Denise Caldwell, NSF Division Director

After an academic career in Atomic, Molecular, and Optical physics, Denise Caldwell joined the National Science Foundation. Between her scientific focus and bureaucratic responsibilities, Caldwell was uniquely situated to recognize the value in supporting the Institute for Quantum Information and Matter in the late 2000s.

The formation of the IQIM at Caltech recognized that quantum information theory had progressed to a point where more sustained research geared toward engineering and experimentation - the "matter" in quantum information - became increasingly important. As one of the Physics Frontier Centers supported by the National Science Foundation, the IQIM is constantly pushing the outer reaches of the field.

Notably, for all the advances in quantum information over the past twenty five years, Caldwell sees current developments as suggestive that the field is just getting started. It is an important perspective, particularly in light of the massive amount of resources that private industry is pouring into quantum information. Wherever the field goes from here, it is clear that such support will be successful as a long-term proposition.

DAVID ZIERLER: This is David Zierler, Director of the Caltech Heritage Project. It's Thursday, March 31st, 2022. I'm delighted to be here with Dr. Denise Caldwell. Denise, it's great to be with you. Thank you for joining me today.

DENISE CALDWELL: Well, thank you for the invitation.

ZIERLER: Denise, to start, would you please tell me your title and institutional affiliation?

CALDWELL: My name is Denise Caldwell. I'm the Division Director of the Division of Physics at the National Science Foundation.

ZIERLER: Denise, do you have a background in physics?

CALDWELL: [laugh] Oh, yes. [laugh] I have a PhD in physics, which I got from Columbia University. After finishing my PhD in physics, in experimental atomic molecular and optical physics, working with what's known as photoionization, using higher energy photons, I completed a postdoctoral stint of a little over two years in Germany at the University of Bielefeld. I returned to the US, and taught for a time at Yale University on the faculty of physics, where I started a research program in photoionization of atoms and molecules that was funded by the NSF (National Science Foundation), actually. Then I moved to the University of Central Florida, again to the physics department, where I continued my NSF-funded research, and also contributed a lot of time to the development of what was at that time CREOL, the Center for Research in Electro-Optics and Lasers. That is now a separate unit within the University of Central Florida.

ZIERLER: I've interviewed Peter Delfyett, and I've learned all about CREOL. It's wonderful.

CALDWELL: I was hired at UCF when CREOL was just established, and I spent about the first five years of my time at CREOL. Actually, I mean, my time at UCF spending a lot of time pulling CREOL together, working with the local laser community, working across colleges at the university. We hired the first director. We hired some of the first faculty. It was a very heady experience. I can honestly say I really like to build things. I like to pull things together. I like to get things started. I like to see them grow and be successful. I left UCF when I came to the NSF in 1995.

ZIERLER: What were the circumstances of changing career course and joining the NSF?

From Academia to the NSF

CALDWELL: Well, I had had a productive period as a professor. I had a research program, an ongoing research program. I had students. I had postdocs. I had collaborations. It was very active. But around 1995, I had got to a point where we had completed the research emphasis that we had been working on, and I was thinking, where do I go next? What areas would I like to go into? I got a phone call from someone I knew who had been a program director–who was a program director—at the National Science Foundation in the physics division, asking me if I might be interested in coming to the NSF on a temporary basis as what's known as an IPA under the so-called Intergovernmental Personnel Act, which is what it stands for. This would be a temporary assignment to work and manage the program in the physics division in atomic molecular and optical physics. I thought I'd kind of come to a turning point, so where do I go from here? This just came out of the blue, I was offered the opportunity, and I decided to take it.

ZIERLER: Denise, I'll just note that at this point in your career, you were already very appreciative of the way NSF can support fundamental science.

CALDWELL: Oh, yes. I would say that my scientific connection to the National Science Foundation started when I was a junior in high school—

ZIERLER: Oh, my goodness. [laugh]

CALDWELL: —because the reason was, at that time, NSF had a program called Summer Science Training Camps. They were sponsored by and funded by NSF at universities across the country. The idea was to bring in promising high school students between their junior and senior year in high school to essentially spend 12 weeks in a summer science training camp. I call it boot camp because—I'm telling you—after 12 weeks of that, you knew what you were getting into if you decided to study science.

ZIERLER: [laugh]

CALDWELL: There was no doubt about it, and so I did that. Oddly enough, that particular training program was in organic chemistry and molecular biology. I grew up in Mississippi. It was at the University of Mississippi, so it was reasonably close by. But it turned out that [laugh] that was the only course in biology I ever took in college. But later on in the physics division, I actually established a program in the division that's now called Physics of the Living Systems, making the connection between physics and biology. The little biology that I knew or could learn, I took [laugh] from that summer science training camp. yes. I knew what NSF could do. Coming here was then an opportunity to see where the field was going, because I thought if I go to NSF, and I have this position, I will be able to see everything that's going on right across the board. Then it would help me think about where I would like to go from here. Well, I arrived in the summer of 1995, the same year and the same time that Carl Wieman and Eric Cornell had demonstrated Bose-Einstein condensation in atoms—and that was in my program. This was closely followed by Wolfgang Ketterle at MIT, again demonstrating Bose-Einstein condensation. It was a heady time. All of this was right smack in the middle of my AMO program. I also arrived when NSF had started a program in optical science and engineering (OS&E). It was an NSF-wide program, covered all of the divisions, all the directorates in NSF, and literally one of the early examples of an NSF-wide activity.

I replaced the person who had been heavily involved in establishing this program, and so I took over where he had left off. It was a wonderful experience because I got to know everybody throughout the Foundation since it was an NSF-wide activity. I met program directors from all of the directorates, all the fields, biology, computer science, engineering. Intellectually, it was just really, really exciting. That combined with the fact that my program was or what was going on in the AMO (Atomic, Molecular and Optical) program was changing the whole field. The division director was really excited, first of all, about Bose-Einstein condensation, and really supported the AMO program. When the opportunity came to stay, I decided to take advantage of it. I really enjoyed it. It's a wonderful place to work. The other program directors I worked with were just, I mean, every one of them was willing to go beyond, just do whatever they could for the benefit of the community. I certainly think it's a real service to the community that individuals are willing to devote their time and their expertise to promoting what they see coming from the community as the most exciting scientific projects, and so I decided to stay. I applied for the position, got the position, and then went on from that point to not only focus on the AMO part, but I worked a lot in plasma physics. Then I established, as I mentioned, the physics of living systems program, working with molecular and cellular biology, drawing on these connections that I had made in getting to know people and getting to know other areas. I established what is now the Physics Frontier Centers program, again working across multidisciplinary types of areas. I spent a lot of my time doing that. I have always found it extraordinarily rewarding to interact across the spectrum of science, because physics does reach out and touch pretty much to just about everything at some point. It was the right move. I really, really enjoyed it. In the end, the opportunity came to apply for the deputy division director position, and I was recruited [laugh] for that, to apply at least, and then moved on to division director in, I think, 2013, something like that.

ZIERLER: Denise, to set the stage, in the late 1990s, were you aware of what people like John Preskill and Alexei Kitaev were doing at the very dawn of quantum information?

CALDWELL: I certainly was, and I looked a little bit at the history of what I remember. In 1999, NSF sponsored a workshop on quantum information science.

