's an easy answer. But at least one way I approach it that I find motivating is, there's a lot of analysis around various public problems that isn't actually particularly well-grounded in what is actually happening. So I'll give an example. So there's an organization called Think of Us, [0:23:54.3] ____, that focuses on foster care youth.

0:23:58.7 KG: And the interesting thing is, they spend a lot of time actually interacting with foster care youth around the actual challenges that are happening around kinship care and foster youth getting pulled from families and everything else. And what they find is that most of the policy discussions that are happening are actually not that connected to what they're hearing from the actual foster youth that they're interacting with.

0:24:31.1 KG: The optimistic piece of that is when they collect that information and they share it with policy makers, it's like electric. It gets very persuasive. So for example, there might be some rule that said, "The grandmother cannot take the kid in because she has a cat. And we can't test for whether there might be an allergic reaction to the kid. So oh, we're gonna actually put them in foster care." I don't know if anybody thinks that that's the right policy answer. And once they can collect those use cases, they can actually advocate for policy changes.

0:25:06.8 KG: So I think one is, what's a structure by which you actually can have organizations that are working really closely with the community that's affected? I think the other one is actually designing high quality ways to then give that feedback. There's another organization that focuses on the tenant complaints around landlords that might be engaging in abusive practices. The challenge is, is that it's really hard to file a violation. Like where do I go? How do I file it?

0:25:43.0 KG: And then, especially in housing, the problem is, is like you don't actually know who the true owner is, and so you might file your application, but actually they are like a repeat serial offender. There's an organization in New York City called JustFix that we work with, where they will actually then kinda unmask who the true owner is and actually be able to tell New York City housing, "These are the folks that are repeat offenders across a range of units that you should be able to take action on."

0:26:14.3 KG: So my sense is that it's not that the public policy response is weak if you can do the work, but often there are not that many organizations that are actually doing that, "We're identifying the problem and then we're structuring that information in a way to actually get it to the person who can make a decision and change what they're doing."

0:26:40.4 KG: I call it the... As someone who was in policy, the biggest mistake people do is to think of policy advocacy as diffusion, like I put something out into the air. And somebody will hear it. So some academic, when I was in these sort of roles, would be like, "I published a paper on this," and then they'll send me the link.

0:27:04.9 KG: And I would be like, "You don't understand my email. You don't understand what the work is, which is, you have to tell me what is the thing I'm allowed to change and why I should change it and what I can do." And if you do those three things, often, people will actually call you back and say, "I didn't know we could change that. And how can I take next?"

0:27:26.2 KG: So one idea I've gotten more animated by, both when I was in government, and especially afterwards, is this idea that we actually need a field of policy entrepreneurship, which is the actual work of taking insights and developing them into ideas, and then actually delivering them in a format to policy makers at all different levels of government in ways that they can take action. Because that last mile problem ends up stopping a lot of good ideas.

0:28:00.4 KG: I met a woman last year who had this really interesting idea around small airports end up being these huge areas for lead pollution. 'Cause the gas that gets used for small airplanes and everything else. And she said, "Oh, if we could electrify some of these places and change the following two rules, it could have a big impact."

0:28:22.3 KG: And I was like, "Okay, so what would you want DOT to do?" And she actually had a proposal. She worked on it and got back. And so we sent it to DOT and that rule is getting implemented. And so there's a huge, I think sense of agency one can have, if you can do that connection. And I think that's one of the roles I try to play as someone who was formally in government, it's just a sense that it's neither impossible nor is it automatic. And that there is this middle space where actually advocacy for high quality ideas can go a long way. I see that all the time.

0:29:02.4 KG: And so that would be the way I would think about this idea of equitable, because I actually think we're not... We're making it all seem very distant. And if we actually worked with lots of different communities and said, "Here are tangible ways to advocate for things you want changed," on a range of things that could actually happen. And often there's a, "If only I... There must be some secret room that I'm not... "

0:29:36.2 KG: And I always say that most policy... The big debate, if you were watching TV, is it's all about prioritization. Is it gonna be these three things that are gonna pass this year through the Congress? But issue four through 100 is not actually debates about prioritization. It's just about debates about readiness. So is there a specific idea I could do that actually allows me to make action on it? And so there's a range of those things right now happening inside the government at all levels, where one can have a pretty tangible impact.

