Hilton Lewis, Electrical Engineer and Former Director of Keck Observatory

In the oral history discussion below, readers will be treated to a behind-the-scenes look at Keck Observatory, one of the most impactful astronomical projects in history. Recently retired as its director, Hilton Lewis explains the origins of Keck as a partnership between Caltech and the University of California, he describes how its revolutionary technologies, particularly in adaptive optics and the segmented mirror design, have profoundly advanced our understanding of the universe, and he explains the importance of connecting the wonders of astronomy to community engagement near and far.

Born in South Africa and trained in electrical engineering, Lewis intuited the importance of the Keck project in its early design stages. After a three year planning assignment at Caltech, Lewis moved to Hawaii where he played an instrumental role in getting the telescope (now called Keck I) operational, from "first light" to hosting astronomers. The enormous success of the first telescope led to the construction of Keck II, and Lewis explains the power of interferometry and capacity of both telescopes to work independently or in concert. After serving as deputy director, in 2014 Lewis was named Keck's director, a position he held until stepping down in 2023. During his leadership Lewis worked to build community relations with native Hawaiians, he witnessed the building excitement around Andrea Ghez's Nobel Prizewinning discovery of the supermassive black hole at center of the Milky Way, and he led the observatory through the Covid pandemic.

At the time of the interview, Lewis was enjoying a full retirement. But, as he reflects below, his excitement for discovery in the universe and the ongoing role that Keck will play will almost certainly pull him back to the world of astronomy in one form or another.

DAVID ZIERLER: This is David Zierler, Director of the Caltech Heritage Project. It is Tuesday, August 22nd, 2023. It is wonderful to be here with Hilton Lewis, formerly of Keck Observatory. Hilton, it is so nice to be with you. Thank you so much for joining me.

HILTON LEWIS: A pleasure.

ZIERLER: To start, would you tell me, please, if you have any ongoing affiliations with Keck, or your title now is Retired or Emeritus Director of Keck Observatory?

LEWIS: Retired. I don't have any formal connections, although I obviously have many colleagues and friends still at Keck, and Caltech and UC as well.

ZIERLER: Hilton, are you enjoying a true retirement right now? Are you involved at all in astronomy or consulting or research?

LEWIS: Right now, I'm taking a break. I still have my intellectual interests, and so I'll see how this pans out over the next several months, and whether this is truly retirement, or I'll find some way to stay engaged.

ZIERLER: More broadly, what is exciting to you in the world of astronomy? What are you following, if only in the news these days?

LEWIS: It's all the big themes that have been present for quite a long time—the origins of the universe, exoplanets, black holes, and of course dark energy and dark matter, those are the big mysteries. Anything around Earth, which is basically all of astronomy, gets me excited. [laughs]

ZIERLER: Hilton, some questions about your own area of expertise, how you got to be the director of Keck. Are you an astronomer? Are you an electrical engineer? What's your field of background?

LEWIS: My background is actually electronic engineering. That's what I studied, and that's what I qualified in. Although most of the technical contributions I made were in software.

ZIERLER: Because you've been in astronomy for so long, would colleagues consider you sort of an honorary astronomer at this point? Have you learned a great deal about what astronomy is, what astronomers do?

LEWIS: I've learned a lot in nearly 40 years of being in this field. I have never presented myself as an astronomer, but I think I understand a lot about where the areas of interest are, and how the science of astronomy intersects with technology, and how to maximize that impact.

ZIERLER: With your focus on software and computation, what have been some of the real game-changing technologies for astronomy in the course of your career?

LEWIS: When I started, mini-computers were the computational environment. That rapidly evolved. Microprocessors became more and more powerful. And so, it was this bringing of what were formally very, very large powerful systems into the realm where ordinary programs and projects could access them. I think that's really where I saw the change. The other part was the whole maturation of the software, software engineering, and software technology, which went from these handcrafted systems that we were working on in the early days to something much more robust, industrial scale. Those were the changes I saw.

ZIERLER: When you stepped down from Keck, what were some of the really exciting initiatives? What was happening at Keck and how might that have influenced the timing of your retirement?

LEWIS: It actually was connected with sort of a stable moment. Stability is never long-term [laughs]. There were a couple of different forces that were going on. One was really what you might call the social environment that was really centered on the issues and controversies and concerns around astronomy on Maunakea in particular, but also Hawaii more broadly, in society and with indigenous peoples, impacts on indigenous peoples. That was one aspect. Although that story is still unfolding, I think we've made enormous progress in the last several years. So, it was a quiet time. We at least have a path forward. A lot of uncertainty, but it felt for the first time in a long while that we weren't stuck. So, that was one driver. I'm very interested in seeing how this social/science intersection plays out. I think it's a fraught path, and a lot of people have to think very carefully about how we are going to put science, at least astronomy, on a sustainable path in Hawaii. But of course, the goal there is not so much to make people love astronomy as to bring our understanding of science and the scientific way, the scientific method, to the broad community. That's one aspect I think that was really interesting. On the technology front, or opportunity front, there's all these fantastic new capabilities that have come online in the last several years and are still coming online. The most obvious one everyone knows about is the James Webb Space Telescope, and what an amazing instrument that is. I think that really is working so much better than any of us imagined. It was expected to be transformative, but we didn't really know—it's the ChatGPT of astronomy. [laughs]

ZIERLER: [laughs]

LEWIS: That's an amazing machine. Then of course there's observatories like the Rubin Telescope that's coming online, and then the extremely large telescopes, the next generation of telescopes that are like Keck but much larger, like the Thirty Meter Telescope and the Thirty-Nine Meter Telescope in Chile. I think those are the exciting enablers. In the science itself, the part that I was really interested in, and it's slowly coming to fruition, is all the multi-messenger astronomy. LIGO, and the way the high-energy physics approach intersects with observational astronomy is very interesting. And, of course, the science itself is fascinating. Then there's the neutrino side of it, with IceCube and others. That's a whole new window—I don't mean to sound trite, but it is a new window on the universe. How that will play out is unknown, and that's what makes it so thrilling.

The Astronomical Triangle of Caltech UC and Keck

ZIERLER: Some overall questions administratively about Keck and its relationship with UC and Caltech. First, perhaps at the simplest level, is Keck an independent organization, or is it a subset of something bigger?

LEWIS: It's formally a 503(c). It's an independent organization with its own board of directors, but it was set up to serve the research needs of Caltech and UC, and later—not very much later—NASA joined as a one-sixth partner in the enterprise. Keck is independent in the sense that it's not directly subject to the rules of Caltech and UC, but the mission is very much to serve those scientific interests.

ZIERLER: What is the partnership like with Caltech and UC for Keck, for the staff members there? How is all of the expertise that comes from these great institutions an asset for Keck?

LEWIS: It really has been a collaboration ever since I began, which is 37 years ago. I have never for a moment really seen the Caltech and UC campuses as something distinct and separate. I've seen us all as having our own focus. Caltech and UC have been the instrument powerhouses, but in the early days of course they also provided a lot of the telescope technology, starting with Jerry Nelson and his ideas and concepts, and Terry Mast, Gary Chanan. Those are the giants, the intellectual giants, that helped set the initial technology path. Then JPL played a huge role in the early management. For me, I've always been immersed in it, and I've really felt as though I'm part of the three campuses, if you like—Caltech, multiple UC ones, and then Keck itself. It's a continuum. It's not, "We're here, and we're getting help from Caltech," or "We're only here to serve Caltech or UC." It has never been like that. It has been a super, super close and very successful collaboration.

ZIERLER: Going on 37-plus years, were you part of the origin story of Keck? Were you involved in the discussions that ultimately led to the creation of Keck?

LEWIS: No, that started a decade before me, with Jerry Nelson and his collaborators. A lot of that was housed in UC in the early stages, and eventually, Caltech became an equal partner. When I arrived, all the initial conceptual work had been done, and we really started with the implementation. I arrived to start the software group that would write the control software for the telescopes and eventually for a lot of the instrumentation. So, certainly not the very beginning, but finding the very beginning of these projects is kind of—tricky. What constitutes it, and all the different branches. [laughs]

The Political Challenge of the TMT

ZIERLER: You mentioned the education, talking about astronomy and Maunakea and the indigenous peoples. When discussions about an ELT project, maybe even before the TMT became codified into an organization, did you see some of the writing on the wall in terms of the concerns, the divide between the astronomers, academic society, and some of the native peoples on Hawaii? Or did you experience that as people began to protest?