ZIERLER: Do you know who got that going? Who at NSF was behind the sponsorship?

CALDWELL: One of the primary movers was the division director of the physics division at that time, Joe Dehmer. Joe also came from a background of AMO physics, and had come to NSF from NIST, where he had worked in the physics laboratory at NIST. First of all, I should say that the people I was funding at the time, people like Ketterle and Cornell and Wieman, and probably almost more than anyone else, Jeff Kimble, was in my program.

ZIERLER: You already had Caltech on your radar at that point?

CALDWELL: I had Caltech on my radar. I think Jeff had started at U. Texas Austin, but I think by that time, he was already at Caltech. But I funded Jeff through the AMO program essentially more in the quantum optics area, not necessarily precisely in QIS.

ZIERLER: Do you remember specifically what Jeff was working on at that point?

CALDWELL: I tried to thumb back through, but we don't have any paper anymore. [laugh]

ZIERLER: [laugh] But it was definitely quantum optics stuff?

CALDWELL: It was quantum optics. The one thing I remember Jeff really working on—and I think it was from this period—was QED and high-performance cavities, resonance cavities, whereby I can remember I think one of this activities was building these really, really, really high-finesse cavities so that you could almost contain one photon in the cavity. It was really pushing the limit of what you could do with these really high-finesse cavities for quantum electrodynamics. I would say that's probably the area that I most recall that he was working on at the time. Probably the defining point when it started to move over in quantum information science was around 2000. There was, again, one of these NSF-wide activities in the year 2000, roughly, called ITR. Information technology research was the global name of this program.

ZIERLER: This was an NSF program?

CALDWELL: This was an NSF program, information technology research, ITR. As I said, it was a very broad spectrum. There was a lot of focus on information technology. The primary focus, I would say, was probably on big computers, and data analysis, and the type of things you normally associate with more classical computing, and classical data needs. But there was a small, little fraction of that program that was pushed by a program director whom I knew very well at that time in the computer sciences directorate that was dedicated to quantum computing.

ZIERLER: Denise, was there partnership with the DOE and their interest in supercomputing for ITR?

CALDWELL: Not to my knowledge. But I was not heavily involved with the heavy-duty, high-performance computing side of the house. I got into it through the quantum computing side of the house. A lot of what you hear about quantum information science grew out of AMO physics,—well, I would say just atomic and optical physics, let's say, primarily, the types of things that Jeff Kimble was doing at the time, and others, and condensed matter physics. That's where the emphasis was, in these really preliminary areas that ultimately sort of started to collect until this whole notion of quantum information science began to well up as an area of particular interest.

Initial Support for Quantum Information

ZIERLER: Now, when the IQI got started at Caltech, when John Preskill and Jeff Kimble and others started to have this idea of formalizing this collective excitement over quantum information, was the NSF involved in that either formally or informally in any way?

CALDWELL: Well, remember that the NSF responds to proposals. In an activity like this, we don't work with any particular institutions, per se. But what we do is we try to identify based upon conversations with members of the community, conversations with people like Jeff and people like John and other members of the community whom you could name—people like Paul Kwiat at Illinois, people like Misha Lukin at MIT, Chris Monroe, Dave Wineland. This is the community who's really moving this field forward, and so we listen to that community. It was listening to that community across this broad spectrum of high-performance computing and things related to it that led to ITR. NSF then looked at where the science is going, and we ask ourselves, what do we see that seems to be an area of research that whereby if you put in a sudden influx of resources with emphasis, you can really make a big advance? That's where the original optical sciences and engineering that came from that I was involved with. It's where the whole nanoscience program at NSF came from, and it's where the ITR program came from. It is the recognition. It is what we hear in the community, and people like Jeff and John, and the others I mentioned in the community, are telling us that this is an area that really is ready. It's sort of ripe to move forward. It's starting to coalesce.

I would say, more or less, that there wasn't a direct connection between NSF and Caltech, or any of these that set it up, but the recognition from what we're getting from the community that this is an area we should be promoting. Then once we make a decision that this is an area to be promoted, then we create something like an initiative, something like ITR or nano or the others that we've done, and we issue a call for proposals. This is wide open call. It's not restricted to any particular institution. It's just a wide open call. This particular one—if I remember correctly—there was a component of this call that was sort of at the institute level. There were individual investigator smaller proposals, and then there was the institute level. I think Jeff, I believe at the time, was the PI. He was the leader or listed as the PI. John was the Co-PI, I think, at that time for the IQI. They competed in an open competition, and got the award. That started off the IQI. Now, a program like this typically will run for five years. There was an award up to five years, and then something else will come along for a big initiative. But, at the same time, what do you do with what you have started? In the physics division, I was instrumental at that time in getting a program established in quantum information science and revolutionary computing as a separate program within the physics division, just like particle physics and gravitational physics. This is quantum information science. Once the institute had completed its five years, and then I think it was John who took over and submitted a proposal to this new program for the continuation of what they were doing. Now, it wouldn't be on the big institute level because it wasn't the same amount of funding. But we did co-fund the new award with the computer science directorate. We funded that for maybe two cycles, I think, led by John. Then the physics division created the Physics Frontier Centers program. Caltech—I think it was still Jeff at the time—submitted a proposal to this Physics Frontier Centers competition in FY 2012 if I recall. I can tell you, a PFC competition is brutal.

ZIERLER: [laugh]

CALDWELL: It is a brutal competition [laugh], and they were funded. Then they'd been given a renewal at the next competition. This is a little bit the history of the evolution of what's the IQIM now.

ZIERLER: Denise, how did IQI in this initial brutal competition come out on top? What was so compelling?

CALDWELL: Science. The Physics Frontier Center always starts out with frontiers, physics frontiers. The word "frontiers" is aptly chosen because it really has to be at the cutting edge of physics. We fund a lot of really, really good stuff in the physics division, but there are certain areas in all of our programs that to a certain extent, to a great extent, push the boundary. And there are some that are more or less ripe. Again, it goes back to this is an area that is ripe, and an extra influx of funds on a larger scale can make a very big difference.

ZIERLER: Denise, was the impulse at this time to support quantum information as a theoretical pursuit, or were people already starting to talk about actually building scalable quantum computers?

CALDWELL: Both. But not just computers. There are a number of other areas where—

ZIERLER: Quantum sensing, things like that.

CALDWELL: Quantum sensing, right. In the physics division, we have a big program in quantum. A lot of our QIS is related to sensing. Some of our AMO is related to sensing, so there's quantum sensing. Then there were all of the issues with quantum networks and quantum communications. Now, within physics, we have mostly been connected to the quantum sensing and the quantum computing part of it. You ask, how did a proposal like this move? First of all, it has to make the argument that this is a compelling scientific frontier, and it is sufficiently compelling that it warrants and will move ahead more quickly with a center-level investment of funds.

ZIERLER: Denise, if we could zoom out where it's not just the science that's compelling but, for the NSF, the determination that the science should be supported because it's in the national interest. I wonder if you could talk about it from a sort of macro political perspective.