0:30:16.6 SP: So what do you envision in this policy entrepreneurship program? Are you thinking... So we have money now from the Ford Foundation to develop partnerships with community organizations. It sounds like with some similar ideas, but focused very much on what are the needs of community organizations vis-a-vis science and technology, and how can we either serve as a catalyst to technical experts, or how can we help them anticipate the implications of a particular emerging technology issue and intervene.

0:30:48.9 SP: But I'm wondering, what do you imagine in terms of who might be a client? What are the kinds of skills that they might need to develop? It seems like part of it is about communication, but it also seems like part of it is also about networking. You have the networks that so many community organizations, for example, don't have as much.

0:31:09.7 KG: Yeah, yeah, yeah. No, I definitely get somewhat about network and somewhat about communication. I guess the piece of it that I really emphasize is this idea of actionable ideas. I'll give an example. So all the time, take the problem of disinformation. There are a 1000 op eds and reports that have been written. But what I do is I just flip to the end and I say, "What are the ideas?" Right?

0:31:38.1 SP: Right.

0:31:38.2 KG: Most of the ideas boil down to, "We more research," or "We should have a panel," or "We should have a commission." But the reason is it's really hard and we don't actually have that many "X should do Y" tangible ideas on this topic. So if somebody actually did, if somebody said, "Oh, I actually have an idea for what the following federal agency or someone could do. And here's how I would... And here's the way we would know whether it's getting any traction."

0:32:04.3 KG: I actually think it could actually have a lot of salience. And so for example, I personally am very animated by issues of lead pollution. Because it's a neurotoxin in the water and in the walls and I think it gets very little public policy focus. Obviously, there's been a bunch of work that's happened in the past year that's passing around lead pipe removal.

0:32:29.1 KG: But the interesting thing I think about it is that I started to just ask what were the... Let's say a mayor was motivated on this topic, but like they had to do it in their existing budget. What are high value set of activities that they could do? And it just takes you a bunch of questions before people say, "Well, you can't replace all the pipes 'cause you're not dig... " So you actually need, "Here are methods you can do to figure out which 1% of pipes are like to be the most likely."

0:33:06.2 KG: "Okay, well, how would you figure out those 1% of pipes?" "Well, we could use these set of data science plus historical record, plus the... " "Oh, is anybody doing that?" "Well, actually, maybe someone should." So what you find is that like if you can actually get to the thing, "Do this thing," then it's not automatic, but often you can actually find people on the receiving end.

0:33:30.8 KG: That's the only thing I would encourage, whether it's your students or everyone else, to be like, "Have they actually gotten this down to X should do Y?" And that Y is not think more about it, convene more people on it. No, you don't get any... My joke always is, most, these large mega reports where you have lots of folks convening, is that it's hard to get lots of smart opinionated people to agree, so they always have three recommendations. Do you wanna guess the three recommendations of a large multi-researcher report?

0:34:06.2 SP: Well, as a reader and occasional participant on these reports, frequent reader, occasional participant, I know that recommendation one is, "We should convene another committee."

0:34:19.7 KG: Exactly, 'cause it's hard to get people to agree. So it's more research is needed, more information is needed, and we should have a national panel, maybe a speech from the president, that kind of recommendation three. And I totally get it. I mean, it's not like any of those things would be on net bad and probably would be helpful, but it doesn't give you the policy maker that much to do.

0:34:46.2 KG: Because if you said, if you write a report on lead poisoning and it's that we need more research on lead poisoning, we need more coordination on existing activities and we should sign a national spotlight, you're like, "Well, that's why I'm reading this report. I'm like looking for actionable stuff."

0:35:01.3 KG: And so that's what I mean by policy entrepreneurship, which is it's doing the next two or three things in the sub that you were hoping that the policy maker would do, but like let's assume that they are cognitively overloaded and you are actually giving yourself the task to be their staff. What would you want?

0:35:27.8 KG: And then what, the interesting you find is that it's a surprising, you get a surprising amount of, "Please help," because there's a range of problems where once you do the leg work, it ends up being quite fruitful, and there's no one yet doing the leg work and you can actually get the ball rolling.

0:36:00.3 SP: It's so interesting that you're raising this. So the graduate certificate program at STPP, which Jason mentioned, so most of our students are scientists and engineers, graduate students, so Master's and PhD students, and they come to the Ford School for the policy training, and partially because they're really interested in being policy entrepreneurs, and the course that I teach, so I see a lot of them in at least one of my courses.

