General Partner, Sequoia Capital
By David Zierler, Director of the Caltech Heritage Project
September 23, 2022
DAVID ZIERLER: This is David Zierler, director of the Caltech Heritage Project. It's Friday, September 23, 2022. I'm delighted to be here with Dr. Shaun Maguire. Shaun, it's great to be with you. Thank you so much for joining me today.
SHAUN MAGUIRE: I'm beyond thrilled to be here. Caltech means a lot to me. It's my Hogwarts. It's going to be fun.
ZIERLER: Awesome. Shaun to start, will you please tell me your title and institutional affiliation?
MAGUIRE: I never say this, but I guess I'm a doctor. I received my PhD. Do you mean my job title? Or my Caltech title?
ZIERLER: Yeah, your job title, where you work, and what your title is there.
MAGUIRE: My job title is I'm a general partner at Sequoia Capital. Sequoia is a 50 year old venture capital firm based in Menlo Park—one of the preeminent venture capital firms in the world that backed, in their early days, Apple, Atari, Cisco, Oracle, companies like that. In more recent memory, companies like Stripe, Zoom, Instagram, YouTube, ByteDance in China which created TikTok and many, many other companies across consumer enterprise, hardware, and all these things.
ZIERLER: Shaun, do you have a sense of the origin story of Sequoia—what niche it was looking to fill when it started?
MAGUIRE: Yes, I do. I am, not quite as much as you, but I'm also a student of history, and I've been a student of Valley history. Basically, venture capital has become this huge industry, but back in the day when Don started the firm in 1972, it didn't really exist. It was basically a bunch of renegade people that just loved technology and didn't want to go work on Wall Street; they wanted to make their career out of trying to build the future the hard way. Don Valentine, the founder of the firm, he had been at some of the top semiconductor companies of the past, including Fairchild and National. Don had mainly been in sales and marketing. He was an incredibly brilliant man and had really good technical instincts, but he was really from the sales and marketing side. I would say a bunch of the other prominent, early investors were really more coming from the technical side. People don't quite give credit, but Caltech's own Arnold Beckman in many ways was maybe the first VC. It's not talked about that way, but Shockley Semiconductor was originally a division of Beckman Instruments. The model was evolving, basically from the time that—Shockley was the first semiconductor company in the Valley, and there wasn't a venture capital model yet, so Shockley was basically a division of Beckman Instruments, a wholly owned division with a bunch of incentives. It evolved over the next ten years with people like Arthur Rock with Intel and others, and it, around the mid-70s, stabilized in the model that we have today.
ZIERLER: Shaun, we'll get to this in real time, but did you always have a business streak, an entrepreneurial streak that you always wanted to actualize? Or was this really a sudden career shift from what you might do otherwise with a PhD in fundamental physics?
MAGUIRE: I think something that's hard for people to understand about me is that I've always been doing multiple things in parallel my whole life. There's not one moment in my life where I wasn't doing three or four things, all at a relatively high level completely in parallel. It's just how I am wired. So, I've been absolutely fascinated by business since I was a little kid. I've also been absolutely fascinated by science. I've also been fascinated by computers, which I would say is slightly different than science. In business, my two passions were—I would say there were three. One was the stock market. I was fascinated by what made these numbers go up and down. Another was industrials. I was absolutely fascinated by where things come from, how energy works, oil and gas, chemicals industry, things like that, pharma. The third was Silicon Valley, where people are making this technology using physics and other things to bring it forward. I always had that passion, but I've had the science passion which really started with astronomy. Astronomy becomes interest in black holes, which leads to people like John Preskill, you know, legends of the field. With computers, it just seemed like the most important technology of the time. I was born in 1985, so I was always trying to mess around with computers. The fourth area is I'm super hyperactive. I have incredible energy, so I've always been doing athletics of some kind, because otherwise I just can't think unless I burn my energy.
ZIERLER: Shaun, in your day to day role at Sequoia, are you bringing your quantum information expertise? Is that relevant to the kinds of things you do on a day to day basis?
MAGUIRE: Very rarely. These things go in waves. Before Sequoia I was at Google Ventures. I led the Series A in IonQ, which is one of these first wave quantum hardware companies, which a bunch of people at Caltech know—like Chris Monroe and Jungsang Kim, the founders of the company—well. So, when I was doing that one, yes, sure, having the quantum background was really important, and being able to do diligence on the company and trying to figure out what the roadmap would be and what the biggest bottlenecks would be for scaling and things like that. But as an investor, I wasn't doing any calculations. It was more helpful for being able to do diligence. My job as a VC is much more about business strategy, hiring people, managing people, understanding human psychology, understanding market psychology, understanding where the puck is going in terms of technological trends, things like that. Physics is very powerful. I use physics a lot. I've invested in a lot of companies. Quantum information not too much. Physics is something I use all the time, because I've invested in a lot of companies that touch atoms. You need physics to understand that. The third thing to say is physics teaches you a way to think. At my job I'm dealing with incredible rate of change, I'm dealing with incredible amounts of data, and physics gives you frameworks to make sense of all this and try to come up with heuristic laws and ways to think about things which are very powerful for investing.
ZIERLER: Shaun, a question I've been excited to ask you since I first reached out: with your area of expertise, as a student of history, I wonder if you've ever thought about some of the parallels between, for example, a Bell Labs in the 60s, 50s, 70s, the middle part of the 20th century—the industrial support for fundamental research and how you might compare that with what Google, Microsoft, Amazon, and Honeywell are doing with regard to quantum information today. Where do you see some of the parallels? Where do you see some really new directions?