CALDWELL: I think there are two answers to this. The first one comes at what I'm going to call the basic fundamental level. The basic fundamental level is pretty much where we are by now, because I'm going to take us, chronologically speaking, up to about 2008, let's say, something like this, whereby you're seeing progressions in understanding the science. You're really beginning to look and see what can you do? What are the possibilities? In 2008—I think it was released in 2009—Carl Williams from NIST was over at Office of Science and Technology Policy, and we prepared a report on the potential of quantum information science for the future, because the idea of quantum computing had been around. But everybody had nixed it. Everybody had said, "No, never happen. You've got so much decoherence, you'll never build anything that's going to work like this." Sensing is on slightly safer ground because it's really kind of a measurement. But quantum computing was bah. It's like cold fusion. It'll never happen. But around 2008, OSTP issued this report, and this was followed up by a workshop that was led by John, actually. He agreed to co-lead, to co-chair this workshop in 2009 that NSF sponsored, to just go back and revisit the potentials for quantum information science. There was a report. There's a workshop report that John and company wrote, and all this time, there was an activity at OSTP. I was an executive secretary, I think, for a working group in the area of QIS. The difference came between, I would say, roughly at that time, that workshop, and I'm going to say about 2015, because the technological advances that we see during that period in what you could actually build, and also in other areas like quantum simulation, and what you could actually realize. You saw, for example, the ion-based computers that people were starting to assemble. There was sufficient technological advance between, I would say, 2009 and 2015 that suddenly opened everybody's eyes that this is not just somebody's imagination, but there are some real possibilities here.

ZIERLER: What were those real possibilities? What jumps out in your memory?

CALDWELL: I would say probably the first—Chris Monroe and Dave Wineland's work at NIST before Chris moved to Michigan and then on—that you could actually potentially control these states--quantum control--if there's a difference between the pre-QIS timeframe and the QIS era, I would say the difference is quantum control.

ZIERLER: This distinction, this would be basically pre-2000 and post-2000? Is that the demarcation, roughly?

CALDWELL: I would put it closer to 2010, because people had been working on quantum control up to then.

The Importance of Better Lasers

ZIERLER: What were some of the advances from 2010—I'm sorry—from 2000 to 2010 that makes that point this dividing point?

CALDWELL: I think I would say better lasers.

ZIERLER: Better lasers. Interesting.

CALDWELL: Right. Just a better—and it's not one thing because all of these things had existed before. You knew how to make cold atoms. You had such control over your lasers and over your traps, for example, that you could at will generate a predetermined state, and then have detection capability that allowed you to monitor that state either by probing with the other lasers or by just watching a Bose-Einstein condensate fly apart. It's not any one thing, but probably the time where the things that people had been looking at kind of began to work on a global scale; that it wasn't just one really refined thing in one particular area that could do something, but there were lots of things you could possibly think about. If you want to build a quantum computer, and you want it to do something, you have to know what your input is. That's where control comes in. To try to describe it without going into technical terms, (I've been working on trying to do this for a long time, and I've never really totally succeeded.) But we talk about, what do we measure? Let's suppose you do an experiment, and you have an atom, and you let it collide on something, and you look and see where it is afterward. What you typically measure is a probability, that is, you just detect it. But you don't necessarily measure a lot of find details. But for quantum computing, you have to know those details, so it's not just enough to know where the thing is. You've got to know what it actually looks like. This is really an understanding of the quantum state, not just, "Well, I know that it's here, and if I look, I'll see something." But I know that this quantum state has to be at this point, and I can read it out. The whole idea is that this state, I create it here, and it goes its own way, and I test it here. How it goes from here to here, and what happens, that's quantum evolution. If you think about it in the Feynman terms, this is his interpretation of quantum computing. But it only works when you can create in advance with pre-knowledge that very, very well-defined state, and you can hope that you can put that state in somewhere that nothing can influence it, that it just can go off and do its thing, and then later on, you can see this is what it did. That process is actually a calculation, and so you go from one to the other. But you can only do it if you can note exactly what you have created to start with. I would say that it wasn't any one particular new thing, but the fact that all of these approaches to control, to confinement, to the reduction of interferences had just improved to the point at which you started to notice a real difference.

ZIERLER: Denise, what were the laboratories doing laser research in the early 2000s that were really at the vanguard of these improvements?

CALDWELL: Well, there were a number of them. I would say, first of all, there was Caltech.

ZIERLER: Who at Caltech? Who was doing this at Caltech?

CALDWELL: It was Jeff. It was primarily Jeff whom I knew. I don't know whether Oskar Painter had arrived there by that time or not. He came more or less from the condensed matter physics side of the house, so I was not that well acquainted with him. Of course, I knew of Alexei Kitaev from the theory side of the house. Experimentally, it was Jeff. In theory, it was John and Alexei.

ZIERLER: What was your sense of how these experimental advances were refining and improving the theory in quantum information?

CALDWELL: Well, I don't know, because I would have to say I think, if you go back to Alexei, some of those first papers—and Jeff also did some theoretical work—to a certain extent, the ideas were there. The ideas were there. If you think about Alexei's papers, these are the seminal notions of people like Alexei, Peter Zoller who had been at NIST for a time and then moved to Europe. The theory or the theoretical notions were really alive. I think the real breakthrough was the experimental realization because if you're going to actually build something, you've got to realize it experimentally. This was always the criticism. You can talk about coherence as much as you possibly want—or lack of decoherence—but until you can actually build it, it's not going to be good for very much.

ZIERLER: This proof of concept that ultimately forms the seed for the IQIM, are you following this in real time? Are you aware of these connections even before this decision in the proposal arrives at the NSF?

CALDWELL: Oh, yes. I have been involved since 1999.

ZIERLER: What's the mode of connecting? Are you coming to seminars? Are you on email lists? Are people just sort of like messaging you, "Hey, Denise, we've got amazing stuff going on"? How are you keeping tabs of all this excitement?

CALDWELL: Well, a number of ways. First of all, my programs, because I had the program. These proposals were coming to my program. This is really the first way that we at NSF get informed of where the science is going, the proposals we get. That's what the proposals were doing. I had the proposals to keep me informed. I had the reviews of those proposals, which meant that I had the expert advice of people in the community who knew something about this. That's number one. Number two, I would go to the regular meetings like the DAMOP meeting, the Division of Atomic Molecular and Optical Physics meeting. I represented NSF at these sorts of things.

Then as an outgrowth of this first involvement in optical sciences in engineering, I was involved across the board in NSF activities that ultimately were related to this in one way or another. I mentioned I established the QIS program in the physics division. Now, I had proposals coming in just for QIS, so I had those. I kept in very close touch with colleagues particularly in NIST, with Carl Williams at NIST. I knew Carl extremely well, and Carl has been a promoter of QIS since the very beginning. He recently retired. As he commented, for over 20 years he and I were partners in crime, promoting QIS within the US government. I interacted closely with Carl. Then, as I mentioned, when Carl went to OSTP, I was heavily involved with writing this document that we first issued. I was executive secretary of the working group. We held meetings with other agencies. A man by the name of Henry Everitt at ONR, I think it was. No, he was in the Army, I think. used to organize regular meetings of all of the government agencies who were interested in quantum, in QIS. I was the NSF representative to these meetings, and so I knew what the other agencies were doing. I knew what they were saying. I've been sort of knee-deep in it [laugh] for well over 20 years now.

ZIERLER: Denise, where is quantum error-correction in all of these developments? Because in order to have the IQIM as a viable project, there needs to be sufficient optimism that quantum error-correction is something that can be dealt with effectively in the long-term.