0:36:30.0 SP: We're teaching them relentless policy memo writing, like, "Make an argument, be persuasive, be concise, and cut it down," and it is, one of the reflections, and I think you sort of were saying this and I wanted to put a fine point on it, is that the conventions of science, like scientific investigation are to be incredibly careful about drawing conclusions.

0:37:00.7 SP: It is just sort of spotlight he limitations of the work that, "Here's how narrow it is. Here's the conditions under which we think this phenomenon works." And they are so... And honestly, this is true of my science students, but it is also something that my other students, my policy students, my MPP students learn when they've taken these courses, is that we have to be very careful, we have to be very nuanced. There are no agents, it's that sort of we talk in passive voice.

0:37:34.7 SP: And yet the policy world, in order to have that impact, you're kind of pushing them really hard in the other direction, and it is... I think there is, must be a moment, and obviously when you're doing it in a classroom, it's probably different than when you're out on your own, but it must be so psychologically jarring for students.

0:37:55.6 SP: And to some degree, I have a science background, so I understand that a little bit. But to go from this extreme nuance to be just saying like, "Right, you have five minutes in an elevator, get to it." There's no time for nuance or caveats or any of that.

0:38:13.9 KG: So, there's a couple of different ideas there that I think are really interesting. One is, so I used to have this dynamic, so my role when I was in OSTP was I was like the teammate. So we would bring in some prized academic and, "Hey, why don't you work with Kumar, and why don't you guys... " There is some initiative or some, "What's that next step?"

0:38:38.5 SP: So often I would get, it was like a very privileged position to be in, and I would be like, "Okay, you know a ton about manufacturing or the future of manufacturing. Let's work together on what you think needs to happen." My lived experience of watching this all the time is that it's very hard when you're like a world expert on something very specific that took you 20 to 30 years to develop that sense of authoritative voice, for you to then say, "Oh, this other thing, I'm gonna read about it for a couple of days," and have an authoritative voice.

0:39:13.1 KG: 'Cause you were taught you had to earn that authoritative voice through a huge amount of prior effort, and the thing I would have to coach them on is, "It's either you or some 22-year-old." That policy keeps moving, and so either you're gonna give your reasoned best judgment, or the system will just pick. And it's not that the substitute is gonna be, we're just gonna wait six to nine months or 12 months or four. So that's one dynamic.

0:39:46.3 KG: But I think the harder part, I think of your question, so one is just like you have to get over it and just try. The harder part is actually that it actually is a learned skill for what I think of it as being able to think about problems at the mid-scale. So let's say you're someone who's been doing cutting-edge work in neuroscience. You might actually... One of the things you train on is, what are the next interesting research questions that you or your lab should pursue.

0:40:23.6 KG: Or you have the kind of view that in general, neuroscience deserves more funding. But if I said to you, "What are like five to 10 really interesting problems in neuroscience that are like, that the field you try to grapple with, if it really focused, you could potentially have a real breakthrough over the next five to 10 years?" Kind of a portfolio approach.

0:40:51.2 KG: That's actually kind of hard because you kind of have to be up to speed on the field, you have to know what are the things that people consider hard, but becoming less hard, where focused work could really make a difference. And what I found is that in policy, that mid-scale thinking is very powerful.

0:41:14.4 KG: Because if we're having a conversation with President Biden and we were saying, "You know what Mr. President? We can bring advanced manufacturing back to American shores, and what we need to do is invest in a set of advanced manufacturing institutes, the great clusters to bring these technologies." You'll get a lot of nodding heads and then they'll say, "Okay, so what are the five topics that are emerging, but where we could... "

0:41:44.2 KG: There's no right answer. There's not like... We can't see into the future. But so you have to be able to talk to a technical expert and say, "Oh," so we had this conversation in 2011, and we said, "Well, additive manufacturing is gonna really matter. Where we're starting to see the early signs that it's working at a very small level through things like the Maker Movement and others. If this hits industrial capability, we're gonna be able to build rocket engines additively rather than subtractively."

0:42:12.4 KG: Well, that ended up being true. Right now, there'll be a huge number of companies that are leading, and many of them are American, right? But that was like a bet we made 11 years ago, and now those companies are hitting in the park and going public. And so thinking about that story is a really interesting, at least in science and tech, ends up being really this interesting story where you neither wanna give in to just a fad, something is exciting today, and you're like, "Well, it's just gonna be exciting."