LEWIS: Keck tried to build outrigger telescopes as part of the interferometer project. That was in the early 2000s. We got a very strong reaction from the local community, negative reaction, which we were completely blindsided by. I would say we had no clue that that was going to happen. We were surprised at the extent of it and the ferocity of it. By 2002, 2003, which is when the outrigger project was abandoned—there was still a component consisting of just the two existing Keck telescopes that continued—we knew there was a serious problem. But at that time, at the origins of TMT (which was then called CELT), that telescope was not necessarily going to be built in Hawaii, so we didn't initially understand those sort of danger signals. But I'd say by about 2006, 2007, many of us, myself included, were concerned that this was not the slam dunk that was being presented. I think we tried to make that clear to people, but—I always joke that every new telescope project knows everything that it needs to know, and no one else can tell it anything. I've been just as guilty of this myself, over several projects. I think our concerns, they weren't necessarily focused enough for people to understand, but there were definitely early warning signs, and there were definitely concerns that we expressed.

ZIERLER: When you served as director and you had to worry about Keck as a whole, the big question about would the TMT be built on Maunakea or would it not, in what ways would it affect Keck, for better or worse, as you were thinking through these contingencies?

LEWIS: Let me just step back a tiny bit from that. Of course as Keck director my responsibility was Keck. On the other hand, I very heavily vested in Caltech and UC's astronomy enterprise. Certainly I understood the importance of the TMT to that. The approach I always took was, I wanted to see TMT built, even though the day that TMT—or the ELT, for that matter, the European version—turns on, it will essentially render Keck obsolete in the areas where it has been most successful. They are aimed squarely at what Keck does very best. In contrast, the Subaru Eight Meter Telescope has quite a different niche that it dominates, and that will continue. Without changing much, Subaru will still be just as effective. But Keck had to, and has to, change significantly. So, it was always a little ambivalent for me, because on the one hand, the thing I was—I don't want to be melodramatic; not sworn to protect—but the thing I was advancing was going to be made obsolete by the TMT, yet I wanted to see it both for the science and for the institutions that I served as well. The bottom line is I always told my leadership team that we 100% supported TMT being built.

ZIERLER: When TMT is built, if TMT is built, how will Keck respond? How will it remain relevant given, as you just observed, it is aimed so squarely at what Keck does best?

LEWIS: You asked me a little earlier why I chose this moment to step down from Keck, and one of them is we have just completed our strategic plan through 2035. The reason for 2035 is it's two years after the lease needs to be renewed, so we made the assumption the lease would be renewed [laughs]. The Strategic Plan looked very carefully at what Keck would have to do. There are a couple of ways it can stay relevant, but it does require a change. One of the very important things is do the things that TMT is not slated to do for a long time, the classes of instrumentation that will not exist on the early ELTs. An area where a telescope like Keck can be very powerful is optical adaptive optics—adaptive optics in the optical wavelength regime. So, there's those kind of really narrow—well, not necessarily narrow; real niches—that we can occupy for a very long time.

The other part that I think is extremely powerful is the way we've used Lick and Palomar in the Keck era, which is you build instrumentation, and you prove the instrumentation technology concepts and the science concepts, at a much lower cost. You do good science, and then once that's demonstrated, then you can reimplement for the bigger telescopes. For example, Keck has an instrument called KCWI that was first built at Caltech 15 years ago, and although the Keck KCWI is a vastly different instrument, you couldn't have started on Keck. It would have just been too expensive. There's an opportunity cost. Your biggest telescope, you can only afford to use on most "sure thing." Now you need to know that instruments will work, that the science cases are real, that you're likely to get a return. When you're a smaller telescope—and this is what the 4- and 5-meter telescopes demonstrated in the 8-meter to 10-meter era—you can take a lot of risk. It's high risk but high reward. That's the way Keck will go.

ZIERLER: I wonder if you can explain just why Maunakea is so good, even in a singular way, for land-based astronomy on planet Earth.

LEWIS: Well, it's very high, so it's above most of the water vapor in the atmosphere. That's one factor, that it enables certain windows of observation you can't do lower down. It's above the cloud layer most of the time. It's very dark. It's still an extremely dark site, although that's gradually being degraded with time. The air flow over the mountain is laminar—it's very, very smooth—and so you can get extremely sharp images, especially with the help of adaptive optics. The idea of adaptive optics is it removes the turbulence effects of the atmosphere. If you start at a bad site, you make it better, but if you start at a great site, you can make it extraordinary. It's really the combination of those three things. There's a fourth one which is not often appreciated, and that's logistics. It's so much easier to get people and equipment to the summit of Maunakea than it is to some of these really remote sites, in the high Andes, for example, and that matters as well.

ZIERLER: I wonder if you can explain the relationship between Keck and space-based astronomy. You mentioned before, when NASA came in, that was obviously very pivotal. Generally the NSF is the supporter of land-based astronomy. How does Keck work in concert with space-based telescope operations?

LEWIS: One thing worth noting is that the same people use all these different assets, space and ground. Many of the UC and Caltech researchers are very active with all the NASA facilities, and also with many different ground-based facilities. So, you have that sort of automatic knowledge of all the capabilities you have, in our human system. Keck was extraordinarily successful with certain telescopes, Hubble—the HST—being the prime example. Because HST could identify things that were really interesting, and then you could use Keck to follow up because Keck was a so much bigger a telescope. You can collect more light. It is this complementarity that really is extraordinary. Another example was the Kepler Mission to identify exoplanets, and now TESS. You can use telescopes to remove the ambiguity about lots of these objects, whether they really are planets or not, and to study aspects of them that you just simply can't do with a small telescope in space. I think JWST will be another amazing example. There, one of the earlier interesting cases was where Keck could be very rapidly pointed as something—a solar system object, one of the planets—to identify clouds on Saturn's moon, Titan. That ability to use a ground-based facility which is generally more flexible than a space one really allows you to plan the next science experiments you want to do in space. This back and forth, that's really powerful in all of these cases.

ZIERLER: You mentioned the lease and 2035. I wonder if you can explain the mechanics. Who is the lessee? Who is the lessor? How does all of that work?

LEWIS: 2033 is actually the lease renewal date. The individual observatories have all to get a lease, or a sublease actually, to continue to operate. It used to be that the University of Hawaii held the master lease. So, the state provided a lease on the land to the University of Hawaii, and then they negotiated subleases with the individual observatories. What's different is that UH no longer will be part of that process. There's a new statutory authority, the MKSOA, Maunakea Stewardship and Oversight Authority, that is charged with negotiating the leases with the observatories. In principle, you replace one lessor with another one, but in practice, it's a very different creature. The University of Hawaii was strongly incentivized to maximize the astronomy component, because their IFA, the Institute for Astronomy, was a major beneficiary of the telescopes. The new authority has to take into account all the other uses of the mountain as well, so it's going to be a more complex process.

From the Observatory's point of view, we want to have uniform subleases or leases; we don't really want to be negotiating different kinds of things. Because that can be, first of all, very hard to coordinate, and secondly, it can be very challenging for the people involved in planning the operations of these facilities going forward. How much do you have to pay? What conditions do you have to provide? I think all the observatories are highly motivated for the leases to continue. It's not a lease-at-any-cost kind of deal, though. It has to be something we can manage. Now, the observatories don't make money on observing time—they're pure research facilities—and so the owner institutions, in the case of Keck, Caltech and UC, have to be able to afford to operate the facilities. That being said, I don't think any of us are planning—no one is planning to walk away. We all plan to make this work. But it's a different environment.

The Two Keck Telescopes

ZIERLER: When we talk about Keck, obviously we're talking about Keck I and Keck II, the two telescopes. I wonder if you can explain how they operate either independently or redundantly or as complements to each other. How does that work?

LEWIS: They essentially operate independently. They were conceived of as an interferometer. Let me restate that. It was conceived that an interferometer mode would be available, which means you combine the light of the two telescopes simultaneously. We did that for a while. The outrigger telescopes that were never built would have made it much more efficient and effective. Eventually it turned out to be too expensive in terms of science return. The interferometers weren't sensitive enough to be of broad enough interest to the science community. You're tying up two telescopes for one observation. So, we've essentially, since we decommissioned the interferometer a little over a decade ago, run them independently, though occasionally we observe the same target with the different instrumentation on the two telescopes, for example the black hole at the center of our Milky Way. But the vast majority of the time, they're just two telescopes on a great site. The benefit you get is they're functionally identical. They're not physically exactly identical but very close, so the same staff can maintain both. You get tremendous efficiencies of operation. What you learn on one, you can apply to the other. There's that kind of benefit that you don't get if you had one in North America and one in the Southern Hemisphere somewhere.

ZIERLER: I know it's not a perfect analog, but in the way that, for example, LIGO has two interferometry sites for the purposes of verifying the observation of gravitational waves, does Keck I and Keck II work in similar ways? Does one telescope see something, and then it's verified when the other does as well?