CALDWELL: Yes. I think John was working on this—again, this is something that we supported, I think, if I remember correctly, through the physics division because we funded what's now the Kavli Institute for Theoretical Physics. If I remember correctly, John headed up a workshop there way back when on quantum error-correction to just see could it be done? It's important. Maybe I did talk for a little bit with John on this one, but he's the expert so don't take my word for it. [laugh] He knows a lot more about this than I do. I think quantum error-correction is important, but I worry a little bit—and this is a little bit, as I say, my own philosophy. I think we make a mistake if we start to think of quantum computing in the same way we think of classical computing. One of the things that I learned was that programming a quantum computer is totally different from programming a classical computer. This is one thing that makes it difficult, and it's not just that for quantum computing—in a classical computer, you get a stack. You've got this machine on the bottom, and then you got a couple of layers that gets to the user. The user writes a program. Somebody has written a compiler. That compiler takes that program, converts it into machine language, and the thing goes back and forth. This is the computer stack. That doesn't exist for quantum, and so the challenge for the field is how do you get from the machine to that user, who doesn't care about the details of the machine? They just want an answer to the problem. But quantum computing is not there yet. Developing the stack is very difficult, and one of the things that I have learned—at least if I understand correctly—is that making the connection between the algorithm and the machine may involve a lot of manipulation of the stack as well. It's not like you can just do what you do with classical computing right now. But you really have to think of the stack as a whole. I think a lot of the work, a lot of the thinking is, how do you build the stack? Now, error-correction we know how to deal with in classical computing. You just do a repetition. Well, you can't do that in quantum, and so the idea is whether or not it can be done, and John had—the proof is that it can be done, I think. There are some companies who are really focused on error-correction. They say, "If we can do error-correction, then we know that it can be done. Theoretically, it can be done." But how do we implement it? The difficulty is that when you start implementing error-correction, the number of qubits you have to add to the system grows rapidly just in order to do it. Here, again, I'm not the expert. I am not a computer scientist, and I would defer to what John or Alexei would tell you about this. But I think that what we might see happen is that error-correction in quantum might look a little bit different. I don't know that. But it's important. From a computational perspective, you've got to trust your system, and so error-correction, knowing that there has been no decoherence is going to be critical, or that if there is, how much it is. That's going to be really, really critical. All of this is connected to fidelity.

ZIERLER: When you talk about critical, just to be clear, there's a very solid line between resolving the problem of quantum error-correction, and actually building a scalable even commercially viable quantum computer?

CALDWELL: It could well be if, in the end, the only way you can handle the decoherence question is through error-correction.

ZIERLER: But there might be a way around that, you're saying?

CALDWELL: There might be. I think there might be, and there might not be. But we have, for example—and John will tell you this—the so-called NISQ machines, Noisy Intermediate Quantum Systems. Anyway, whatever, NISQ, John will tell you what it means. But these are machines that are noisy, and so the community now is trying to ask the question, what can we do with noisy machines? Can we actually perform a reliable calculation with noisy machines? People are really working on that, and trying to just see, what can we calculate? If you can calculate something with noisy machines, then you can think about probably at the scientific level making some progress and doing some real science with just the noisy machines. However, if you want to go to the universal machine--so are you thinking about a machine that a scientist would use in a laboratory by taping and twiggling and making it work, which is usually the first step, and something that your general user can use, the universal quantum computer? Probably for the second one, I think it's in the second one that error-correction is going to be most critical. [Ed. note: Caldwell subsequently requested that technical statements on quantum computing should be derived from the experts in the field; what she provided here is what she referred to as an "outsider's opinion."]

ZIERLER: Denise, this is an extremely important point, and I think this'll be our last question for today's session. There's so much more to talk about. Hopefully, we can meet back again. It's an incredibly important observation you're making. If I'm understanding correctly, what it sounds like you're saying is that, ultimately, the proposal and the successful pitch to NSF to jump in and support IQIM with both feet is that for you, for NSF institutionally, having a game plan to demonstrate that quantum error-correction is resolvable is not really a hang-up. In other words, you don't need to see that perfect pathway from getting to where the field was in 2010 to however long it's going to take that says, "If you can demonstrate that quantum error-correction is a resolvable problem, then we will support this." That was not part of the calculus.

CALDWELL: I'd have to go back and get John to show you the proposal. We weren't—even in those days--we weren't that far along. The proposal was, just like in all of the Physics Frontier Centers proposals, the proposal would have to be, say, this is a cutting-edge area of science. It's the cutting-edge area of physics. We know a lot. We have got the right group and the right team, and we've got it appropriately put together that we will significantly advance this field.

ZIERLER: But it's about fundamental research, and not translation?

CALDWELL: That's right. It's about physics. But there's one other part of the Physics Frontier Centers program that you can't forget, and that is the workforce preparation part, because that proposal, and what they would have to demonstrate, was what they would do and how successful they believed they would be in training the next generation of young people—mostly young—who would move this field forward.

ZIERLER: Because this is a multigenerational scientific effort?

CALDWELL: If it's a frontier, and you're on the frontier, you might push the boundary but, frankly, you're probably not going to be around to see [laugh] where it's going to go.

ZIERLER: Meaning that in 20 years, graduate students will still be excited about this?

CALDWELL: That's right, and I think you see that. I think you really see that. You go to some of these PFCs - and I've been out to the Caltech one a couple of times - and the students are bouncing off the walls.

ZIERLER: That's right. That's right.

CALDWELL: They're just so excited about this. This is what you have to have. When it comes to our Physics Frontier Centers as a whole, I would say that this is one of the most successful aspects of all of the centers. That you go to a meeting—when we do our reviews, our periodic reviews, you go out there, and you meet—we always meet with the students. We kick all the faculty out, and have the meeting with the students and postdocs. The level of excitement—

ZIERLER: [laugh]

CALDWELL: —and the ability is just amazing.

ZIERLER: Denise, on that note, it's a perfect place to pick up for next time where we'll discuss the actual ideas that came together for the proposal to make IQIM.

[End of Recording]

ZIERLER: This is David Zierler, Director of the Caltech Heritage Project. It's Monday, April 11th, 2022. I'm delighted to be back with Dr. Denise Caldwell. Denise, wonderful to be with you again. Thank you.

CALDWELL: Good to be back.

ZIERLER: Denise, we're going to pick up right at that point where we left last time, where you get word that IQI is thinking about developing a proposal to bring matter into the equation to make a formal NSF center, IQIM. Just to set the context in Washington, where are you at this point? What's your position? What are you responsible for?

CALDWELL: I'm going to look very quickly at my dates just to make sure. I think I've got them. The first PFC would've been around 2010.

ZIERLER: What's that? That's NSF speak. What is PFC?

CALDWELL: The Physics Frontier Center.

ZIERLER: There you go.