0:42:49.5 KG: While at the same time, you don't wanna give in to just pessimism that like, "Oh, we're just stuck." There's like, breakthroughs are happening all the time and you wanna be open to that. Some of the people I worked with were there during the early '90s, in the Clinton administration when the internet was just an early idea, and they just, they have all these experiences of people in senior policy making levels being like, "Why is this important? And why should I care? And what are you talking about?"

0:43:23.9 KG: But they were actively curious about this thing that was emerging, that the research community is excited about, was starting to get... And that kind of open curiosity to emerging areas ends up being really important. So that part I find really hard. I'd be really curious how you guys think about it. There's a skill you have to develop to be able to approach fields at that five to 10 years out range versus, this is an interesting question today, versus this is an interesting question at some point, but I don't know when.

0:44:00.1 SP: And I would add to that maybe also, this is an interesting question, given real-world constraints that I'm sort of being faced with, whether that's from policy makers and budgets, or if whether that's from communities, organizations. You talked about your interest in lead and of course, in Michigan, we know about the real consequences of that, and I know that the university has been... University researchers have been very engaged in the community in exactly the kinds of things that you're talking about, but in super challenging... That have been super challenging for them.

0:44:30.7 SP: So Jason has a question. So he says, "I very much like the idea of engaging broad scientific communities in defining a portfolio of mid-scale opportunities. Much of the direction of current policy appears to be aiming at large-scale multi-field or convergent innovations."

0:44:50.2 SP: "What do you think are the challenges of engaging multiple communities to define mid-scale interventions that might span the interest of particular fields? And how might you avoid the problems that you identified in the context of large multi-disciplinary policy reports and panels?"

0:45:07.1 KG: Yeah, I think it's a great question. I think "interdisciplinary" sometimes just gets thrown around. It's like a, just as a headline.

0:45:15.4 SP: Oh yeah. [chuckle]

0:45:19.2 KG: And so convergence can have that same feel to it sometimes. So I think there's a couple of different methods one can use. So one method that I find very powerful just as a conceptual method, is what are methodologies or tools or platforms that are considered obvious in one field, that are considered odd in another field?

0:45:41.0 KG: If you take that methodology or platform or way of working into a related field, you start to get interesting reactions. So I'll give an example. So in a number the physical sciences, you have this notion of the telescope, the collider. You have this idea of large, persistent physical infrastructure that the field identifies and builds, and then the job of the researcher is to identify really good experiments and then run those experiments on those, on that equipment.

0:46:13.3 KG: One thing I find really interesting is like, we'll take that analogy of common research infrastructure and think about what are fields where we make the researchers do all of that on their own. So an example I think is education, right? So in education research we'll focus a lot on, the individual researcher design the instrument, recruits all the study subjects, runs the study and then publishes the result.

0:46:40.5 KG: And so you don't actually have anything that looks like a large persistent architecture where all the researcher does is design the experiment, and you can have thousands of experiments running at the same time. And so then you ask the question, well, could you actually design such a system in education research?

0:46:57.5 KG: One of the things that we've been doing under our Learning Engineering program has been saying, "Oh, some of these digital learning platforms, which have millions of learners, can actually be that test bed where researchers can much more quickly design." So I find that analogy to be really useful.

0:47:16.8 KG: A similar one is like, what does use of the cloud and self-driving look like in biology? So most researchers in biology are running their individual experiments in their lab. It has been very hard for the field to conceive of, I can actually design the experiment and just, and it runs in some third party area, and it can actually run at the capability, so I can be able to get a huge amount of orders of improvement. That's something that one can aspire to.

0:47:55.2 KG: So that's one form of interdisciplinary I find interesting, which is, what are methods and capabilities of one field that compare with another, and what could you get out of it? Similarly, one area that I think in general that is an area of under investment is most of the time in research, we are under-investing in the platforms and tools and data sets that power those fields.

0:48:28.4 KG: So individual researchers might develop and fund those tools, but actually directly pointing at those tools ends up being like an accidental innovation. One thing that I think is very powerful is to think about if you were trying to accelarate a field, what are the sets of tools that could rapidly improve your ability to run and test ideas in much more powerful way?

0:48:53.6 KG: One emerging idea that I know a number of fields are looking at are, in what ways could you use things like machine learning to actually predict the next experiment? So right now we sort of, what is the next experiment? You actually don't... That's a decision space that's left to the individual researcher.