LEWIS: No. Even the LIGO one is not that simple—the fact that an event occurs at both sites, it's a check that you have a real event, but more importantly, you use that as a way of locating where the object is, the general area on the sky that it is in. The Kecks aren't used simultaneously to do either of those things. The only time we would use them in concert is if you want to use different kinds of instrumentation on the same object at the same time. Because the thing about ground-based observing, the atmospheric conditions matter a great deal, so if you observe something tonight and you observe it a week later or a year later, it's not necessarily that easy to correlate. But if you observe them effectively simultaneously, it's a powerful additional tool.

ZIERLER: I wonder if you can comment on the politics of observing time—who gets it, how that's negotiated, what role Keck has to play in which astronomers get to use it for which projects.

LEWIS: Keck plays a very small part in that process, because the science utilization is governed by the participating institutions—Caltech, UC, NASA, and there are a number of smaller scientific collaborators. That process is supposed to be meritocratic, so it's based on the proposal and the ability of the team to carry it out. There has been a large amount of effort over the last decade to try and remove biases of various kinds, particularly gender biases which were well known effects. There has been some real progress. I don't think we're all the way there yet. The problem is, it's a small field. If you had a million astronomers, you could randomize things. When you've got 30 people, 20 people, it's a lot harder to really not pay attention to who's doing the research. And of course experience comes into it. An experienced team is more likely to succeed than an inexperienced team. So, there are these different sociological effects that groups are trying to deal with in order to assign time fairly. The role that Keck has is to confirm that the science being proposed is technically feasible. If someone wants a particular filter—this is a trivial example; it wouldn't really happen in practice—but that we have that filter on the instrument. But it's more than that. It's the kinds of observations that must be feasible. We really just play a technical verification role.

ZIERLER: You mentioned previously the accessibility factor makes Hawaii a great place, as opposed to some of these rural places in South America. I'm smiling because that doesn't mean it's not a massive trek up the mountain. What staff members are permanently stationed or are there as much as possible? Administratively, where are other members of the Keck team located, if not on top of the mountain?

LEWIS: You actually implied this, but I'll just state for the record—no astronomers go to the mountain. No users of the telescope go to the mountain. About half of them go to Waimea, our base facility, on the Big Island, and the other half are all remote, effectively Zoom. Actually not just effectively; literally. [laughs] The only ones who go to the summit are the regular day crew who are there for maintenance purposes and to reconfigure the instruments for different kinds of observations. Then when a new instrument is being delivered, installed, and commissioned, there's a flurry of activity. Typically, you'll have 20 people on the summit in a given day, a little less on some days, a little more on others, but when a new instrument arrives, you might get 40 people go up—lots of engineers, technicians, scientists. The majority of the staff are actually at the headquarters in Waimea. Keck has a staff of about 145—144 after I left. [laughs] Roughly; it changes. Most of those are engineers, support staff, administrators, and staff astronomers who are the ones based at the headquarters. The purpose of the staff astronomers is to support astronomy. Keck is not a research facility in its own right. The research entities are Caltech, UC, and NASA along with UH and some other scientific collaborators.

From Base to Summit

ZIERLER: I wonder if you can narrate the journey from the base to the summit. What's it like? What precautions or preparations should you consider? Then what does the world look like when you're on top?

LEWIS: Okay, I'll start with that—fantastic! [laughs]

ZIERLER: [laughs]

LEWIS: I actually gave a whole talk on this subject, because folks had a lot of interest in it. First of all, the physical travel from headquarters is about an hour to the midlevel facility, Hale Pohaku. That's at 9,000 feet. Then, depending on how acclimated you are—if I were to take you up, we'd stay there for at least an hour, sometimes an hour and a half. That's just to allow you to get adjusted to the altitude. Then it's a further about 35 minutes to the summit itself. The first part is a dirt road, but the last part is paved, the last few miles. There's a paved section, but it's not all the way, by any means. If you were going up in the evening, which very few people do—our night crews have three people typically—you'd stop at Hale Pohaku, have dinner at 5:00pm, and then before sunset you'd start your journey up the mountain.

Of course the summit is always chilly. It hovers around zero Celsius most of the year, 32 Fahrenheit, and it gets as cold as 20 Fahrenheit. Very occasionally I've seen it a little colder than that. But windchill can be something, so if the wind is blowing, it's miserable. [laughs] When you get up there, it's definitely high altitude. It's 14,000 feet. Most people are okay but about one in five or six get a headache. Sometimes people feel very poorly, but the remedy is pretty straightforward—you bring them back down to Hale Pohaku. It's very rare for someone to be seriously affected. It does happen. I think in my entire time working at Keck, one person died from altitude sickness, and probably a million people went up the summit in that time, so you're unlikely to have a fatal encounter. But as I said, about one in six feel a little off, and a few of the others feel worse—when I take tour groups up, probably every five or six groups, I'd have to have someone brought down, to Hale Pohaku. So, you get up there, it's cold, you feel a little lightheaded, but when you get to look around, well the summit itself is spectacular. You can see the background in my zoom session is sunset behind the two Kecks. It really looks like that, so it's very otherworldly. Then of course when you go inside, the office facilities are at a normal temperature, like the control rooms, the labs and so on. But when you go into the telescope area, it's kept at the midnight temperature, so it's frequently freezing or below freezing. Then you get confronted with this gigantic structure, the telescope, which is just extraordinary. We used to have a site superintendent who, when we'd bring visitors up there, he'd say, "Welcome to the Sistine Chapel of Astronomy." That actually captures it. Most engineers and scientists are rather pragmatic people, but it takes everyone's breath away. And, if you have a poetic soul, it's even better. [laughs]

ZIERLER: All of the instruments in the domes—the space is obviously finite. How does that work? Does one instrument come down only when it's obsolete? Is it put in storage and it comes back based on a new project? How does that process work?

LEWIS: There are multiple locations on the telescope where you can bring the light that's gathered by the telescope to a focus. We have instruments in most of those. There is one, for example, behind the primary mirror—there's a hole in the mirror—and that's called a Cassegrain focus, so you can mount an instrument there. Or you can mount it on the side and use another mirror to redirect the light out to that side. So, there are multiple places. There're reasons for using one location over the other, to do with factors like loss of light, the mass you can support, and so on. But nevertheless, there are more instruments than there are active foci, so you can't have them all bolted onto the telescope at the same time. So, there's storage and a complicated switchyard that allows you to move instruments in and out of the beam in order to change the instrumentation. It does happen that eventually an instrument gets old enough and obsolete enough that it gets taken out of service and into retirement. In all the time I've been there, I think we have decommissioned two instruments. It's very, very slow—people are very hesitant to let them go. What happens is that as the technology develops and new instruments come in, people tend to shift to those, but there's always one or two science cases that can be optimally done with the old instrument, so you're very reluctant to cut off that little community, because they're still doing great science. It's a perennial debate. You asked me earlier about one of the things that has to happen at Keck to compete with the TMT, and a big part of that is deciding to decommission instruments. Because we have to find ways to save costs and to operate more efficiently, and it's inefficient and costly to constantly change instruments.

ZIERLER: These days, of course, everyone is talking about artificial intelligence and machine learning. I'm curious, with your background in software, have you seen Keck and more generally astronomy embrace AI, and what might this mean for the future of land-based astronomy?

LEWIS: Certainly the data sets have gotten larger and larger, and so machine learning has just become essential to understanding those properly. At the campuses at the universities, there has been a very strong effort in all kinds of areas, for example around exoplanet research and stellar modeling. It's a very, very active field. At the telescopes themselves, which is more about control systems, people have been much slower to embrace it. This really came home to me because in March this year I called a meeting of our staff to talk about the new AI tools. ChatGPT was the thing that woke me up and I realized how it was revolutionizing so much. This was three or four months after ChatGPT was introduced. The response I got from staff was very telling. I'll tell a story in a moment, and I'll go back and explain why I'm sharing it. The staff we had who were non-technical—HR, IT—I don't know if IT is non-technical, but I mean, they don't write software—those folks were kind of blown away with the potential, and what we were reading in the media, what kinds of things we might be able to do. But the hard-core technical people—engineers and particularly software engineers—were much less swayed by this. They said, "Oh, we've seen this before. It's just repackaging old concepts with new labels."