CALDWELL: IQIM is one of the physics division's Physics Frontier Centers. This is a program that we initiated in 2001 to build a centers program—as we talked about, I think, the last time we chatted—focusing on essentially the frontiers of physics, as we see it, to really develop a scientific program involving multiple individuals under the argument that by working together, we could advance the field much more rapidly than just giving all these people an individual award; that there would be some value added through bringing them together in a center. This is the Physics Frontier Centers program, and it was established in 2001. The Physics Frontier Centers program has an open competition every three years, and it's open to the entire community. There's no statement in advance that it has to be in a certain area of physics. It only needs to be in an area of physics for which the physics division has the primary responsibility. It's an entirely open program. At that time, I was already the Deputy Division Director in the physics division. But I also, as part of what I was doing at the time, managed the competitions. I managed the program, that is, I managed the competitions. Then when the awards were made, I exercised oversight over the new centers. When the competition opened up, Jeff Kimble—if I remember correctly—Jeff was the PI. He was, at that time, the director, I believe. I'd really have to go back and look. But I had known Jeff for many years, and had funded him. I believe Jeff was the director. John was, of course, involved. They submitted a proposal. It was reviewed. The process for reviewing these awards is extremely, I would say, brutal. It takes about a year to go through a competition, starting sometime around August, with what we call pre-proposals. That is, there's an idea that is put forward in the form of a proposal, and then the physics division assembles a review panel of experts to look at those proposals, which are actually proposals for proposals, in a way. Then there is a recommendation as to which of those groups look promising enough to be invited to submit a full proposal. At that time, it was IQI. That group that had been the IQI was invited to submit a full proposal. Full proposals were extensively reviewed first by just ad hoc review by experts in the field. Then after that, a select subset of the proposers submitted would be invited to the NSF to make a presentation in front of another review panel. This is known as the reverse site visit,. Basically this is, "Here's what we're proposing. Here's how we're going to manage it. Here's what we're going to focus on. Here's why we're going to make a difference if we do it this way." Then that panel makes its recommendation. Then based upon all of this input, the division makes a selection of which are the most promising proposals that should be funded. When that decision was made, IQIM rose to the top. I don't remember where it was on the list, but there were more than one. I think we funded probably three new ones that year. I'd have to go back and check. Anyway, it rose. It just rose out of this larger group, and was selected for an award.

Promoting Fundamental Science and Applications

ZIERLER: Denise, I want to ask a question that zooms out a little bit, and gets a sense of the things that motivate NSF to support particular endeavors and not the other. In the way that IQI became IQIM, which indicates that we're getting beyond a purely theoretical endeavor, and we can start to think about translation and application and, eventually, this is going to have social utility, how important is that at this juncture that NSF is not supporting purely a fundamental science endeavor? Of course, it's that. But it's a fundamental science endeavor that, at some point in the future, does hold extremely exciting progress and prospects of transitioning over into applications? Is that something that particularly resonates at that juncture that allowed the proposal for IQIM to rise to the top?

CALDWELL: Not particularly for the Physics Frontier Centers program, because there is a continuous and ongoing discussion about the value of basic science; the argument being that everything ultimately down the road that has the utility that you're talking about originated somewhere in basic science. You discover something. Just like quantum mechanics or just like the photoelectric effect, you discover something. Then as you learn more about it and build it out, you recognize that it's going to have uses. Now, we have programs in NSF, like the Engineering Research Centers programs that really focus on that ultimate end use, let's call it. The Physics Frontier Centers program does not specifically require that there be a specific end use that is outlined when the proposal is submitted because, the argument being, if it is the right kind of basic science, if you've got really good people, really smart people who are working on it, sometime in the future, there will be that end use that will come out of it. But you don't ask upfront that you have to define what this is going to be used for down the road specifically. If you think about something like IQIM, your argument could be that if you go back 100 years, and start with quantum mechanics itself, Max Planck introduced the notion of the quantum to just describe an experimental observation, and that was the only way he could make it work. He hated it. I read somewhere that he spent a great deal of his life trying to prove himself wrong, but he couldn't do it. On the other hand, Einstein recognized that there was something real about this quantum. It wasn't just a mathematical artifact, but it was something physical. That ultimately became the photoelectric effect. If you look at the 20th century, and where we have come with the 20th century, we are where we are because of quantum mechanics. It goes back to the very small. Once you could start in the very small, you have to do quantum. If you think about what would be the value of an investment in quantum at this level—because, after all, it's been around for 100 years—you could say, "Well, we know everything there is to know." Well, the thing about it is we don't. There are things that we don't understand, and in the late '90s came the notion that you might actually be able to use a quantum system to do computing, and this is where John comes in. He was very influential in it. If you could do that, then you could start to imagine many things. Also in the '90s, there was a man by the name of Peter Shor, who developed an algorithm, Shor's algorithm. If you could implement Shor's algorithm on a quantum computer, you would wipe out virtually all of the standard encryption security that we use today, because it's all based on the fact that what you have to do just takes a lot of computer time, and you don't have that much time. On the other hand, quantum is going to cut this down immensely, so it becomes a bit of an issue. But if you take it a step farther, you can think about, well, if I can do that—and it comes back to the notion of control—that once I can control, then I can start thinking about doing quantum chemistry—and people do. They do cold atom chemistry now.

The quantum computer, of course, comes in for these calculations that are just next to impossible, for many things, particularly as systems become more and more complicated and complex, you would like to be able—you could do things that you couldn't currently do, either at all, but you could do them better. Why is it important to understand this basic science? Those are the arguments. It's not so much that I would lay out a specific, "I'm going to build this, and then I'm going calculate that, and then we're going to put it on the market." That's not the way it works. But if once I understand all of the ramifications of this, and once I demonstrate what the possibilities are, then I open the door. Think about it as a door opening; not a specific but a door opening to—as we see it now—an entirely new technology. You open this door. Certainly, the importance of what the proposal contained, the importance of quantum information science derived from the fact, well, first of all, there were some interesting basic science questions, but it was so important that you answer those questions because you could simultaneously open this door to unknown possibilities—and that's what we like to do. That was part of it. But the other part of it—and I mentioned this last time—is that the Physics Frontier Centers program is really also heavily oriented toward training the workforce. Students, postdocs, and the opportunities that are offered to students and postdocs are a critical part of this Physics Frontier Centers program. When I say "critical," I do mean critical. It is our students, the products of these centers, who are those individuals who now will go on to take the next steps. Education and workforce development from day one for one of these centers is just as important as the scientific thrust. I've been to Caltech, and I met these students. I met these postdocs. They're fascinating.

ZIERLER: You saw the energy. You saw their focus.

CALDWELL: You see the energy. It's just you see it, you feel it, and it really carries you away. That's what characterizes a center, when you put together a center. Now, IQIM had a bit of a head start because they had the IQI, the institute before that then morphed into the center, so they had a bit of a head start. If a group starts out from zero, it takes about two to three years for a center to become what I would call an organic entity, to where it really becomes recognized for its own existence. That's what brings the students in. You talk to the students at many of the centers. You talk to the students, and they will tell you, "Well, I'm part of the IQIM." They don't say, "I'm part of the physics division or I'm part engineer." "I'm part of IQIM." Then when that center achieves that, that's success because, now, you are recognized. You're an entity. You go out, and I can go out, and say, "Oh, well, there's the IQIM at Caltech," and everybody immediately knows what I'm talking about. I would say that, if that answers your question—

ZIERLER: Absolutely. Denise, to understand the context here, when you're reviewing the IQIM proposal, is it apples to apples? Are there other quantum information projects out there that NSF is considering, or it's a whole wide variety of physics areas to fund, and this just was compelling at the time?