MAGUIRE: That's a great question. It is something I've thought a lot about. I think Bell Labs, one of the key things, they basically had a regulated monopoly. You're not supposed to say that these days, but it was important, because when you have that incredible amount of predictable free cash flow, it makes it really easy to go pump tons of money into the R&D. I also think we were living through a pretty incredible period in semiconductor technology. I think Bell Labs won nine Nobel Prizes, and there was a lot of stuff that was pretty adjacent to Bell Labs. It wouldn't be relevant to the business model in a parent company. It wouldn't have been relevant in a five year time frame, but relevant in a fifteen year time frame. So, they were able to do really amazing work that was not too far away from the core business. Out of the three you mentioned, I think Google is the only one that has a lot of parallels. It's not a regulated monopoly, but they have—not supposed to say this—but they have a monopoly on search. It's actually breaking in some ways right now via Apple. Google will do better than say, Meta or Facebook, but Apple has changed the way ad search works recently and made it a lot harder for their competitors, and they're getting a lot of ad market share, so it'll be interesting to see what happens to Google in that context. But, just by having this incredible profit engine that you can pump into these other ideas, I think there are a lot of similarities. One thing: I think a lot of the things they were investing in were not related to their core business. They were investing in flying cars—Kittyhawk—just two days ago said that they're shutting down. They're investing in a lot of things that were not—biology, life sciences—I actually think it might be bad to invest in things that are so far away from what you're doing since you don't have the core expertise. I actually think that with Google, they've lost a lot of the goodwill internally. They've lost a lot of the goodwill of public markets. Quantum computing and AI are existential, and I think a lot of internet research efforts of Google, those are both Bell Lab equivalents. Those are things that Google should be investing like crazy into, because those are existential risks to their core business on a 20 year time frame. Honeywell I don't think is a great comp; they don't have the same profit engine that Bell Labs has. They've always been more of an R&D firm and government contractor. One thing that I think is close to Bell Labs from a different direction right now is Deep Mind. Deep Mind is now owned by Google, so I think that is a good one. Deep Mind has basically been going across all of science and trying to apply machine learning in science, so it's a much closer thing to the core business model. That's another area where Google has done an incredible job, is machine learning research. It's close enough to the core business that it's a very smart strategic thing to invest in. I know that's a long answer. I'll pause.
ZIERLER: Relatedly, I wonder where you see all of this investment in quantum information within the broader context of venture capital. In other words, if all of these companies are pouring billions of dollars into quantum computing without anyone really having a truly well-developed sense of what this technology will be used for, is that common? Is this like a common narrative in venture capital? Or is it something entirely—
MAGUIRE: It's super common. It's incredibly common in the history of technology. Just think of a few examples. Look at solar. Solar starting in the early 2000s—2003 to like 2012 got incredible attention both from VCs but also from government subsidies. The vast majority of the individual solar companies failed, but the whole category has been incredibly successful. Now, power cost per watt generated by solar is roughly one-tenth of fossil fuels at this point, depending on where you are. That's not the fully burdened, fully level-ized cost, but just the instantaneous production cost. And that all comes from a huge amount of money that got poured into a basket of approaches, and those things were all able to compete and evolve. Very few VCs made any money on solar. They lost a lot of money, but the category has been very successful.
Another example is fiber-optic communication, where in the late 90s, early 2000s, there was an incredible amount of venture capital money and government subsidies that went into building fiber infrastructure. One of the most high-profile ones was Global Crossing, which was this company that was the fastest company ever at the time to get a billion dollar valuation. It led to this overbuilding phase with fiber, which was actually a prerequisite to have all the internet businesses of the next decade built on top of it. I could go on and on. I think the key lesson here is that there can be certain industries where almost all of the VCs lose almost all their money on the investments because there's too much competition and the science is moving too fast, but that actually is an important part of getting the future to arrive faster. I personally believe quantum computing is going to be similar to solar. There are probably a thousand solar companies started in that ten year window and at most two or three of them that are meaningful today. I think it might be a similar thing with quantum computing. As someone who loves quantum information, I'll be thrilled with that. I think it's a good thing.
ZIERLER: Does the comparison hold up insofar as with solar startups, we knew what solar would be good for, right? It gets us off fossil fuels. It's a stable energy source. Can the same be said at this point for what quantum information, what quantum computing will be good for? Or is it still premature?
MAGUIRE: I would say they're very similar, and with solar, it wasn't as clear. It wasn't as clear actually, if you go back 15, 20 years, that solar would be able to get to the price levels it's at today. It wasn't as clear that you'd be able to go to cheaper instantaneous power production than natural gas, for example. Because it's an extra three factors of two you had to get. Moore's law had to keep running for an extra five years, and no one knew how long it would run for. I was paying pretty close attention back then. It wasn't as obvious, but it was obvious there would be certain niche applications of solar. It was a good investment for governments. We had to have a basket of renewables to fight this thing that was starting to happen with global warming. With quantum computing, I would say there's already a lot of applications that are pretty clear, and then there's also a whole bunch of things that maybe you can't say the precise algorithm, but on the other hand it's pretty obvious quantum computers will be important. Physicists say all the time, "Simulating physical systems: quantum computers are clearly going to be important for that." It's a tautology, but it's also 100% correct. It's obvious that for things like material science, when quantum computers are powerful enough, they will play an important role in material discovery. I would say it just doesn't matter. We don't need to know the exact algorithms that are going to run. It doesn't matter. The hardware is going to be really valuable.
ZIERLER: From your own perspective, do you tend to think of this in somewhat of a horse race metaphor? Do you see Google, Microsoft, IBM, and all these companies—are they racing toward a singular finish line? Or are they doing something different?
MAGUIRE: Yeah, I don't think they're racing toward a singular finish line. A few things: one, I think there are many finish lines; two, I think the future is non-deterministic. I'm not talking about on a physics level. I mean for all intents and purposes, even if it's deterministic, it's such a complex system no one can predict, and I don't think it's yet set on—the fact that humans used rockets instead of some alternative technology to get to space is in part a function on when World War II happened and when the Cold War happened. If we wouldn't have had a world war until 200 years later, we would have gone down a very different technical pathway, and then maybe would have been using something other than rockets to get to space. We would have gotten there later with the different technology. Some of these things are so dependent on so many other variables. First of all, I don't think it's racing toward the same goal, but even if it was, I don't think anyone knows what that goal is, and I don't even think it's set. I think it depends on a lot of things that play out. It's path-dependent over what happens over the next ten, twenty years.
ZIERLER: You mentioned on a day to day basis you utilize your expertise in quantum information fairly rarely. Institutionally, is Sequoia involved in the quantum information space at all?
MAGUIRE: Yeah. Sequoia invested in Rob Schoelkopf's company, QCI, before I joined. Rob is another legend of the field.
ZIERLER: Yup. Did you know Rob? Was that a connecting point to Sequoia?
MAGUIRE: No. I do know Rob. I did know Rob, but it was not a connecting point for me. It was a happy accident.