It took me back to when browsers first were introduced to the general public. Many people in science had been using the internet, or the ARPANET originally, and we had used email, and we knew how to search for information online. Then, the first browsers were released, and we said, "Big deal. So it's a little easier to do it," but we completely missed this transformation that was coming. I mean literally within a few years the whole world was using the internet, gone from 100,000 scientists and engineers to a billion people. We certainly could not imagine the kind of transformation that was upon us. It's very much like that, I believe, with the AI tools that we're seeing today, in that we've been using a lot of the machine learning tools, and deep learning, and we've applied them successfully to particular problem sets. But there's something else in the air now, this sort of vast integration, this exponential increase in capability. It's really hard to know how it will transform astronomy, but clearly it will. It's going to transform all the technology side for sure—design, operations, maintenance, the sort of bread and butter of running a telescope. I'm sure it will have a dramatic impact on our ability to process and understand these giant datasets that we're now generating in astronomy with petabytes of data in our observations.

From South Africa to Hawaii

ZIERLER: This has been a great tour of Keck and astronomy in general. Let's go back and establish some personal history. Where did you grow up? What was your childhood like?

LEWIS: I was born in South Africa, but as a young man moved up with my family to what's now called Namibia. At that point it was just ruled as another part of South Africa. I grew up in a small town of about 50,000 people in the middle of nowhere. The big nearest cities were a thousand miles away, over dirt roads. So, pretty remote. But, I had a standard education, I wasn't deprived in any way. We studied mathematics, and physics, and chemistry, like any other kids. But we didn't really have examples of any of the really modern technologies. The area I lived in was all about mining and agriculture. That's sort of what I knew. But when I was about 15, I won a place in a national science competition and was sent off to Johannesburg, which is one of the biggest cities in South Africa. One of the places we went to visit was the NASA deep-space tracking station at Hartebeesthoek, outside Johannesburg. That was my first exposure to a telescope of any kind—it was a radio telescope. But it was also my first exposure to scientists and engineers in high technology. One of those people—I don't know whether he was a scientist or an engineer—took an interest in me and spent about an hour talking to me, while the tour wandered around outside. And that's really what sparked my interest in astronomy and astronomy technology.

I always was interested in technology. My father was an engineer. But I had no idea how it could be applied in such an extraordinary way to understanding fundamental—well, to me, fundamental—physics. Not that deep-space tracking stations are fundamental physics, but it's much closer than mining. [laughs] That impressed me. I really was amazed because that's the first time I saw atomic clocks, and computer systems, and all that kind of stuff. There was a parallel aspect that didn't surface for a very long time: it struck me only years later how powerful these kind of mentoring opportunities are. I mean, it changed my life completely. When it became my turn, when I got to more senior leadership, and some kids would come through Keck, I'd always take the time to sit and talk to them, because you never know when you'll light that spark with someone. That was very much my experience. That's what got me interested in astronomy. Then I kind of didn't pay attention to it for a while, went off to university, became an engineer, and was actually on my career path in the industrial world. I had enrolled in an MBA program, and then saw a job ad for an engineer to build a telescope, in New South Wales, Australia. [laughs] I thought, "Oh, I'll do that." I remembered how I was very interested in astronomy. I said, "I'll do that for two or three years. Then I'll go back for the MBA." [laughs]

ZIERLER: [laughs] Growing up in South Africa, what was your awareness of the racial politics? Did you understand what apartheid was? Did you live in the middle of it?

LEWIS: Absolutely. I grew up in what would have been called a very liberal family. But when you're in a system, it's very hard to see what you're in, and until I went to college, I really didn't understand it. I grew up in a place where different races weren't allowed to swim in the same swimming pool. We certainly didn't live in the same neighborhoods. We were forbidden to go to what was called the township areas, which were on the outskirts of the city. In fact, the town I grew up in, I described as having 50,000 people—when people would describe it, they would say, "It's a city of 50,000 White people and 50,000 Black people." But the economic power was of a city of perhaps 51,000 people. Because those who were Black did not have opportunities—they really had very, very little. That's the kind of environment that I grew up in, as a kid. But when I went to college, the scales fell away from my eyes. That was the first time I understood what I was inside. And that's what impelled me to leave South Africa. We had compulsory military service, the draft effectively, and as soon as I finished that, I left for Australia, because I did not want to live in or raise my family in a country that had those policies.

ZIERLER: Obviously when you made that decision there was no vista by which apartheid would end anytime soon. There was no Nelson Mandela on the horizon.

LEWIS: He was in jail, of course. But there was never going to be change—it was going to be "never in our lifetime". And, when people spoke about change, they spoke about it in apocalyptic terms—rivers of blood. I didn't want any of that. I left. I left my family behind and went off to a new place. I had no desire, whatsoever, to be part of the South African system. Even though I benefited from it, because of my skin color.

ZIERLER: Tell me about the adventure in Australia. What was the opportunity? What was compelling to you there?

LEWIS: Well, it wasn't South Africa; that was the first thing. [laughs] Culturally, it had a lot of similarities, right? Similar climate, English speaking. I was interested in technology at that time, and there were a lot of technological opportunities. But what kept me there was I helped build a telescope in the north of New South Wales. It was sort of, if you like, a prototype for Keck, very similar. Not the primary mirror, which is a key part of the Keck telescope, but the whole way the rest of the system worked. The entire telescope I built was only marginally bigger than one Keck segment. There are 36 segments in the Keck primary mirror. But I loved it. I felt like I got back to what I really wanted to do, even though I didn't know I wanted to do it [laughs] at the time. It was just very exciting. I loved building something that had real value. Before then I worked for Control Data, a computer company, and I worked for a company that did electrical distribution networks. Those were great jobs, interesting technical problems, but it didn't hold a candle to doing something that has a purpose like that of astronomy. Not to me, anyway.

ZIERLER: To clarify, was this before or after your education at the University of Cape Town?

LEWIS: After. I got my engineering degree in Cape Town.

ZIERLER: And then when was the draft? That was before or after college?

LEWIS: After. We used to have a choice, but I went after college. I spent two years in the Navy, and after that, I worked for a few months in South Africa and then left.

ZIERLER: When you focused on electrical engineering at the University of Cape Town, did you have any inkling that this could be applied toward astronomy? Did you take classes in astronomy?

LEWIS: I took classes in astronomy and I took a lot of classes in physics. In fact, one of my professors was George Ellis, who was a collaborator with a number of noted astronomers, including Stephen Hawking, in cosmology. So, I was deeply interested in that, and I was deeply interested in physics. My father pressured me very heavily not to do physics. He felt a PhD was a waste of time.

ZIERLER: [laughs]

LEWIS: He was an engineer. He said, "Why waste time on an academic career? In fact, even doing electronics I remember was a big fight with my dad, because he felt it was like too—unimportant [laughs]—compared to generating energy and transporting it [laughs] around the world. I actually did four years of physics and applied mathematics as well, so it was always a latent interest of mine. But there weren't really the opportunities in South Africa at that time, so I didn't see my way forward in that.

ZIERLER: How long did you spend in Australia?

LEWIS: Not very long. I was only there about five years, a little under five years. I spent a good deal of that building a telescope. The reason I left was because of Keck. Someone at Keck wrote to our group asking if anyone was interested in joining them. And I was. We had finished building the 2.3m telescope at Siding Spring. I won't bore you with how I actually found the ad, but my boss did try and keep it hidden from me. [laughs]

ZIERLER: [laughs] To go back, that interesting comment you made about it's difficult to pinpoint the origin story of these big scientific projects, were you aware of the planning stages? Had you known who Jerry Nelson was? Did you have any idea that there was a big project about to happen?

LEWIS: Just before I saw the ad for the job, I read a piece in TIME Magazine, and that sparked my interest. This was billed as the largest telescope on the planet, which it was. It was an interesting article. I'm sure it mentioned Jerry Nelson but I had no knowledge of him. I did know about Cliff Stoll who was on the project and who wrote the book The Cuckoo's Egg. He was the person who caught the first hacker, or first person to catch a hacker maybe, at Berkeley. Cliff had written a letter to me saying, "I'm consulting on the software for this project. I've got some ideas. What do you think of them?" He told me what they were. I wrote back and said, "Those are all insane. None of them will ever work." [laughs] Later I met Cliff, and he's a fantastic guy. And it's true; those particular ideas would not have worked. But that was my introduction. So I had this little small introduction. But I did get to know Jerry very well. I worked very closely with him. And that was, I would say, one of the great experiences of my life, my working life.

ZIERLER: Did you ever get a sense from him, did you ever talk on the level where he explained his founding vision for Keck?

LEWIS: I don't think that's quite the right way to think about how Jerry saw Keck. He had this idea for how to build really large mirrors, the segmented mirror approach, and he worked for a long time to prove it could be done. He saw Keck as an implementation of that, not as an end, in itself. Jerry was an astronomer, but he was also an amazing engineer. I wrote an article about him in Scientific American, an obituary, which if you're interested you might read. He was an engineer's engineer. Because he thought about everything from first principles, and he really taught those around him to approach it that way. No formulaic approaches. It was extraordinary. I had never been exposed to someone like that. I mean, he was genuinely brilliant, but also the way he approached problems was almost childlike in the way he thought about them—he was very inquiring about things. And he didn't take anything for granted, ever. I found that really astonishing.