0:49:14.8 KG: But you could actually map out the next 10,000 possible experiments, and so this is happening a lot with materials, where you could say, "Oh, these are the next set of combinations of materials that are worth testing." And that's rapidly improving the rate at which we can find new materials with new properties.

0:49:32.9 KG: So the thing that jumps out to me is like, "Oh, that's... Could you actually develop a decision framework for the next experiment?" Well, take that idea and could you take that idea to other fields and say, "How could that change what we're capable of doing?" So I find that to be most useful, which is, what's obvious in one feels radical and different in another. And then the question to ask is like, is that just because of practice?

0:50:01.8 SP: And it also seems like these are, in all of these cases, that these are particularly fruitful places that philanthropies could actually play a really crucial role, right? Because so often the National Science Foundation is siloed and they're structured by scientists and university disciplines and departments, that then shapes the programs at these funding agencies.

0:50:29.5 SP: And so creating this kind of convergent work, or creating incentives for producing platforms or tools, when the kind of tenure-track rat race doesn't necessarily support that. It seems like that's a unique place that philanthropies could potentially play.

0:50:49.5 KG: Totally. I think we're definitely in this kind of moment for science where there's been a, just in the past year or two, there's been a flurry of energy around bringing the experimental method to science itself and saying, "What are different models by which we could be supporting science?"

0:51:10.0 KG: Some of this... A lot of them have historical roots, funding researchers not the ideas, funding tools. And I think that's actually really healthy because a lot of the ways that we fund science globally actually come out of the United States, and those ideas actually got crystallized to a set of structures that we built right after World War II, and have become the modern system for doing science.

0:51:35.6 KG: I think there's room for continued experimentation, innovation. So I think one example is that like all our research agency in the US are largely define themselves as field-led. They say, "We ask the researchers what are the most important areas for investment and we drive it."

0:52:00.1 KG: But then they've been recently polling of actual researchers and researchers will say, "You know, if you do not have to follow the requirements of your grant agreement, would you do what you wrote or do something else?" And 40% say they would do something else.

0:52:12.4 KG: So then it's like on one level, the research community takes pride in the fact that it actually... The researchers are actually the grant-makers themselves, NSF, NIH are on rotation, researchers are actually making those decisions. But on the other hand, researchers themselves express a lot of frustrations about the system that they are in charge of. And so it means that there's actually a lot of room for improvement.

0:52:33.2 KG: Because it's actually a conversation the research community is happening with itself. It's not some political prerogative that is driving the way the research community is doing its grant funding. That's actually a decision that the community is making for itself. Which means on the optimistic side, it can actually change that view and actually develop new systems and new methods. I actually think this is a healthy moment.

0:53:03.4 KG: Sometimes Americans are famous for complaining, but the nice thing about complaining is you get results faster. I actually think that this is an area where the American spirit of complain, if you don't like the thing, complain and then actually start to come up with solutions, is a really healthy moment. Because, I don't know, we're living through this moment where a lot of what faces us this century is gonna be science and tech-oriented, and we're just gonna need the system revving.

0:53:41.1 KG: And so taking it as a given or locked, I don't know if it serves any of our benefits. For folks who are long-term players, I think there's a healthy amount of, what is it that we are living through, and I think it's actually we're living through the beginning of something.

0:54:02.2 SP: So, something on that note, talking to you on Wednesday January 26th, I cannot let you go without asking you about the COMPETES Act. Now the conversation has restarted about this COMPETES Act, that many see is potentially revolutionizing R&D. But of course there are some clear political dimensions to that. I'm just wondering what you think about it?

0:54:35.1 SP: What you think are the limitations, especially given what you've been talking about, what you think the promises in that? What would you like to see in that kind of legislation?

0:54:47.5 KG: I'm generally positive. I think there is a number of great ideas that have gotten into the process, and I think the push to increase NSF funding is a really good one. NSF historically, even though it funds a ton of the computer science research that happens in the country, quantum research, range of other areas, has substantially lagged NIH in funding.

0:55:12.9 KG: It was mostly an historical accident. The two agencies were running line with each other into the mid-'90s, and then just because of a set of circumstance the NIH was put on a doubling path. It's actually been a interesting history. I think NSF should start to catch back up, it's gonna be a very important agency in the work that comes ahead. I'm very supportive of that.