ZIERLER: [laughs] Shaun, let's establish now some context. Where were you for your undergrad?
MAGUIRE: My academic background is pretty unusual. To answer the question: bachelor's degree from USC, University of Southern California, but there's a lot more to the story. Basically, starting in eighth grade, I got really disillusioned with school. I went to public school in Orange County, California. I had a really horrible experience, to be honest. I kind of stopped going to school. I missed more than the legal number of days in the state of California due to three or four factors, so I was just kind of sat on my computer and doing my own things. Then, sometime in tenth grade decided I just had to leave school, so I took this thing called the California High School Proficiency Exam which is a GED equivalency, and left, and went to community college for two years while my friends were finishing high school. I transferred to USC, and I was only there for two years. It was entertaining.
ZIERLER: Then what happened?
MAGUIRE: I was really into computers as a kid, and really passionate about physics. I didn't really have much of a formal background in it or anything. I wasn't really going to school, so I wasn't doing math competitions or anything. When I got to USC, I randomly was walking—I studied math. I was randomly walking through the halls of the math department, and there weren't many math undergrads there. I'm probably making this up, but it felt like 20 kids. It was a really small major for a school that big. I was walking through the halls, and there was a professor who had a math competition sheet on his door—this guy, Richard Arratia—the competition was the Putnam competition, which is the North American college math competition. I didn't even know about math competitions. I had literally never done one. I had never seen one of these. In high school, I didn't know about the IMO, USAMO, AIME, or any of these things. I saw these 12 questions and sat down outside his office and started thinking through how to solve these. I had a pretty good intuition about how to solve them and ended up talking to Professor Arratia. He started mentoring me. He had, every Wednesday morning, a group of literally three kids that were interested in these things, so we would go and he would teach us some mathematical problem solving. I had some aptitude. The professors at USC told me I should graduate and go somewhere better, so I went up to Stanford and started grad school there, actually in the statistics department. My passion, especially coming from that background, was in probability and combinatorics, but really theoretical probability I just found absolutely fascinating. So, I went up to the statistics department at Stanford, which is one of the top places in that, and at Stanford is where I fell back in love with physics.
ZIERLER: Did you think about quantum information at all at Stanford? Did you have any interface with that world?
MAGUIRE: No. My physics passion was in ninth grade. I became really interested in the solar system. Sorry, not in ninth grade. When I was nine years old, I became really passionate about the solar system. I had this unbelievably lucky thing: one of my friends' dad was a local community college professor. He taught computer science and astronomy. His name is Doug Borcoman [?]. He was an amateur astronomer, and sometimes with my friend Brandon, he had like an eight inch telescope, and we'd go look at stuff in the sky. When I was in sixth grade, NASA had this program called SAREX, Satellite Amateur Radio Experiment. There are some videos of this online; it's pretty hilarious. Basically, NASA was doing this program—you could learn ham radio and ask astronauts aboard one of the shuttles, ask them a question in ham radio as they orbit the Earth. So, I did this and got to ask a question to the astronauts, and that honestly made space really tangible to me. It gave me intuition for the distances and the speeds. That kicked off a whole new passion in space, and that led to learning about black holes and getting absolutely fascinated by black holes. Before black holes, a prerequisite to understand them is you have to know some general relativity. A prerequisite to that is special relativity. I literally couldn't sleep as a kid, I would just think about special relativity. I would do these thought experiments. There was one—I think it was shortly after I was 13—for about six months I couldn't sleep at night. I was thinking about if you had three space ships that were traveling in a line, so spaceship A, B, and C. If the two ends were traveling away from the one in the center, each at the speed of light—so A is traveling away from B at the speed of light, and B from C at the speed of light—how the hell could A and C not be traveling away from each other at more than the speed of light? I spent six months really trying to understand that, and I couldn't understand it. It took me years after to really understand it. It's easy to understand the calculations, but it gave me this really deep—the answer here is that space is not flat, and your intuition for flat geometry is completely wrong. You can't have spaceships traveling away in a straight line from a Euclidean geometry perspective. This all happened in a curved space, a Minkowski space. It gave me a really deep intuition for that, and that led to a passion for black holes, and I came back to it later.
ZIERLER: What's the connecting point from Stanford to Caltech? When did that happen?
MAGUIRE: I was at Stanford for a year and a half. Candidly, with my background of 1.8 GPA in high school and an F in algebra 2, beggars can't be choosers. Some of the USC faculty had good connections at Stanford and basically got me in on letters of recommendation. But once I got there, I got over a 4.0 or whatever in my first year. I got A pluses in a lot of my classes. It was still pretty easy for me. I took a lot of tough graduate math classes. I was probably taking eight classes a quarter. I was really doing a lot. So, at that point, I wasn't just a beggar anymore. I won an NSF Graduate Fellowship and NDSEG Graduate Fellowship, and I kind of in my head had this realization that I was only at Stanford doing probability because that's the thing that I got recommendations for. What I was actually most interested was space. At Caltech there was this guy, Jerry Marsden—Jerrold Marsden—who is an absolute legend in space physics. He had founded this tiny little department called Control and Dynamical Systems. Jerry was this weird physicist where he had a really elite pedigree, and a lot of people with really elite pedigrees would go into the really cutting edge stuff of quantum gravity, or string theory, or high energy physics in general. Jerry was one of these rare people that decided, I'm going to go back to the fundamentals, go back to classical mechanics, and try to understand that really, really well and figure out important things there. Jerry had just done incredible work in understanding our solar system, orbits, trajectories for space crafts, and things like that. That was my passion, so I went to Caltech to work with Jerry. It was a crazy thing, but Jerry, in my first year, had a medical complication and died during my first year. I didn't know exactly what to do. I had this strong background in probability, so I went into the math department and started working with someone—Nikolai Makarov, who's a legend in math—to do some theoretical probability work. Then I got recruited to work at DARPA by Regina Dugan. Regina was a Caltech alum who was the Director of DARPA. She recruited me onto a pretty crazy project related to the war in Afghanistan. I had to say yes to it on the spot, so I went to DARPA for a year and a half full-time there. Then it led to starting a company, which did DARPA work for the next seven or eight years. It was when I was at DARPA, that's when I got exposed to quantum information. It tied together all of my passions, just all of them—black holes, computers, all of these things. I literally emailed John Preskill from Afghanistan. This is a true story. I emailed him from Afghanistan and said, "I'm coming back to Caltech. I have a fellowship, so I don't need any funding. I've been reading your notes from Afghanistan." I sent him a picture. He didn't take me as a student, but he told me to come to his group meetings, so I did. I was lucky enough to work with him. I love John.