ZIERLER: Had the segmented mirror concept been implemented anywhere before, or this was entirely brand new?

LEWIS: Brand new. There was someone who was experimenting with the idea I think in the thirties or forties, but it was really Jerry, and a guy called Terry Mast, who were instrumental in figuring out how to do this. Then there was another scientist, Gary Chanan, who's the only one of the three still alive, who figured out how really to position the segments so the mirror worked as an appropriately shaped surface. Those three guys were the ones who made it happen.

ZIERLER: Do you remember landing in Hawaii, what your first day was like?

LEWIS: Oh, I do. I remember arriving in Hilo, and seeing tiki torches for the first time, and saying, "This is amazing! This is fantastic!" [laughs]

ZIERLER: [laughs]

LEWIS: I went there a few times before we relocated—we were based at Caltech for the first three years and then moved over. I did go over occasionally, because there was a site office. But I didn't come to Keck to go to Hawaii. I had no interest in Hawaii. I had been to many beautiful places. I didn't have an objection to it, but I was there to build a telescope, and my goal was to build a telescope and then leave. But, I got there, and there was a lot more work to do than I had realized, and then, you know, I had children, and children anchor you to a place. And, I never left.

Design Planning at Caltech

ZIERLER: You said the first three years you were actually at Caltech?

LEWIS: Yeah.

ZIERLER: What was happening there? What were you doing at Caltech?

LEWIS: There was a project office, the corner of South Wilson and California, so right at the corner of campus. There were about 20 of us, mostly engineers, who were involved in the design. The management personnel for the program came out of JPL, so that group was running a lot of the big contracts. Software we did in-house. A lot of the mechanical stuff and electrical stuff, we did the requirements and basic design, but building those were done as subcontracts to outside vendors. But the software we designed and wrote ourselves. We had a tiny little team doing it, actually.

ZIERLER: Was the original plan for you to be based permanently at Caltech, or you knew eventually once everything was up and running you would move to Hawaii?

LEWIS: No, that was always the plan, to move. Some of us came back to other faculty positions or staff positions at Caltech or JPL, but I was part of the Keck staff, I wasn't seconded, so I knew that I was Hawaii-bound. It's just that I had no intention of staying. I just thought, "We'll build it, commission it, and I'll leave" [laughs]. But it's one thing to think that; it's another thing to do it, actually. [laughs] Then the operations aspect became really fascinating. I had no idea what it would take to make not just a big telescope work, but to get the research output, it's an enormous confluence of factors that have to all succeed. When I started, I was only interested in the technology, and to some extent the science. I realized as I was working on it how much the science drives what you have to do with technology. Getting that judgment right of which requirements are needed and which are not, that's a very sensitive function, and many observatories have not got that right, though some have. Keck was one of them that did. Then there's all the sociology. How do you get people really to care about it? Keck is renowned, I think rightly, for its service ethic toward astronomers. It has always been the observatory where people will not give up—where the astronomy, the actual observing, is regarded as completely crucial. You can build the machine, but if you don't observe, it's not worth anything. And if you observe but you can only observe for an hour, and then it's down for an hour, and it runs again, or it shuts down for an hour, you don't do anything either. Keck is really focused on serving the science, and the scientists who carry out that science.

ZIERLER: What was happening on the mountain for those three years you were at Caltech? Were they already starting to build?

LEWIS: Right, they had started the construction, the excavation. By the time we relocated over, the building was—the enclosure was in place. We were finalizing some internal stuff. Then we started installing equipment—drives and electronics and various systems to make the Observatory work. And eventually, of course, the mirrors.

ZIERLER: We should clarify here, in these early years when we're talking about Keck, there's no Keck II; it's only the Keck Telescope.

LEWIS: Right.

ZIERLER: Was there the sense that the next telescope was in the offing? Was that part of the planning?

LEWIS: It was part of the planning at the leadership levels. People like Jerry Nelson, Gerry Smith, knew about that, and they were careful not to design out that possibility. So, the excavations were made so that you could add a second telescope. But for those of us who were, quote [laughs], "doing the work" by which I mean the junior engineers—

ZIERLER: In the trenches.

LEWIS: We were the ones who wondered what the heck the leadership does [laughs] and why they're always in the way, we were unaware of that planning pretty much. In fact, when we finished Keck I, and the Keck Foundation decided they would make a grant for the second Keck, many of us, perhaps all of us junior engineers, were opposed to it. We said, "This is ridiculous. We haven't even made the first one work, and now you want us to build a second one. This is a distraction." But of course we didn't understand from our lowly perch that if someone offers you $100 million to build a telescope, you don't say, "No, I'll take it 10 years from now." From a strictly engineering point of view, it seemed like the wrong thing to do, but in hindsight, quite apart from the political and financial factors, it was a great idea. Because we had the original team there; we all knew exactly what we'd just gone through. So, we started building the second telescope. I'd say we got the two telescopes up and running much faster than if we had built them really sequentially. But of course it was at a high cost. High human cost as well.

ZIERLER: You mentioned an early red flag based on the objections of the native communities in Hawaii with the outrigger telescopes. During the construction of Keck I, was there any sign of this whatsoever? Were there any misgivings that you were aware of?

LEWIS: Absolutely none. I certainly wasn't aware of any. And I never heard anyone talk about that. That's why I say we were really blindsided when the public opposition materialized to the outrigger telescopes. We had just not anticipated that.

ZIERLER: Was there an outreach component from the beginning? Did Keck make pains in terms of employment, in terms of opportunities, to include the native community in this project?

LEWIS: Not the native community, specifically. There was always outreach, but it was always seen in terms of STEM outreach. We always had people going into the schools to teach and mentor. I went to give talks at the local schools. We had kids come in. We gave public lectures. We always wanted to be part of the community. Indeed, most of the staff live in the community, and for extended periods of time. I'm not an exception. But that's how we saw our role. And I think that's how we saw the role of science and technology for the community, in terms of education. There were other warning signs, of course. In the fifties, maybe even in the early sixties, science was seen as salvation for everything. And then, that changed, right? There was Rachel Carson. There was the opposition to the Apollo program. Then after that, the total drop-off of interest. Then of course we've come to this era which is, if not anti-science, at least significant parts of our community are anti-science. All along, people were dutifully marching off to school to try and motivate kids, and tell them how great and important STEM was. But we missed something along the way. I think that happened to us in astronomy, too. We didn't really see it as our role to transform community, or that we had a role in this broader drop-off of interest in math and science. We just felt we had something great to share, and we wanted to share it. We always were preaching to the converted. We always talked to those people who were really interested already. We never actually had a program to change—to do hearts and minds. [laughs] And that's changed. Many of us have seen—back up—I think we always understood that a lot of our funding is public. Even the money that comes through Caltech and UC, it's essentially public money that is funneled through the universities. So, we had an obligation—everyone always felt we had an obligation—to give back in some sense. But we didn't see it as integrating within our community in a really deep way. That has definitely changed, and this whole experience has taught all of us that we were approaching it in a high-handed fashion. It wasn't our job—it was someone else's job—to worry about these things. Our job was, follow the law—we were always scrupulous about that. Many if not most of us are environmentalists at heart, care a lot about the environment. Anyone who suggests we would hide pollution, or a spill have no idea of the kinds of people we're talking about, because our people care very deeply about the environment. But community was a little farther away. That was somehow someone else's responsibility. And that has changed. That has changed.

ZIERLER: Hilton, when you moved to Hawaii and you were involved in the building, what were some of the big engineering challenges, from the mechanical to the civil to the electrical, once Keck I started to get built up?