CALDWELL: It's entirely open. As I mentioned, when we write the solicitation, we do not say upfront, "We want to see a proposal in gravity, or we want to see a proposal in quantum." We don't specify a particular area of physics. We say that "you tell us what your compelling area is, and you tell us why it is compelling." The competition is wide open. The final panel has to struggle with whether or not the arguments are made that what I'm here proposing for this particular field is really going to change the way we think about physics. That's the way it's decided. We don't choose in advance. We just try to pick the most compelling cases for the science, the education and workforce, and the value of the center. That's something that's really important because if you look at the IQIM—John and Jeff and Oskar—they could go off and have their own programs. They're good. They could get funded, they could run their research, and probably be successful. Why would they want to be part of the center? What do they get? What is this—we call it the added value—what is the added value that comes from deciding we're all going to work together on what it is, and what we're proposing to do in this center?

ZIERLER: Denise, what stands out in your memory in the way that the proposal successfully articulated why it needed NSF to do what it had proposed to do? How did it put that together?

CALDWELL: That would [laugh] be hard. It's impossible, certainly, at this late date to state the specifics. If you can go back and talk to those guys, they should have access to the reviews from those years, and so you can read what the reviewers said. That should give you a good deal of insight as to what some of those factors were that you were talking about. That's been [laugh] a long time ago. [laugh] I couldn't tell you what the specifics are. I can only tell you they must've been compelling; otherwise, we wouldn't have funded them. [laugh]

ZIERLER: Denise, there's so many intangibles besides the binary, "We'll give you money or we won't give you money." Besides just the funding, and what that allowed to do in terms of building the center, funding postdocs, hiring people for salaries, and things like that, the intangible of just having NSF's imprimatur, NSF, the National Science Foundation saying, "This is something that we want to fund," why was that so important for the development of IQIM, in your view?

CALDWELL: Well, it's really hard to know. We like to highlight our Physics Frontier Centers because we like to say that our Physics Frontier Centers—I mentioned the IQIM—but for those areas in which the centers are active, I can go into that community, and I can mention the center, and everybody knows what I'm talking about. This is why we say that when we pick these groups to fund, they should be having an impact of the sort or of the level that they will be recognized by the community. If I mention any of the centers, the community knows what I'm talking about. In other words, they are recognized as among the leaders. That's what we like to see because then that tells us that it was the right choice for the field because we expect to see progress. We tell our centers, "You have to reapply every six years. If you get funded every six years, you're going to have to go through an open competition, that is, no holds barred; everybody is welcome to submit. Somebody can submit to supplant you. Somebody can submit to an area that's just better. If you look at the centers who've been funded over a period of time, if they had been successful, when they come in after six years, they should have morphed. They should have demonstrated that they're still on the cutting edge. In science, that means a sort of renewed way of thinking; a renewed focus; bringing in new people—very often, just bringing in a different cast of characters that we haven't brought in before. But our successful centers have been around for a long time. Typically, as they move from one six-year period to the other, they will change their emphasis. They'll follow the science. As the science moves, they move with it. Like I said, it becomes somewhat of a living thing that grows and expands and changes as the years move ahead.

ZIERLER: In the way that IQI was a theoretical endeavor, and the interest in bringing the condensed matter experimentalists and theorists on board, in what way did the proposal or the plan, as it was articulated, make it something new and not just an addition to what IQI already was, in other words, that IQIM would be something more than simply an addition based on adding matter to the IQI endeavor?

CALDWELL: Well, I think it changed it. It changed it. We have a number of centers—a very small number—that purely focus on theory, and they do great work, very valuable work. I would say the majority of the Physics Frontiers Centers, however, have an experimental component. If you talk about value added, you're now coupling—closely coupling—experiment with theory. In other words, you're having these people talk to each other. It's not the theorists go off and develop a number of really good-sounding and highfalutin ideas, and somewhere out there some experimentalist will pick it up and use it. On the other hand, you could make an argument that by bringing the experimentalists together with the theorists in this value-added relationship that I talked about, you're talking to each other all the time. If there is a new theoretical notion, it could be tested more quickly. Now, you're talking about moving things ahead and more rapidly—moving it ahead faster than it would move if I had to do a two-step process. I say this a lot for a lot of the multidisciplinary things that I could have a group here, and they develop something, and they publish it, and somebody picks up on it, and moves it on the next step. On the other hand, if you've got the two groups working together so that, immediately, I don't have to go through that public step—just that I don't have to go this way, I can just go this way, and my students can just go that way. You might get a situation where you have a theory postdoc working with a graduate student who's an engineer. This is the beauty of the center. It creates these connections; these interrelationships. That's the value-added part of it. When you do that, it becomes more of a team effort that really moves the science forward at a much more rapid space. Now, in this particular instance where there was condensed matter physics, I knew Jeff from his AMO work; not necessarily condensed. But it was the choice of the group to combine the more condensed matter approaches with the theoretical approaches. It was kind of the natural thing for them to do. We've got these people, we've all agreed we want to work together, and here's what we're going to do.

ZIERLER: NSF loves the proposal. They accept it. What happens next? How fast does the timeline move at that point?

CALDWELL: It takes about, I would say, probably about a couple of months. [laugh] I think, if I remember correctly, it was a Saturday morning. I was working in the office, and I called up Jeff at his office. I don't know why I thought he would be working on Saturday morning. [laugh] But, anyway, I called up because sometimes you have to sort out things. There's a submitted budget. Sometimes, you don't understand the budget. You have to ask questions about it. There are a number of steps you have to go through. I had some questions. I was working out all the numbers and everything. You also have so much money to spend on these things, and so I have to make sure that I maximize my portfolio. [laugh] I think I called one Saturday morning to ask Jeff a question, something to the effect that, "You're probably not here, but when you get a chance, I need to talk to you a little bit." I don't know whether he picked up the phone or what, but half an hour later, I get this call from Jeff [laugh] who I guess was working on Saturday morning. [laugh] There's a bit of a discussion of any sort of related questions that we have, and then centers are funded through what's known as a cooperative agreement. In other words, we don't just make an award. We make an award, we step back, and we don't do anything. But we actually issue what's known as a cooperative agreement, which means that we stay in much closer contact about what our expectations are and about what you're doing. Putting together that cooperative agreement can take a little bit of time. From the time we finished up until we are ready to go, —and then we don't make the awards. This has to go through our grants and agreements office, and they officially make the award to the university. But I think the first one, probably the last review was held in May—something like that—and the award went out the door probably in July. That is more or less typical, I would say. But it does take time because it takes time to negotiate the agreement, and then it takes DGA time--grants and agreements—just time to issue it. Two months is usually to be reckoned with, just for the paperwork, basically.

ZIERLER: What's the funding structure? What does that look like? Is it a lump sum for a year that's renewable? Is it a multi-term commitment? How does that look, initially?