ZIERLER: Did you officially unenroll from Stanford at that point? Could you have gone back?
MAGUIRE: I had officially unenrolled from Stanford a long time ago. I left Stanford with a master's degree and went to Caltech. I started at Stanford in 2007 and moved to Caltech in 2009.
ZIERLER: When did you first appreciate the connections between black holes and quantum information? When did that start for you?
MAGUIRE: It's what Stephen Hawking is famous for, but I didn't understand at all the stuff Hawking had done. I just kind of knew this thing, the information paradox, and all of that. When I had thought about it—I'm going to tell you, this is the 100% truthful version.
MAGUIRE: When I was a Stanford and when I first joined Caltech, because I had such a weird background, I didn't have the background yet to actually be able to think about the problem or really understand the problem statement. I viewed that field, the stuff that John was working on, as the absolute top of physics, and I didn't think I had the background yet to be in that world. I didn't even know the prerequisites to be in that world, so it took an extra few years. By the time I came back to Caltech in 2012 after all this, I had been able to self-study and knew, for example, general relativity on a basic level. I originally self-studied quantum mechanics, and I was able to have some intuition. I honestly didn't feel like I deserved to be in that world, and I didn't know enough to even know how to get started until I was coming back. Now at this point I'm maybe a 25 year old or something, I think was when I was coming back to Caltech.
ZIERLER: In 2009, just to clarify—
MAGUIRE: 26 actually. I came back in 2012.
ZIERLER: Just to clarify, when you came to Caltech, you were already admitted, but it was not certain at that point that you'd be John Preskill's student? That sort of developed over time?
MAGUIRE: Correct. When I came to Caltech, I was going to work with Jerry Marsden. I originally joined in the Control and Dynamical Systems Department. It was a tiny department. I was the only student in my year that joined that department, the only grad student. It was a tiny department.
ZIERLER: Once you started going to group meetings at IQI, what were your impressions? What were people excited about?
MAGUIRE: That was another thing, is that I am a—there's this joke at Caltech (MIT does this too): How do you tell the difference between introverts and extroverts at Caltech? The extroverts are the ones who look at your shoes when you're talking.
MAGUIRE: By Caltech's standards, I'm an extreme extrovert. By normal human standards, I'm an introvert. The other groups I had been in, they weren't groups. Working with Professor Makarov, I was his only student. Every week or two, I'd go talk to him, but there was no one else at Caltech I could talk to about the work. It was unbelievably lonely. It was really lonely and solitary. When I first went to John's group, it was like 20 people in the meeting, once a week getting together, people having lunch together during the day sharing ideas, people working on many different topics, working on the future of computing, algorithms for that, hardware for that, working on black holes, working on fundamental quantum mechanics, paradoxes in quantum mechanics, things like this, condensed matter physics. There was this incredible energy and camaraderie there, and it was addicting, especially for me coming from—I had only been exposed to solitary research before that.
ZIERLER: So, this was a real vibrant social scene separate from the science? [few minutes pause] When you got to the group meetings with John, what were some of the big debates that were happening? What were people excited about at that point?
MAGUIRE: I joined the group in 2012. The field has moved so fast. At the time, a recurring theme through the group is that Kitaev had done a lot of really interesting work and people were trying to understand it continuously.
ZIERLER: Was Alexei accessible? Could you find him? Could you talk to him?
MAGUIRE: I love Alexei and couldn't think more highly of him. Alexei is a mathematician. I think that for a lot of people that come from a pure physics background, it's hard for them to talk to Alexei because he really is talking as a mathematician. Because I had come from a math background, I found that yes, I could talk to Alexei. He was always accessible. Alexei traveled sometimes, and I think he was very protective of his time in that he wanted you to meet him when he would say, but he would always make time for you. So, if you say, "Hey Alexei, there's something that I would really like to understand that you worked on. Can you make time for me?" He would say, "Be here at this time and place." He would always offer that. I think a lot of people were always too afraid to even ask him. Alexei is really introverted. Alexei is not going to just go hang out in the hallway at the blackboard doing his work in a public space, inviting people to come up and start talking to him. You need to grab him when he's around and set up a time, but he'll always do that. I think there's a second thing. I think some of the physicists didn't quite understand the math language that he was using, but Alexei is a path breaker. Maybe five years later the physicists will go learn the math required to talk to him. That was one thing, Alexei. That was one theme. Another is that Jeongwan Haah had just done this three dimensional error correcting code work, or was just finishing. Jeongwan was just finishing his PhD at that time, and that was a really exciting result that a lot of people were very interested in. I can't remember the exact other things in the very beginning when I joined the group, but I can tell you the themes over the whole ten years or whatever.
ZIERLER: Coming in in 2012, did you recognize the transition from IQI to IQIM? In other words, the experimentalists joining matter to the theory, did that register with you at all?
MAGUIRE: I think I was a little unusual in that I was pretty social. I had a lot of friends, so I was already hanging out with a lot of the matter people, like Oskar Painter's students, who obviously has been a big part of IQIM. He had a couple PhD students that were about my age or a little older who became good friends. There's this guy Amir Safavi Naeini who's a professor at Stanford now, and Alex Krause, and Simon Gröblacher who's a professor at TU Delft. These guys all became my good friends. At Caltech, everyone talks about the science all the time. I think that on the grad student level, the evolution from IQI to IQIM wasn't that big of a deal. I think maybe on the faculty level it was a bigger deal, because it changed who was on the committees. So, now, your default is sitting next to experimentalists as a theorist at events and on committees and all that. I think maybe on the postdoc level it had an impact because we started to have a lot more seminars and all of that, that would have people from both experimental and theory world. But as a grad student, especially a social one, you already knew a lot of those people.
ZIERLER: What was the process from John inviting you to the meetings to actually becoming his student? How did that play out?