LEWIS: The biggest challenge was actually integrating all of those fields. A telescope is an engineer's dream—[laughs] or multiple engineers' dream—because you have all of it, right? The civil, the structural, the mechanical, electrical, electronic, software, optical. It's an amazing collection of different technologies and disciplines that have to work together. I was a software engineer but I had to be concerned about how the foundations were settling, because I was doing some of the analysis of where the telescope was pointing. Or, that the gravitational vector was deflected by Mauna Loa, this giant volcano next door. So, you sort of get involved in areas which you have no idea about. As an electronic engineer you think about circuit design, or a software engineer you think about real-time programming, but if you are to be successful, you have to embrace all these different disciplines. And, they all talk different languages. So, it's a very challenging systems problem. That's really it. The individual disciplines can be tough, right? The optical designs are very tough. Some of the software criteria we were trying to meet in the area I was expert in were very, very challenging high-speed real-time control systems. But that was dwarfed by the systems problems. I can think of many cases when we were trying to commission something, some capability, and then you'd get a really bizarre effect that you're chasing, and you might think it's software, but it might turn out to be a bearing that's stuck somewhere. Or you think it's clearly a structural issue, but it's an optical issue! [laughs] There are only a handful of engineers, I think, who had the right kind of training to have that kind of broad systems view, and Jerry Nelson was foremost among those. And he wasn't even trained as an engineer! [laughs]

ZIERLER: [laughs] Hilton, I asked you to narrate what it's like for humans to go up to the mountaintop. These are enormous instruments. These are huge domes. What were the technical challenges of getting these things all that way up to the mountain?

LEWIS: You're right, some of them are very big and heavy, so you had to cut things up into smaller components to bring it up. Weather was an enormous factor, in fact still is. From November to March, you don't know when you're going to be able to operate. And there are high winds at times. So, there are environmental factors that are extreme. But then also working at altitude and in the cold. You could design something in the lab at Caltech, for example, take it up to the summit, and it operates quite differently. Of course we know about temperature and pressure, but there's other things that are more subtle. Like electronics doesn't cool the same way, so your power supplies that are perfectly suited to operating at sea level just don't work at the summit, or they don't work reliably. I think one of the most challenging technical things we ever did was we started installing high-powered laser systems for adaptive optics. The first one we got was actually a design that was used for isotope separation, from Lawrence Livermore Lab, a dye laser. Lawrence Livermore built it for us, and their teams came out to commission it, and we had to learn how to do it as well. I remember when I called up to get a review—typically the way these things work is when you think you're done, you get a committee of experts to come in and give you their opinion so you don't fool yourself. I called up various colleagues I knew in defense who had used these kinds of lasers before, to have what we called an operational readiness review. The guy I had tapped for the committee chair laughed at me and said, "I won't serve on your committee, because there's no way a dye laser can ever be made operational." This was after Livermore had spent hundreds of millions of dollars and decades developing them [laughs]. And we had one. We didn't take that as a suitable answer, and we did actually make it operational, but it's an example of how these complicated systems can really be tough to make work at that location.

ZIERLER: This is a very complicated concept, but I wonder generally if you can explain how the engineers respond to the science objectives, for what the telescope is supposed to do. How do the scientists communicate these wishes to the engineers, and how do the engineers make it happen?

LEWIS: The approach we adopted, and I think that all big telescopes everywhere have used for decades now, is really rooted in the way that NASA had transformed itself in its early days. It's turning these loose science goals into hard technical requirements and then managing those requirements carefully. That's the language that's actually used. There's a whole discipline of systems engineering around this. The scientists will have use cases, things they want to do, and then you try and break that down to what does the equipment actually has to do, in terms of things like vibration, expansion, temperature, all these factors. Then the challenge is that as you build it, first of all you learn a lot about the things you think you can build and whether it's possible or not. And, the science evolves. So, you're constantly in flux because you can't build it overnight. It might take you seven or ten years. Of course, the technology evolves as well while all this is going on. Managing that process is the key to success or failure, and I've seen both sides. If you're too rigid, you end up with things that are obsolete on day one. In this game, obsolescence or technology prowess really determines the science you can do.

I have an example which I kept up on Keck II actually, to show my engineering colleagues—once I was part of the leadership and no longer did anything useful, in the judgment of my junior colleagues [laughs]. It was a system that was used for infrared observing, and we spent many, many millions building it, and close to seven, eight years. We used Lockheed Martin Palo Alto Research Labs, a really renowned lab, to build it. It was delivered, working, but we never, ever used it. Because it took so long that in that time infrared detector sizes got larger, and by the time we got the device, you could just read out the detector and get the same result at a fraction of the cost and more effectively. So, you have this situation where if you're inflexible about your requirements you get things built at cost but they may be obsolete on the day you get it, and if you're too flexible you never deliver anything because you're constantly chasing your tail. I found that's a really interesting part about building instrumentation for science. Because science is moving fast, and there's tons of competition, and then the underlying technology is developing all the time. That's where I think judgment is really important. And a little bit of luck, actually.

ZIERLER: You mentioned earlier that the original plan was to build the telescope and then onto your next adventure. Do you remember what it was that convinced you this was a long-term proposition, and you'd be there perhaps for the duration of your career?

LEWIS: I never intended that, I must say, but there were always interesting things to do, and that's why—it was always more interesting to stay than to find something else to do. [laughs] I think what happened is I hadn't appreciated how interesting and challenging the operational aspect of an observatory is. You can build it, and it's full of promise when you build it. It's like launching Webb Telescope. But what really happens is after that. Of course the science can't happen until you're operational, and becoming operational for a big system takes many years as well. Finding ways to expedite that, finding ways to wring out that little extra performance, is really intellectually very satisfying. Then there's the tremendous satisfaction of being part of these great discoveries. One example of that is the Nobel Prize of Andrea Ghez, for getting the experimental evidence for the black hole in the center of our Milky Way. Andrea started showing up when I had been at Keck for four or five years. She was a postdoc then. She was embarking on trying to detect the black hole in our Milky Way by looking at the orbits of the stars around it. We had an instrument that couldn't actually do that, but if we modified it, had a shot at doing some of the science she wanted. So she came to ask. I was head of software, and so I had to authorize the changes being made for that instrument. And I didn't want to do it. It was our prime instrument, and we had only just got it working, and we had hundreds of astronomers at Caltech and UC who were wanting to use it all the time, and she basically wanted us to do something that risked breaking it. Of course I asked around, and people said, "She's crazy. She's wasting your time. You totally should not follow that." But Andrea is very persuasive, and she did [laughs] eventually wear me down, and we worked together, and we assigned people, and she got this result. She committed 30 years of her career to doing it, too, on what was an extremely long shot. It was eventually successful. It's that kind of thing that just is hard to look away from. There're many other examples of incredible science, whether it was the dark energy work, or galactic evolution, all the exoplanet stuff—I mean, it just kept happening. It was like never—it was never over. The adventure was never over. It's hard to leave that.

Telescope Success Times Two

ZIERLER: I wonder from your perspective what the decision-making was like that led to the authorization for Keck II. What was going right with Keck I that convinced the powers that be to make this second telescope?

LEWIS: The main thing that went right is it actually worked. The Keck Foundation took a gigantic gamble building Keck I and there were many people in our community that were skeptical it would ever work. There was a noted optical engineer, probably the best in the world, at UC, and when he saw the initial data, he said, "They faked it." He was convinced it couldn't work, and when he saw the data he was convinced there was fraud afoot. But there wasn't. [laughs] It really did work. The fact that it worked, and it worked as well as it did, kind of took everyone by surprise. I think what was originally a gamble suddenly was seen as a sure thing. It was to cement the legacy of the Keck Foundation, I think, that they were willing to make the second gift. But we had no right to expect it. It was too big a gamble. But the fact was that it worked. And, it was built in seven years. It was very, very fast, compared to TMT, or GMT, or ELT, which have been going for over two decades, and they're still a long way from completion.

ZIERLER: What aspects of building Keck II could you simply hit "repeat," where were there lessons learned that you could apply, and what was brand new for Keck II?

LEWIS: A lot of the big mechanical structures were simply repeated, but some of those were from different vendors. A different group made the telescope structure for Keck I and Keck II. But all the mirrors, we just used the same process. Of course, by the time Keck II rolled around, we had ironed out the problems—the mirror fabrication in particular was very, very challenging in the beginning and at one time threatened to sink the whole program, because we were spending many hundreds of thousands of dollars a month, and nothing was coming off the production line. But on the software front, which was what I was responsible for, we took the chance to deliver generation two. We completely rewrote all the software. We had to write the first version because people needed it to commission everything else and to demonstrate the telescope worked, but we knew that we had taken many shortcuts, and so when Keck II came around, we said, "This is an opportunity to, quote, ‘do it right'." We've since rebuilt it two more times [laughs], so I don't know how right we got it, but we were able to take advantage of more modern technology every time.

ZIERLER: A question of some terms of art. I wonder if you can explain the difference between first light, and when the telescopes actually start doing science.