CALDWELL: Normally, what it is, I said it's a cooperative agreement. The cooperative agreement is written for a number of years. In this case, it would've been written—at that time, we could fund no more than for five years. It would be in place for five years. We had a system whereby if they wanted to apply for a renewal—since we only have a competition every three years, if they wanted a renewal, they could get a supplement for a sixth-year while they competed for renewal. The normal cooperative agreement would've said something like—it is the NSF intention to fund this award for a total of five years, and a total of a certain amount of money. I've forgotten what the first one was for the IQIM. I think they were getting something like two and a half million a year maybe. John should have the number. He should have the records. But let's just say it's the NSF intention to issue this cooperative agreement for up to five years for a total of, say, $12 million over the five years. Now, you don't get all of the $12 million because what will happen is there is an agreed release of funds. When the CA is issued, there will be a certain amount predetermined for the first year that will be released. This is also part of what we negotiate for the cooperative agreement because we also say that then it is the expectation that, at the end of each year, there will be a review. Now, it might be a site visit review. It might be just an NSF-only site visit. It might be a site visit with an external panel. It might be just an annual report. But all of this goes into the cooperative agreement. At that time, I think, we were doing site visits about every year. You get the award. Let's say your award comes out in July. Along about next April or May, there will be a site visit. We would come out with a review panel, and we would review the progress, meet with the students. Contingent on that review, then we would release the next year of funds. The funds are released on a year-by-year basis up to the maximum based upon a successful review each year. This is all agreed upon in advance. This is the way normal cooperative agreements are done.

Funding the Matter in Quantum Information

ZIERLER: What happens then? IQIM is founded. It's funded. What are they allowed to do as a result of this funding? In other words, obviously, it's the salaries. It's bringing people in. But what does all of that mean in terms of what IQIM can now accomplish?

CALDWELL: Well, they better accomplish some good breakthrough science.

ZIERLER: [laugh] Exactly.

CALDWELL: That's the first priority. [laugh] In other words, when we go out there with that site visit team, that site visit team had better start getting really excited about the progress that's being made. But we also talk with the students. We meet with the students. Usually, they'll put up a poster session, and the students will go around and talk about their posters. All of that is part of the review. Then the review panel will write a review, a report, and they might be finding things that they didn't particularly like. But if they did, we could then sit down with the director and maybe the executive committee or however they're structured, and we would say, "Look, all of these things are going really well, but you got some problems here. This, you need to focus on. This, you need to start to think about as you go into the next phase." Typically, how well are the projects coming along? You've selected out of a broad spectrum of projects what you might work on. Why this one? What's so critical about this one, and what kind of progress are you making? Students, to what extent are these students demonstrating this excitement? Are they really moving ahead in their own preparation? All of these things are reviewed as part of their review, and basically what we expect is a demonstration of progress toward reaching the goals of the center as they express them.

ZIERLER: Now, when you say that there might be things that the NSF might not like, what would that look like? Obviously, that's not the case. It remains a phenomenal partnership between Caltech and NSF. But just to get a sense of, like, what are the things, what are the boxes that NSF is looking to check on these site visits?

CALDWELL: For example, let's say your review committee sees an area that you're investing in, and they're not persuaded that that particular area is that promising. Now, that doesn't mean that they can't continue it. But what it means is that your panel has said that they have a question about why have you chosen to invest in this particular direction? They might say, "Well, they have their own reasons." But all we're telling them is, "Your panel is telling you it doesn't look very promising." I think I had one center once where—it wasn't Caltech—but I think I had one [laugh] that the students were complaining about the increase in the health fees. [laugh] It has nothing to do with the center but I guess they figured maybe we could do something. It's usually things of that ilk, if they're real problems. For example, you might go—though it's never happened—but you might visit a center, and the students are not even talking to each other. Then you got a problem. Now, that's a big problem. Your center is failing in that particular instance. But, like I said, I've never encountered that. Most of the little things that we don't like are of the type that—it's not so much that it's wrong. It's just that we think you—based upon what the panel is telling us—we think you could do a little bit better. The whole idea of the site visit is not really to criticize, but also to help them. They have an advisory committee too, and the advisory committee is also there to help them get better. You can improve virtually anything on the edges. The only thing about the NSF, about our site visit, is that they report to us. We always tell the centers, "You need your advisory committee because they work for you. If I come out with a panel, they work for me. You should really talk to your external advisory committee. Make sure you have a good committee, and let them help you."

ZIERLER: Denise, beyond the year-to-year assessment, to step back and take the bigger picture, how does NSF measure progress in something like quantum information where there's really no agreement on when quantum computers will become viable, what exactly the role is between the fundamental research in quantum information, what the timeline might look like. As the NSF reassesses the relationship, how does it make sure that a center like IQIM is making progress year after year? What are those metrics?

CALDWELL: Well, it goes back to what I mentioned at the very beginning: that there's no requirement that you have a certain product. There's no specific requirement that you come out with something. However, what have you learned from the basic science point of view? You've been at this now for five years. You come up for renewal. You've been at this for five years. What have you done in five years? What do we know now that we didn't know five years ago, and how important is it? That's one of the things that we look at. Now, those are judged by perhaps review papers. But, many times, there will be things that are picked up by others outside the center and, moved on. It could be something that's picked up by industry, and they want to work with you because they're ready to. Doesn't John work with Microsoft? Is it Microsoft?

ZIERLER: Amazon.

CALDWELL: Amazon. There are things. You see it starting to have impact outside of this particular group making up the center. In the scientific community, others are jumping onto it. They're recognizing it. You can honestly say that over a five-year period, the science due to the work in the center has visibly progressed in a substantial way. Now, it's hard to quantify that. But if you look at the successful center, five years down the road, there's going to be a demonstrable impact of that five years of investment, and this is where it gets into the renewal. If you've made a demonstrable impact, and you're up for renewal, then we have more confidence that what you're next proposing will have that same impact five years down the road.

ZIERLER: Denise, as quantum information was really exploding as a science endeavor in the 2010s, and what had started in many ways at Caltech, there were new research projects, new research centers happening at Maryland, at UC Santa Barbara, at Waterloo. How did the proliferation of academic quantum information initiatives influence the Caltech-NSF partnership in quantum information?

CALDWELL: Again, we look at the individual. For us, we're happy to see these other things come along because it's telling us that this is an area we should be investing in. That's important. But just because there's one over here, it doesn't mean we're not going to do that one over there. At this particular level, we don't use that as a criterion for evaluation. In a way, we're happy to see the Maryland center come along. We're happy to see the CUA, the ones that we have, because they're sufficiently different. It's not like they're all doing the same thing. They're sufficiently different. The focus is different. But it tells us that we're investing in the right field because this is a field that's moving. The fact that we chose to do this, we like to think, means that we have a pretty good intuition of where the science is going. But it's not a situation whereby you say, "We will either do this one or that one, or we will look and see if that one is better than this one." But we evaluate the proposal for what they propose to do.

ZIERLER: Denise, I asked about the proliferation of quantum information in an academic setting. What about when major companies like Google, IBM, Honeywell, Amazon, when they start to get involved, and they're throwing massive amounts of money at this effort as well, does that change things for NSF?

CALDWELL: I would say that in the case of quantum particularly, I think it only gave us more impetus at NSF because quantum information science is still sufficiently in the early technological stages that there is a lot to be done that will take it over the hump into a technology of daily use, even if it's somewhat limited daily use. There's a long way to go. For us, we felt very good that NSF had made this investment, that we were called on by the NQI to continue this investment. These companies need us. It's great. I'm glad to see that they're investing. But they still need the basic science. Companies don't do basic science the way they once did. They're not the old Bell Labs. They just don't do it anymore. There are still so many basic science questions that need answers in order to move the field forward. They only stand to benefit. But the other place where they stand to benefit is we're going to train the workforce they need. Out of these centers are going to come some of the leaders that they will need to bring into their company to move it forward.