MAGUIRE: My read is John is just testing your commitment. After the fact, I would say my post hoc analysis is that almost anyone that shows up for three to six month, you kind of default become his student. And what happens, the wave function collapse moment is when you need an advisor to sign something—there are certain things at Caltech where you need an advisor's signature, so the first time that happens, when you've been going to his group meetings for a few months, you kind of go to him and say, "So, I need this signature. Will you be my advisor? Will you sign this thing for me?" At that moment, he becomes your advisor. But I think he's testing people's commitment, which I think is a really smart strategy that not enough people do. It's really easy to go say, "Admit me; I'm going to work really hard." It really is easy to say that, but doing three to six months of repeatedly showing up repeatedly and having something to say every week because you learned something new, that's what actually matters and that's actually hard to do. So, I think John has a smart system.
ZIERLER: Shaun, coming in in so many ways as an autodidact between the math and the physics, what areas did you have to play catch-up for quantum information, and where did you jump right in alongside your cohort?
MAGUIRE: I'm always playing catch-up. There's always more to learn, so I'm always playing catch-up. I didn't know anything about quantum information. I still don't—no one—I don't know anything about quantum mechanics. Definitely had to learn all that. Eventually with where the work went, I didn't know anything at all about quantum field theory or string theory, which I had to learn over the next few years. The only area where I actually knew something was probability, which was an area that I had spent five years or whatever, so that was an area where I knew something. I don't know, I was learning the rest.
ZIERLER: As you were surveying all of these ideas, where did you see a niche? Where did you see an opportunity to work on your own project, do original research?
MAGUIRE: I think someone doing theoretical work in what I call "hard" fields—a PhD student doing theoretical work in pure math, or in quantum gravity, or high energy physics, or whatever, those are really hard areas to do original work. There's so much prerequisite knowledge, it takes so long to get to the point where you can actually make a contribution. It took a long time. It was basically learning, reading papers, talking to lots of people, going to group meetings for a long time. It took a long time to get to that point. I can be a little more concreted if it's helpful, but I'd just say in this field, in quantum gravity, it's really hard to do an original contribution without three to five years of having learned the foundations. Whereas there's some areas, like in combinatorics, where you can do—or like today in machine learning—you can do original work in three months. We're so nascent in those fields that if you're just really smart, IQ will get you far, and in three months you can do some original work. In some of these other fields, it takes years.
ZIERLER: As you got comfortable in the field, where did you see an opportunity to contribute? What was some of that original work for you?
MAGUIRE: It was simply in having a stronger math background than some people. I had been interested in this field called hyperbolic geometry. I have always, in science, I'm attracted to people that have been out of the box. Those are my heroes, my role models, the people that have done things very differently than other people. Just to give you some examples of people from different domains, in mathematics there's this guy, Bill Thurston, who pioneered hyperbolic geometry. He was also an interesting, out of the box human, so I found him really exciting. Feynman is the classic Caltech person. Another example would be: in quantum there's a guy, Yakir Aharonov, as in the Aharonov-Bohm effect. He's a professor I believe now at Chapman University, spent the early part of his career in Israel. Aharonov is one of these guys that's always doing things very differently than other people. I think that, to put John in that category, one of the things I always really admired about John is he had changed fields many times and risen to the top of many different fields, like, started off in really high-energy physics, dark matter work, hardcore high-energy physics, and then he moved to Stephen Hawking style quantum aspects of black holes I would say was the second major area. Then the third major area was quantum information in the 90s, and has now been the glue that has tied many things together. I admire John as someone who's fearless enough to go be at the top of one thing and then jump and do another field where they're a relative novice. So, I've always been attracted to people like that. Bill Thurston was this guy who's work—I had just been fascinated by the guy, and I read a lot of his papers. So, I tried to bring some of the hyperbolic geometry ideas into this field.
ZIERLER: On a technical level, I wonder if you can explain, what was the relevance in this field?
MAGUIRE: I will do my best to explain the arc here. The arc was that Hawking and others had come up with this information paradox that was basically saying that the general relativity and quantum mechanics make different predictions about the end-state of a black hole. In some very crude sense, one says that information is conserved, the other says that information is destroyed. That then led to a lot of evolutions over time. One of the big evolutions in the early 90s was this thing called the holographic principle. Lenny Susskind, but actually really, based on a lot of John's ideas as well. John was a huge part in this holographic principle idea. He said that maybe nature has this weird property that sometimes you know the physics of what's happening in some region of space, maybe all you need to know is what's happening on the boundary of that space. That's kind of the core intuition of behind the holographic principle. That happened in the early 90s. In the late 90s, Juan Maldacena had a big breakthrough there. He showed that in a specific sub-version of string theory, that that holographic principle would hold exactly true, and this result, I think it was in 1998, but in the late 90s became called what's called AdS/CFT, anti-de Sitter space, which is on the general relativity geometry side, and CFT, conformal field theory, which is on the high-energy physics quantum field theory side of the equation. That happened, and then in 2015 there was this thing called the firewall paradox. That was the next huge jump in this area. I would almost say in a lot of ways it was similar to Maxwell's demon paradox, which was in the late 1800s. Maxwell's demon was first in the statistical mechanics domain or thermodynamics domain, but it was what first brought the concept of information to physics in a tangible way. It's the first time that information had to be considered in physics. This firewall paradox really sharply showed that quantum information will play a fundamental role in resolving, in terms of understanding the nuance between general relativity and quantum mechanics, just in a really sharp way. That was a very exciting time, so a lot of people both in quantum information and also in high-energy physics, people all came from those two extremes and all came to the same problem. So, I jumped on that bandwagon, joined that group of people. That paradox really came from AdS/CFT line of thinking, and for these things like anti-de Sitter space, very intuitively, very naively—to back up one step, you can think of objects in space, for example, any two dimensional surface: it's either got negative curvature, zero curvature, or positive curvature. That's a global statement about the object for any surface or three dimensional manifold, etc. There's some technical definition that gives you that, but you can have negative, zero, or positive curvature.