LEWIS: Even first light is controversial, because there are many, many first lights. For Keck, we defined first light as when we'd first see light through the telescope, on a detector somewhere, with nine of the 36 segments, so a quarter of the primary mirror. That was chosen because nine segments made a primary mirror was as big as Palomar. It was equivalent to the 200-inch. But even the first light, I remember when that happened, the project manager at the time, Gerry Smith, who's still alive and who I saw just a week ago, Gerry declared first light. We saw light on a detector. Of course we'd seen it with six segments too, and we'd seen it with nine segments, before we declared official first light. The project scientist, Jerry Nelson, refused to acknowledge it – the image quality wasn't good enough. So we had first light, and then we did some stuff, and then we had first light a month later again. [laughs] And so, even first light happens multiple times. But technically, it's the first time the entire system is working and you can generate—for telescopes—images of science quality. Though you're not actually doing science. Then of course there's the time you put your first instrument on the telescope after that, and you get light through that instrument. You get first light for the instrument. But when you actually start doing routine science—then, it's science. But it's a continuum, for sure. Even first light—like we recently had first light for an instrument called KCRM recently, built at Caltech—of course the real first light happened two nights prior, because the team wanted to make sure everything was working. They just kept very quiet about it. Then we had the official first light. [laughs] But science is what really matters. When it's science, not just for the commissioning team, but any scientist can apply—when you know enough about the instrument that you can design your experiment, it's reliable enough you can actually carry it out, and you can find something new—then you have something important.

ZIERLER: As you explained, Keck II happened because Keck I was working. It had proved that it was working. What had been proven at that point? That it had achieved first light? Or had Keck I already started doing science in order for Keck II to be approved?

LEWIS: I don't know when the actual approval decision was made. It might have actually been before we had started doing science. Though there had been some preliminary science. The instrumentation had been delivered—Tom Soifer and Keith Matthews at Caltech had delivered an infrared camera. Steve Vogt at Santa Cruz had delivered a high-resolution optical spectrograph. And Bev Oke and Judy Cohen of Caltech had delivered a very sensitive Cassegrain optical imager and spectrograph. Those instruments had started being used by at least the commissioning teams. Part of commissioning is that to prove that you can do things, that it's working, you have to look at the most challenging objects and show that you can do the kind of science people want to do. So, that had happened. I don't recall if we had actually scheduled the telescope for routine observing—I rather think not—when the decision was made.

ZIERLER: How did your day to day change, how did operations at Keck change, once we had two telescopes?

LEWIS: Oh, once we had two telescopes? Well, there had not been enough hours in the day when there was one, and there were far fewer [laughs] when there were two. It was very, very intense. Because the first several years of operation, we'd just throw up one problem after the other. You peel the layers of the onion. You'd work all day, you get home, grab a bite to eat, and the phone would start ringing as soon as it got dark, and with two telescopes, it rang twice as often. [laughs] It took a while to settle down. It became very intense; put it that way. But eventually it all settled down, and once you solved the problem on one telescope, you had solved it on both, and that was great.

ZIERLER: When did astronomers start to come regularly to Hawaii? When did that process really begin?

LEWIS: 1993, because yeah, 30 years ago from this year. But astronomers only were going up to the summit for about a year, and then we decided all observing would be done from headquarters. People were coming to Hawaii to use the telescope, but not going up Maunakea after that.

ZIERLER: Was that a logistical decision or were there technological developments that just made it make more sense to do it from the base?

LEWIS: Physiological because people really do not work optimally at the summit. That had been known for a long time. People had been doing astronomy at other facilities for years at Maunakea. But at that time, round about that time, it became technologically feasible to observe remotely routinely. So right from the very early days of Keck, we realized you could run remotely. You had to have enough bandwidth to get images down and data down, but that was about the time it became feasible. It wasn't without controversy. There were many people who felt they were losing something of inestimable value by not being at the telescope. But once they actually observed from headquarters, and they realized they didn't have a headache, and they didn't have the fogginess, they never went back.

ZIERLER: Of course a cartoon perspective of an astronomer working with a telescope is that they're physically—they have their eye right next to the eye piece. Obviously that's not what's going on. There's really no difference between an astronomer working at the summit versus at the base in terms of how they interface with the instruments.

LEWIS: Right. What people felt were lacking was—first of all, it's ground-based astronomy, so the atmosphere and atmospheric conditions matter a lot, for example where the clouds are. And so, of course astronomers were walking outside, and taking a look, and getting an impression of how clear the skies were, and they felt that losing that would be an enormous loss. The second part was, the operators were up at the telescope. It's often a very close interaction between the astronomer and operator because the operators understand all the vagaries of the machine, but they also need to understand what the astronomers are trying to do in order to optimize the observing. Perhaps you can't observe below a certain elevation, or you can only observe below a certain elevation for a certain amount of time, and that stuff is hard to communicate efficiently when you're in separate locations. I think it was the second one which astronomers, traditional astronomers, were most concerned about, but once they tried it, they started learning how to compensate for that. This long predated Zoom, but effectively many of the same arguments you've heard about Zoom, in the pandemic period, we were hearing back in 1993.

Pathway to the Directorship

ZIERLER: Some questions as they relate to foreshadowing to your directorship. When do you start to take on increasing administrative responsibility? When do you get a sense that you might be headed for executive leadership at Keck?

LEWIS: I always was in a leadership role. I came to start the software group, and I headed that right from the beginning. But I was having a tremendous amount of fun with just the technical side of things. It was only in about 1998—so I'd been there 12 years—when I took a sabbatical. I went to ESO, our main competitor in Munich, spent six months there. At that point I think I felt that perhaps it was time to do something different. I had been doing very much the same kind of work for a long time. My group had grown larger and larger, so management issues were coming to the forefront. I was no longer the person writing the hardest code or trying to understand it. I came back from that sabbatical with a very different perspective. Because ESO is a much bigger organization than Keck, and they were organized—it was a bigger research institution—they were organized differently. I saw the possibilities of how you could extend your grasp, you could achieve more, if you had more people working in the same direction. I also saw the challenges. Because I had been, as I said, what I'd classify as—not junior engineer exactly—but someone in the ranks, and I had run up against various frustrations. I think when I was pulled out of it, it was the first time I saw that there were ways of being organized differently that could be good for the organization and for the engineers and technical people.

When I came back, I discovered that the organization had closed up around me, that there was no actual place for me, because someone else had been filling my role and he had done a great job. I felt very spare for a while. But it turned out my bosses had seen potential for me to take on more of an executive role, so within a few months—we had had an external review, by Wayne Kennedy of UC and—Jenkins of Caltech—I can't think of his first name. They had identified a leadership need at Keck and they identified me as someone who could fulfill that need. That's when I was promoted. I became an associate director. I was responsible for larger and larger parts of the engineering effort. I had also always cultivated the science connection very carefully, partly because of Jerry Nelson who had introduced me to many of the scientists. Of course I had worked alongside many of the scientists as we were commissioning the telescope and instruments, people like Sandy Faber and many of the Caltech colleagues that you know. I had that network of people, so I could go between the two worlds. I found that's a very powerful way to advance an observatory. I'd say it really started early 2000s, and soon I was promoted to deputy director, with some partial responsibility eventually over essentially the full operation of the facility, more like a chief operating officer, and then director, at the end of that period, about 10 years ago.

ZIERLER: You mentioned Andrea Ghez's 30-year commitment when the things that she was after were hardly a foregone conclusion. Did you witness the building excitement for what she was on to? Were you aware of this in real time?

LEWIS: I think so. Actually, her particular story paralleled the development of adaptive optics. It's slightly ironic—the actual science she got the Nobel Prize for predated adaptive optics. People forget that. [laughs] But there was this huge technology development that was going on. The former director at Keck used to say it was the first major innovation in telescopes since Galileo. Because it really changed the kinds of things we can do. Now, he was very enthusiastic about AO, adaptive optics; not everyone was. So, there was this huge excitement around the technology development, and Andrea was right in the center of that. I was definitely aware that she was onto something really important, and we started seeing results early on. I didn't know it would lead to a Nobel Prize, but we knew it was important. But it's also important to put it in context, because there were other amazing discoveries going on at the same time. The dark energy discovery was one of those. Saul Perlmutter, Brian Schmidt, and Adam Riess were coming to Keck, and other folks too like Alex Filippenko, who were well very known. So, it was in the air, and we would get to talk to the scientists. They wouldn't speculate much about what they were doing, but they talked about the kinds of science that they were doing. It was a very heady time. Right from the beginning, as soon as Keck became available to the community, it just transformed astronomy. There was the amazing stuff that was done with Hubble, to identify the most distant objects in the universe that Keck played a crucial role in. There was all the pioneering work around exoplanets. Every day you turned around and there was another very exciting field of science being explored, and none of them were visible when I was studying physics and astronomy back in college, which didn't feel that far away at that time. [laughs]

ZIERLER: You mentioned the parallel development of Andrea's work and adaptive optics. Is that to say that even conceptually AO was not baked into the origin plan for Keck I and Keck II?