ZIERLER: Denise, it sounds like what you're articulating is that the federal government, because it's obviously not profit-motivated, it can exercise patience. It can articulate a long view where it's not about the quarterly profits next year or even five or ten years. It can see farther down the line than that. Ultimately, the companies need that long-term perspective.

CALDWELL: Yes. That's a good way of putting it, I think, and it's particularly true for an emerging technology.

ZIERLER: Now, one person, of course, the person that got us together, the connections that Spiros Michalakis is able to make, tell me about why he's so important in managing this wonderful NSF-Caltech relationship. What is he able to accomplish over these past 10, 12 years?

CALDWELL: Well, I think that, first of all, the IQIM has one of the most delightful outreach programs I know. Their outreach program is just fantastic. It's imaginative. It's creative. It really knows how to take advantage of all of the modern technologies.ZIERLER: Including Hollywood. It goes to Hollywood, and makes movies quantum.

CALDWELL: I know. Not everybody has that. But apart from that, anybody else could have—you need that spark of creativity. That's what I see a lot in Spiros. He has that spark of creativity to see what will be interesting. What will make an impact? When I talk to these people, what do I want to communicate? He also knows the material. This is another thing. He understands. He knows enough of the science. He may not be actually doing the research, but he knows enough of the science to be able to communicate it. That's important because it's not like he's just someone who is writing—I've dealt with some science writers, and it's really hard because they know nothing about the science. It's really hard to communicate a concept. But he's sufficiently versed in the science that he knows what he's trying to communicate from a scientific perspective, and that's important. He's very creative. If I remember correctly, he was the one—does IQIM still have this seminar that's organized by the postdocs, and the faculty are not invited? [laugh]

ZIERLER: [laugh] Where they can talk freely, right? [laugh]

CALDWELL: That's important. That's really, really important because it means it gives the postdocs—it goes back to this organic entity that I was talking about. It gives them investment. They're making the investment too.

ZIERLER: Denise, for the last part of our talk, to bring the conversation right up to the present, to answer that question that you already posed at the five-year mark, at the renewal, at the ten-year mark, of all of the things that IQIM has been able to accomplish as a result of this partnership with NSF, what stands out in your memory as, like, historically significant that when we look 50 years into the future, when we're writing the history about the quantum information revolution that's happening right now, what are the things that have happened over this past decade where you'd say, "This is really what made this possible"? What stands out in your memory?

CALDWELL: Well, it's really hard for me to know because I stepped out of management of the PFCs in 2012, when I became the Division Director. One of the things that sticks out in my mind, first of all, there's the overall success of the center. But John has been an incredible—I don't know what the word is. I hate the word "proselytizer" but—

ZIERLER: He's a visionary.

CALDWELL: He's a visionary in the field. John and I have been saying it every time we're both at the same meeting, I mean, we always wind up at the same meetings. I would say that John, in addition to just great science across the board, John has made a tremendous contribution, I think, in keeping this field—helping it gain the recognition that it has gained because he's recognized as being able to explain it, what it means, and to talk about it. For example, if you want to pin something that he's done, it would be the error correction. If I remember correctly, I think he chaired that KITP meeting--he was one of the organizers--on quantum error correction. Demonstrating that this can be done and it does make a difference, I believe, goes a long way.

ZIERLER: Denise, as other federal entities have moved in to support quantum information--for example, the Department of Energy is now a key supporter at Caltech and elsewhere—how has that changed things for NSF? Has it clarified NSF's unique role? Have there been partnerships within the federal government to strengthen support for quantum info across the board?

CALDWELL: Well, I think that it hasn't changed our emphasis and focus. The NSF, we have always had our—at least I say that I certainly have always had my vision of where I think this field should go, and where NSF's role is. I don't see that has changed, when we fund the people who are still making the difference. We fund our Physics Frontier Centers that are still making a difference. We've added our Quantum Leap Challenge Institutes, which are still making a difference. The DOE, of course, uses the labs a lot. We still are the primary point of contact with the academic community, and our PIs are members of the academic community, and that's how we have access to the students, and that's not going to change. To a certain extent, I think, the DOE coming in and making its own investments with its own particular emphasis, bringing in the labs, just expands what can be done. I talk with the DOE a lot. We have the Subcommittee on Quantum Information Science. We communicate. We talk about areas where we overlap, areas where they have their particular focus, and we have our particular focus, and we try to move ahead in a coordinated fashion, because if we coordinate, then we can expand the investment. The same is true for industry, by the way. They can also expand the investment. What's probably often not talked about very much is quantum information science has, for a long time, had a huge investment from the security side of the house.

ZIERLER: That's right.

CALDWELL: A lot.

ZIERLER: Quantum cryptography, it's a big deal.

CALDWELL: Yeah, I mean, for the longest time, and one forgets that. Of course, they don't talk about it [laugh], and so that's why nobody knows about it. The security side of the house has long been interested in this. As industry has become more interested, then you see the economic side of the house start to grow.

ZIERLER: Denise, two last questions, one specific to you, and then we'll end looking to the future more broadly. Given your role in making all of this possible, what are you most proud of?

CALDWELL: Well, first of all, I didn't solely make it all possible. I had a lot of help. I would say that what I am most proud of is that, more or less, the germ of an idea that the community had over 20 years ago, when we had the first workshop, has now become something that is very real. At that time, it was speculation. It was—

ZIERLER: It was a dream.

CALDWELL: --dreaming. It was dreaming. Now, it's real. Now, there are so-called quantum computers that actually compute things. We see sensors that are much more sensitive than was thought possible. I would say that seeing that happen has probably been—and knowing that, all along, I've been involved in it, and had a role to play, and made a contribution—or I like to think I've made a contribution—to me, that's really the most gratifying part of it because my job is not to do it, but my job is to try to help those people who want to do it, do it. By doing that, and seeing how far it's come, is probably the most rewarding part of it.

ZIERLER: Then, finally, last question, a theme in our conversations, the interplay of the fundamental research and how, ultimately, that relates to applications and the creation of scalable, commercially viable quantum computers. What are the things that you, NSF, the community in general are most excited about as both quantum information and quantum computation continue to merge and mature as a field?

CALDWELL: Well, I would say that what we're most excited about is that this is the beginning.

Now is the Beginning

ZIERLER: Wow, yeah.

CALDWELL: This is the beginning. When you're at the beginning, the possibilities become just wide open. What I'm most excited about is that we've got it this far, and now that it's this far, now we are really in a position to take those big next steps. When I see the future, that's what I see.

ZIERLER: It's almost like these past 20 years, it's almost prelude, is what you're saying?

CALDWELL: Yeah. [laugh]

ZIERLER: [laugh]

CALDWELL: Yeah, to a certain extent, and that takes us back to the role of basic science.

ZIERLER: That's right.

CALDWELL: It's a prelude to something, and it requires patience. You have to be patient. But I think we're at the beginning, and the next generation who comes along and takes it over, they're going to have a grand time of it.

ZIERLER: [laugh]

CALDWELL: [laugh]

ZIERLER: Well, Denise, on that note, it's been a great pleasure spending this time with you. I'm so grateful that we're able to capture your unique perspective on all of these things. I'd like to thank you so much.

CALDWELL: You're quite welcome. It's been my pleasure as well.