A lot of people, their intuition for space or geometry is that we live in flat space, but if you live on a sphere, that sphere is what we would call positively curved. I could explain the technical definition, but that's neither here nor there. A saddle is what we call negatively curved. This anti-de Sitter space, it's like living in a space-time where you're stacking a bunch of negatively curved manifolds on top of each other. That is basically hyperbolic geometry. AdS, anti-de Sitter space stuff is basically just doing hyperbolic geometry, and there's a bunch of really concrete ways to make that precise. Mathematicians have studied hyperbolic geometry to death and have learned incredibly beautiful things. One of the things they've learned—there's this famous essay from a guy named Mark Kac where he asked, "Can you hear the shape of a drum?" That was the question, and what he meant by that was if you could take boundary measurements around the sounds you'd be hearing on a drum, or the heights of waves moving through a drum, could you uniquely figure out the shape inside? A spherical drum makes a different sound than an oval drum, which makes a different sound than a square drum. So, that was the question. It turns out the answer is no. There are certain shapes that have different—they have the same eigenvalues, same to the Laplacian, with different geometries. There's a very similar result in hyperbolic geometry which basically says the eigenvalues that correspond to waves moving in negatively curved space follow these very specific rules, and there's some really beautiful aspect, and there's actually a relationship between those things—between those waves and the eigenvalues that come from the waves, and the geometry in these geometric data. My PhD was basically making a bunch of connections between these ideas.
ZIERLER: What kind of role did John play in all of these decisions? Was he a hands-on advisor? Did you talk to him a lot about these things? Was it related to what he was doing at the time?
MAGUIRE: First of all, I could not love John more, could not be more grateful to John, could not think more highly of John. It's an interesting thing, because I think John changes many people's lives. I don't think it's an accident that John's group has been the central node in quantum information over the last 20 years or so. It's funny because John has a very—it's maybe not the style you would think. The way John works, is it's really a Socratic style. It's almost a minimalist style. I would say there are two parts to it. John asks incredible questions. There's two things. One is people respect John so much that you don't want to disappoint John. You want him to respect you; you don't want to disappoint him. It's like, some parents rule out of fear; some parents rule out of love. John rules out of love, and you don't want to disappoint him. You want to live up to his name and reputation. So, John never tells you you're wrong. He never tells you you're stupid. He never tries to make you feel stupid. He just asks really good questions that lead you on a journey. That's the way I would describe it. Over the course of three years, maybe once every two to three weeks he'll ask you a question that is almost like the series of questions is taking you on a journey that he wants you to go on, but he doesn't tell you explicitly what journey you're going on ahead of time. He knows where you're going. You don't know where you're going. He doesn't tell you where you're going. He just gives you breadcrumbs along the way when you need them. It's a very interesting style.
ZIERLER: Shaun, to zoom out from your specific research, what were people talking about with regard to quantum gravity during this time? What was seen as the holy grail? What could quantum gravity actually achieve?
MAGUIRE: Others know this stuff better than I do, but last I checked, there are only two places in nature that we're aware of where quantum mechanics and general relativity make different predictions about what should happen. One is what happens with the end-state of black holes. Another is how the universe was formed, like the big bang. The big bang one is—somehow people don't really talk about that. People don't really use that as an example. I think it's because it's just in some ways it's unknowable. It's too far outside of our tools right now, and we really don't know what direction to go. But it's only those two places where we know that quantum mechanics and general relativity make different predictions. The goal of quantum gravity is to reconcile these discrepancies. In some ways there's a parallel to the past. For example, the thing that motivated quantum mechanics, I think there were three main categories of discrepancies. One of the most famous ones was the photoelectric effect that Einstein won the Nobel Prize for his explanation of. When you're looking at light, there are certain ways where light very clearly behaves as a wave, and there are certain ways where it very clearly behaves as a particle. In all the classical physics, optics, Newtonian mechanics, etc., and classical electromagnetism, that didn't make any sense. So, that was one example of something. Another was the way black body radiation happens. That's another example of something where it didn't make sense with the classical treatment. Trying to understand these discrepancies led to quantum mechanics. It may seem like a really small thing, but it's the only clue we have that there needs to be something new. So, we need some better version of physics that can interpolate between quantum mechanics and general relativity and be consistent with these two things, these two points that don't fit the data. We call that quantum gravity. We still don't know much about quantum gravity but we're making some progress. There are a lot of candidate theories that fit under that umbrella. String theory is one. Even within string theory, there are many different branches and ideas. There have been these big evolutions, these big jumping points, and I only mentioned some of the ones related to the information paradox. There have been a lot of other big breakthroughs related to string theory over the last, like, 40 years. But in 2015, this firewall paradox was a huge jump, because it created a bridge for the quantum information people to talk very precisely to the high-energy physics people. Before, there was too much incompatibility in the languages these fields would use, so it was just hard to even communicate. So, that became the most exciting thing by far in quantum gravity, and now the field is on a journey to unite the fields even more closely. There has already been a lot of great results there, and I'm sure there are more to come. Some of the big ideas—one that John was involved in was bringing in the ideas of error correcting codes, that nature might be behaving like a quantum error correcting code on some really fundamental level. Another is this idea that people have called ER = EPR—Einstein-Rosen equals Einstein-Podolsky-Rosen. It's basically this idea that somehow wormholes and entanglement—so wormholes on the general relativity side, entanglement on the quantum side—are very deeply related to each other. There's this other thing called holographic entanglement entropy. There's been a bunch of these big ideas that the whole field is unpacking with the goal being to understand nature in a much deeper way. If I were to guess what would happen, I think it will probably lead to a new set of equations that capture nature on a deeper level than we have today.
ZIERLER: Shaun, I'm curious in graduate school if you interfaced at all with string theorists who of course are convinced that string theory is the likeliest path to developing a theory of quantum gravity.
MAGUIRE: Of course! There are a lot of people in that camp. I also, though, I think a lot of string theorists have gotten a bad rap. I'll say something that can get me in trouble. In my job as a founder of companies and partner at Sequoia and all this, being on lots of boards, I deal with the media a lot. One of my deep firsthand experiences that the media often time wants to build people up to tear them down, and I've just seen it. It's a universal thing across many different fields. They'll build someone up and then they'll tear them down. I feel like that's what happened with string theory. The string theorists weren't asking to have all these articles written about them and get all this hype in a lot of ways. The media built them up. Then the next version of that is to tear them down and make them seem like they were too arrogant, like, "Oh, it's not working." I think most string theorists have been—most, not all, some of them have been very arrogant—but the vast majority have been very measured in how they've thought about string theory and the current state of string theory and all that. Anyway, a bit of an aside.
ZIERLER: In your work on wormholes, just to clarify, are these toy models? Or are you thinking about actual wormholes?