0:55:38.6 KG: I think the piece that I would, when we encourage lawmakers and policy staffers is, I think we just need a lot more experimentation built in into the way that we do this work. So for example during COVID one of the experiments that ran was called Fast Grants. Really small grants, but that happen in really short amounts of time. NSF actually has a capability to do really fast grants through its existing mechanisms, but Congress kind of actually actively encouraging it on priority topics to set that money aside and actually get it out the door, I think is a really strong encouragement.

0:56:19.8 KG: Similarly, I think places like Operation Warp Speed, where we're doing these dedicated investments around an advance market commitment, I think need... We should be thinking about where the next sets of AMCs that should get developed. There's lot of great talent provisions in there, but things like high skill immigration end up mattering a lot.

0:56:40.4 KG: There's things to work on, but the encouraging thing that I would say is there's lots of ideas that the field had, that have made it into some of the base bills that are getting proposed. It's just a sign of this theme that we've been talking about which is, ideas come from somewhere, and often they come from a student that writes up a paper or someone who works on it, and that gets lost once you see it in some bill text.

0:57:06.3 KG: But especially for any of the folks who are watching and listening, if they're eager to get involved, I'm happy to be a resource and very encouraging of folks raising their hand in this kind of way.

0:57:17.8 SP: That's super helpful. I'm curious what you think, do you think that the COMPETES Act will be successful? Do you think it's gonna get passed? Are there particular parts of it that you think are likely to get cut out of it as it go through the process, that are perhaps more challenging to particular political groups, or?

0:57:36.5 KG: I will be inevitably be wrong because it's like...

0:57:39.4 SP: That's okay. That's okay. [chuckle] We're not leveraging any money on this.

0:57:44.2 KG: I think the general view is that something will pass this year. Part of that is because we continue to have a pretty sizable problem when it comes to the semiconductor shortage. Currently, whether you call it the USICA or the COMPETES Act, we're gonna need to pass something that starts to address the fact that we have a large systemic risk in production. And that rather than just solving this immediate problem, we should be investing in the next generation of technologies that allows more of that production to happen here.

0:58:26.6 KG: So I think that discrete issue, I think is gonna be an engine that drives this conversation forward. I also think there's a healthy interest both in the House and Senate to do something. There's been some areas that they've had to work over, but I think they will work over. So I remain positive that we're working towards a bill that will pass.

0:58:52.3 KG: I'm a big viewer that you should make it better irrelevant of whether you think it will pass. Because often these things, when they eventually get legs, they will move very quickly. At that point the train is done, and so the moment before is the moment you have.

0:59:09.1 SP: Right. I think it's really important to pause and just note something that's implicit in what you're saying, which is that there is actually bipartisan support in R&D. They're specifically, at least to my knowledge, supporting increasing NSF's budget in this really substantial way.

0:59:31.4 SP: I think, especially at this moment when we're especially politically polarized, and there's a lot of charges about whether or not people distrust science and what Republicans in particular think about science, etcetera, etcetera. I think it's really important to remember that, I think for all of us, that it's a more complicated picture at the very least, right?

0:59:55.4 KG: Yeah. I think this is a hidden history that doesn't get talked about enough. A lot of biomedical research, a lot of scientific research has been championed by and passed by Republican Senators and House members and their staffs. It's a sign of national greatness, and it's a way that the United States leads the world. I think it's incredibly important.

1:00:26.9 KG: I think people come to it for different reasons. Sometimes it's fears about America losing its leadership potential, and sometimes it's about how do we solve very particular problems. But at its core, I actually think this is a pretty healthy and dynamic community that ends up being quite bipartisan, I think that's to its testament. What I have found really important is that you can get a lot of agreement when you're specific.

1:01:00.3 SP: Well, on those words of wisdom, we'll close now since we're at the hour. But I just wanted to thank you so much. I've really enjoyed our conversation, it's been so wide-ranging and you have expertise on such a variety of issues, it's been great to really pick your brain. So thank you very much.

1:01:15.8 KG: Amazing. Well, thank you for having me. Thank you for what you're doing. I think advocating for and teaching this generation will matter a ton. I think it's one of the highest ROI things we can do, and if I can ever be a resource to the school or to the students, folks can always look me up and I'm happy to do it.

1:01:31.6 SP: Wonderful. Thanks.