MAGUIRE: We don't know. [laughs] Math is math, and there's a pretty good history of the math becoming reality. I think these are actually wormholes, and that's a huge point of disagreement. That's a personal, philosophical—it's like a religious conversation. I think there are probably actually wormholes, but it doesn't matter for the sake of the work. An equivalent thing is in quantum mechanics, people still debate the interpretation around wave function collapse and things like this. Do we live in a many worlds thing? There are a lot of different interpretations, and it's more in the realm of philosophy.
ZIERLER: What did you see as your primary contributions and conclusions with your thesis research?
MAGUIRE: I gave you the whole long story, but to give you the very simple story, the simple story is that AdS/CFT has been this really interesting thing in physics the last 20, almost 25 years. On the AdS side, that has a very deep relationship to hyperbolic geometry, which is something mathematicians have studied very deeply. Mathematicians know a lot of things; I don't think we're yet well-known enough by the physicists. My PhD is a very toy regime of three-dimensional gravity, two dimensional quantum sides—AdS, three; CFT, two. In that world, there is a deep relationship between the waves allowed in the space and the geometry allowed of the space. You can interpret that as a lower bound of the masses of particles allowed in the space. In that toy regime of three dimensional anti-de Sitter space, there's a concrete relationship where the more curved the geometry is, the more tangled the geometry is, the higher a lower bound would be on the masses allowed of particles in there. We live in a space where photons have a mass. We say they're massless, because if they were at rest, they'd be massless. But photons are always moving, and they have a mass when they're actually moving from a relativistic perspective. Imagine having a relationship between the masses of photons and the shape of space.
ZIERLER: Besides John, who else was on your committee?
MAGUIRE: John. Alexei Kitaev. Oskar Painter. Mark Wise.
ZIERLER: Anything memorable from the defense?
MAGUIRE: Those are days you don't want to remember.
MAGUIRE: I was insanely nervous.
ZIERLER: Did you get any curveball questions? Or did some interesting debates come up?
MAGUIRE: It's one of these weird things. When I proposed to my wife, I blacked out that day. I don't really remember any of it. I was so nervous. It's kind of the same thing. Honestly, I kind of blacked out. I don't even really remember.
ZIERLER: To foreshadow to what happened next, were you on a trajectory of pursuing an academic career and then some opportunity came up? Or did you know in the back of your mind that you'd be doing something besides academia?
MAGUIRE: I would say, a long time ago, I had to make the decision that I would go in another direction. But going back a long ways, going back to when I first started at Caltech, I thought I would probably be a professor, but when I went to DARPA, that was the moment when I had to choose between the two. Not completely explicitly, but a little bit subconsciously and implicitly. So, I basically made that decision a long time ago that I wouldn't do it. When I came back to Caltech, I had started a company in 2012, and it ended up being a relatively successful company. I sold it for a billion dollars, all of that. I was kind of doing both: doing the company and grad school. Everyone was telling me—people on both sides didn't understand why I was doing both. It's a little unusual in that on the company side I was doing it because—the reason why I was doing the company, in a lot of ways, is I got lucky. I got lucky in that when I was leaving DARPA, we came up with an idea. It became a program. It became a $110 million program. I had the opportunity to win an award—originally a $10.5 million contract to go build some of that thing that I helped come up with the idea for. It's hard to say no when DARPA is willing to give you money to go build some really advanced technology with really brilliant people. So, I felt like if I'm ever going to do something in business, I'm never going to get a shot this good, so I kind of had to do that in my mind. On the other hand, since I was a little kid, my passion was black holes and space. I really did the PhD for myself. My goal with the PhD was just to get to the cutting edge of knowledge in that field, because these things had kept me up in the middle of the night as a little kid, literally, for a long time. I felt like I just had to get to the cutting edge. It was this weird, internal drive. I just had to get to the cutting edge. Honestly, after getting to the cutting edge of knowledge there, it's weird. I think some people would be different than me, but I don't feel like I have to be the one to push it forward. Just knowing where that edge is, is enough. Being able to stay on top of it and having a lot of my friends be the ones pushing it forward, it's kind of enough for me.
ZIERLER: So, it was in some ways really a purely intellectual pursuit for you, then?
MAGUIRE: Honestly, yeah. It was purely just my curiosity.
ZIERLER: So then what happens next? The day after you defend, are you not looking at postdocs? Are you not looking at faculty appointments?
MAGUIRE: The day after I defended, I flew to Israel to get married, literally. Actually the day I defended, I flew to Israel to get married. Then, we went straight on my honeymoon. The day I got back, I went to graduation. That day, I was working. I had three jobs. Honestly, at the end of my PhD I had three full-time jobs. I was a partner at Google Ventures at the time. I was chairman of my company. I had moved from a full-time operating role to chairman, and I was finishing my PhD. I was doing all three.
ZIERLER: And when does Sequoia enter into the mix?
MAGUIRE: Sequoia enters shortly thereafter, but basically in the summer of 2019. I'd been at Google Ventures for three years, and I had the opportunity to move to Sequoia which is the best venture capital firm in the world, so it was hard to say no to.
ZIERLER: What were you doing at Google? What was some of the work there?
MAGUIRE: I was doing the same thing. Basically, I was investing in companies and taking board seats.
ZIERLER: The point of connection to Sequoia, how did that happen?
MAGUIRE: The point of connection to Google Ventures was simple. With my cybersecurity company—I really helped start many companies, but the cybersecurity company one—which was called Qadium, but then we renamed it to Expanse—that's the only one where I was really full-time with my company for many years. In many ways was the core person that drove it in the beginning, if not the core person. So, we raised a bunch of venture capital. I got to know a lot of funds. We met with and got term sheets from pretty much all the top firms in the Valley. That's how I got to know Google Ventures. The only firm that we never pitched was Sequoia, because they had a competing portfolio company, so we didn't want to give them the data on our company or something about us. So, I didn't really know anyone at Sequoia, but I was getting recruited by other firms.
I had been very lucky to meet this guy, Patrick Collison, who's a pretty famous founder now. Patrick started a company called Stripe. Founders Fund, which is another venture capital fund, invested in both of our companies. I met Patrick at a Founders Fund event many years ago. It's actually a directly relevant story, so I'll share it here. The way I met Patrick is pretty funny. It was a small event, call it 50 people. Founders Fund had flown us to an island off Vancouver Island in British Columbia. Patrick was talking to an extremely famous founder. People would know who he is and know the companies he started. But they were talking about quantum computing. This other founder was saying some things I don't think are correct and saying it in a really arrogant way to Patrick, and Patrick was pushing back and was correct in his understanding of quantum computation. I jumped in the conversation, because honestly I didn't like the way the guy was talking to Patrick, and Patrick was right. I tried to do the John Preskill Socratic method: ask some questions that would reveal the other person didn't know what he was talking about. Then that person got mad and left the conversation and left Patrick and me talking. We became friends from that. Patrick is a huge lover of physics. He dropped out of MIT. He was a physics major. So, we became friends. When I was at GV I invested in Stripe. When I was getting recruited by other funds, Patrick was aware. People were doing references with him. He told Sequoia, "You guys should hire Shaun." He emailed Mike Moritz, who's a legend in venture capital, and, Michael Abramson, and they ended up giving me a job.
ZIERLER: Shaun, with the entrepreneurial culture at Caltech, I wonder if your work has given you a broader perspective of the kinds of ways Caltech ideas, Caltech faculty and students are involved in technology ventures.
MAGUIRE: I wouldn't say that Caltech is the most entrepreneurial place. I would say that Caltech is more scientific—
ZIERLER: I meant relative to where it was maybe 20 or 30 years ago, not relative to Stanford of course.
MAGUIRE: I'm actually not so sure about that. I think Caltech might have produced a comparable number, or maybe even more high-impact companies in the past. I think, sure, the volume of companies is greater now, but Caltech had its hand in some pretty legendary companies in the past. For what it's worth, I think it's really important for the world to have places like Caltech that are so focused on science. I think Stanford is the other extreme, where Stanford in a lot of ways is just like, you go to Stanford because you want to start a company, and it's going to be the stepping stone to starting a company or joining a hot startup. Many would disagree with me, but I actually think it takes away from the quality of research at Stanford. Stanford does amazing research, but Stanford has a lot of faculty and a lot of money, and I actually think Caltech has higher quality research per capita. In my opinion, no question. There's a lot of amazing faculty at Stanford; I'm not trying to knock Stanford. I have a degree from Stanford. I think that's actually a part of the magic of Caltech: it's the only elite undergraduate and research university in America that is just so focused on science.
ZIERLER: That's pretty cool. Shaun, for the last part of our talk, just one retrospective question and then one going forward. What has stayed with you from IQIM and Caltech in general? Just in terms of the way you approach business, the way you understand science, the way you think about the world?
MAGUIRE: Many, many things. Some of it is implicit. Some of it is subconscious. I don't even know some of the things that I know are there, but I'll tell you some of the things that I'm aware of. One of the things is Caltech is a very humbling place. I think everyone that's been at Caltech, it has to lower your ego. It raises your ego in some ways, but it has to lower our ego in others. I think that's a really good thing. I think that as an investor, it's actually incredibly important. As an investor, you want to have intuition, but you also need to check your intuition with lots of diligence on things. Let me tell you, that's one of the lessons of studying either quantum mechanics or general relativity: your instincts are oftentimes wrong, so you have to actually go do the calculations. So, that's one thing that is really powerful. Another is just the network of people. I made a lot of my closest friends from Caltech. I've backed some people I knew from Caltech's companies. I've brought some people from Caltech into companies I've worked with. I've used people from Caltech as expert diligence when I've looked at companies. I think another thing that's very powerful about Caltech is that—it's actually something that we have in common at Sequoia—is that Caltech forces you to raise your ambition. When you go and you're around such incredible, brilliant people that go on to do such amazing things—being around so many Nobel Prize winners for example, or knowing that a couple people in your class are going to go win Nobel Prizes, it forces you to say, "Well, if they can do it, what's holding me back? They're human too. They're not that…" They're really, really smart, but having that exposure really raises your own personal ambition. At Sequoia, we have a lot of these flywheels, if I'm honest. One of the things that's a flywheel: because Sequoia has so much historical success and so many legendary companies in our portfolio, when our founders—just as a very recent example, Sequoia had invested in a company called Figma. It was just announced last week that Figma is going to be acquired by Adobe for $20 billion. The founder of Figma is an amazing 30 year old kid who also really loves physics and computers. When the Figma acquisition happened, it caused a lot of our other portfolio companies to raise their ambition. They said, "Man, I love Dylan, but like, I can do it, too. I don't want to sell for a billion dollars now, I want to sell for $20 billion. If Figma grew this quickly, we can grow this quickly." I've already noticed in the last week, I've had many founders in our portfolio come to me, and it's raised their ambition. So, I think that's something that's really powerful about Caltech.
ZIERLER: Finally Shaun, going forward, do you have a fluid view about your relationship with academia? Do you stay on top of the literature? Do you think about one day becoming a professor? Or are you a lifer in business and VC at this point?
MAGUIRE: One of the things that I am proud of in my own life is I've been willing to change course quickly even with limited data when crazy opportunities come up. So, I really love academia. I love Caltech, etc. I think I will definitely have more involvement with Caltech at some point in my life. I don't know what shape or form that will take. I'll trust my instincts when something comes up. With the literature, I am really, really busy. So, I don't stay up perfectly, but I do try to stay up with the really big results. One of the things that's interesting about the journey of being a PhD student is that you work so hard to get to the cutting edge. Once you get to the cutting edge, it's not that hard to keep up. The actual edge doesn't move that quickly. You have to claw your way from hell to get to the edge. Another thing too, to be very candid for me, I have very broad interests. I have been really interested in machine learning, and in cryptocurrencies, and in robots, and in space, and in physics, and other things. I try to keep up with all those fields. Right now, machine learning is probably the field that's moving the fastest, so right now I'm actually probably spending more time reading the machine learning literature than, say, the quantum information literature. Five years ago, quantum information was moving way faster than machine learning. These things change over time. They ebb and flow, so I try to go where the action is.
ZIERLER: It sounds like it's always exciting for you, no matter what it is though.
MAGUIRE: Someone's got to do it.
ZIERLER: [laughs] Shaun, this has been a great conversation. I'm so glad we connected. Thank you so much.
MAGUIRE: Thank you.