LEWIS: Absolutely it was not. It was not. AO really came to the fore because of defense conversion. When the Cold War ended and the Berlin Wall came down, that's when a lot of these technologies became available to science and astronomy. There were certainly people who were aware of AO possibilities, but it really burst on the scene in the 1990s and came to fruition in the early 2000s.

ZIERLER: What is the value of AO in a national security framework?

LEWIS: Astronomers use it to reduce turbulence, so you can see things very sharply when you're looking up, but the atmosphere works the same both ways, so you can look down through the atmosphere, and if you can remove the turbulence, you can see things incredibly sharply. The other part that defense was very interested in is if you want to shoot a laser up to destroy a spacecraft, the atmosphere will blur it out. So you want to pre-distort things so that your particle beam or your laser emerges from the top of the atmosphere perfectly collimated and can do whatever nasty work you want it to do. Then of course the other purpose the military used it for was to observe things through the atmosphere—a spacecraft going overhead, for example.

ZIERLER: This is a technology, this is a program, that's essentially classified, that astronomers are not aware of these capabilities?

LEWIS: It all started in the 1950s. There was a guy called Kolmogorov who did a lot of the early work that was needed. Astronomers were always aware of it, but the technology and the costs of it were just far beyond astronomy as a field. The defense force spent vast sums of money on the underlying technologies, everything from the lasers you'd need, to the deformable mirrors, to the computing capabilities and detectors. Astronomy was the beneficiary of all of that, once it became declassified.

ZIERLER: Was there a worldwide race among different telescopes to adopt AO, or was Keck at the forefront? What did that look like in real time?

LEWIS: There were some smaller telescopes that had been doing work ahead of us and had fielded systems before us, but it gets progressively more difficult to field it on larger telescopes for technical reasons. Of the large telescopes, Keck was definitely the first to the game. Equally important, which goes back to an earlier time, is we were the first to make it an operational capability. It's one thing to field it, to demonstrate it works on a particular night for five minutes. It's another thing to integrate it into routine science, night after night, month after month, for 10 hours at a time. At Keck we did exceptionally well on that. Since then, other large telescopes have done just as well. The VLT, in particular, is fielding more sophisticated systems than we have. But, it's an arms race; there's always someone ahead for a while. But Keck was definitely the first of the large telescopes to get there—and the first to use it for significant scientific output.

ZIERLER: When you were named deputy director, what new view did that give you, both of the Observatory and the policy decisions, whether in Washington or UC or Caltech, for how all of these things came together?

LEWIS: I definitely had not appreciated the role of long-term policy in the success of the Observatory. I had always seen it as a series of projects. We'd have funding for an instrument or funding for some program, but I hadn't really understood how you have to put in place the enablers for those kind of projects. For example, the interferometer, the Keck interferometer, that was really an effort by folks to deliberately increase the NASA involvement at Keck, because people at Caltech and UC understood that funding the observatory for the long term meant we had to have another major funding partner. That level of policy or policy planning had been pretty much invisible to me as a lower level manager. It was also the first time I was exposed to our board in any significant way. I had met and knew many of the individual members, but hadn't interacted with the governance of the organization. It's a part that's largely invisible to most staff at Keck, that level of decision-making, and I dare say to most of the scientists at Caltech and UC as well, but it's critical for the long-term future. Short term, one understands that you have a budget, you have resources, you know what's coming down the pipe, but how to make sure that continues decade after decade, that's where I got my trial by fire as a deputy director.

ZIERLER: Was it as deputy director that your administrative responsibilities really overtook your engineering responsibilities, or did you always retain a hand in engineering questions and work?

LEWIS: I think as deputy I still maintained a fair degree of involvement in the technical side, mostly through reviews that I chaired or set up. It's really when I became director that that was no longer possible. I was very much involved with the operational decisions as deputy. That was by design, as well. The director at the time, my boss, was an astronomer, and so he focused on astronomy policy issues and fundraising, and I focused on keeping the place running on a day-to-day basis. It was a transition. You start off with your feet firmly planted in the technical and gradually you have less time, and there's other questions that are higher priority that you have to get involved with. For example, questions around safety or risk management, those are things that started taking more of my attention. I wouldn't call those hard technical issues in the way an optical design would be, but of course they have a great deal of bearing on the kinds of things you can do. But it was really as director that I effectively had to leave behind the engineering aspects in almost all areas.

ZIERLER: The protests in 2014 at the base of the mountain, for the beginning of what was planned to be the construction of the TMT, did you witness that? What were your thoughts as this was unfolding?

LEWIS: I did witness it. I had two competing thoughts about it. One was for the safety of our own staff and the well-being of our own staff. These are not just things you witness on TV. When you actually go through it, and you're in a hostile crowd, and people are yelling at you, or you go to a hearing and people are very combative, that's tough on individuals. I saw it as my role to ensure that our staff were not unnecessarily exposed to that, that at least they had the support that we could offer them, and that they knew it. So, I had two thoughts. One was that, "How is it that we're here?" I'll tell you that by 2005, 2006, we were really pretty uneasy. It felt to me like we'd lost a decade that we could have dealt with a lot of this. Even at that point it was very hard for us to get through to the decisionmakers around the TMT about the Hawaii situation. I think it was viewed simplistically—but Hawaii is unique. To try and understand what's going on in Hawaii from outside Hawaii is very difficult, I think. Even inside Hawaii it's very difficult to understand it. But it seemed to me that we were getting wrongfooted a lot. Our own actions weren't optimal, and we didn't understand the situation properly. We were slow to react, to everything. Whereas the folks who were opposed to the TMT were highly organized. We underestimated their organizational abilities for sure.

ZIERLER: Did it impede Keck operations directly?

LEWIS: It did, not for an extended period, but we definitely lost some time, and it increased the risks that we faced. It had a bigger impact on our staff. I would say back in 2015, there was a lot of concern amongst the staff, and many people started thinking there was no real way forward. People didn't like being accosted for wearing their Keck t-shirt to the store. There was definitely—I wouldn't go so far as to call it a siege mentality, but definitely people felt that their acceptance in the community was under threat. And we lost people over that. That's why I said by this year, there's a different feeling. We have really made some very big efforts to try and understand where those opposed to the telescopes were coming from, and to try and meet some of their concerns. But that wasn't the case so much back in 2015 and 2016.

ZIERLER: What is the selection process like for the director? Who makes the decision? Who are you talking to in this process?

LEWIS: It's not quite as opaque as selecting the Pope, but—

ZIERLER: [laughs]

LEWIS: —the process is there's a selection committee, and they are given some broad guidance about what the board is looking for in a director based on previous job descriptions and the environment at the time. Then they end up selecting a slate of candidates and winnowing it down to three or four. Then those candidates are sent to the board of Keck to select one. That's the process. They're well along in that process at this point, but they haven't yet made their selection.

ZIERLER: What did you articulate as your leadership role or your leadership vision for Keck when you were named?

LEWIS: It's important to understand that I was a non-traditional director, because traditionally directors are astronomers.

ZIERLER: You were the first engineer director?

LEWIS: Right, certainly of a large telescope. My vision was partially the continuation of the enormous success that the Observatory had had. I wanted to ensure that we could continue that for the decades ahead, so it wasn't a one-trick pony. The elements that I articulated around that were that the science mission was central. We weren't going to move away from that. And the way to do that was to empower our staff in Hawaii to properly support that mission. One of the issues that I had recognized we had early on was we were quite siloed. The different areas and divisions all worked for their own interests. What I really wanted to do was change that, so that at least at the leadership level, the first interest everyone had was the Observatory, not their department or their section. I actually made quite a few changes in the leadership, and in our way of decision-making, so that the top level, their first allegiance was to making sure the science mission was met. The second part I had was to start transforming, for the advent of the TMT and the ELT. Back when I was hired as director, we were supposedly only a few years away from the TMT coming online, so I saw that as an urgent need. As it happened, I was at least 15 years [laughs] too soon. But I saw that as likely to—if we didn't immediately act on it, that that would significantly diminish what we could do in time.

ZIERLER: What was the budgetary situation that you inherited? Was Keck on stable ground at that point?

LEWIS: Yeah. It always has been. The operations budget is well guaranteed. It's the budget for building new things that you have to go out and beat the bushes for. Part of doing that is through philanthropy. Part of that is these long-term policy decisions around either scientific collaborations or supporting NASA or the NSF interests. It's the latter, funding for new things, that I really tried to focus on improving, because I knew the former was funded by Caltech, UC, and NASA, and it's long-term stable. As long as we keep doing great science, that's not going to be called into question. But the new development was very challenging, with TMT being close-in, because we knew that the federal dollars are tight, and philanthropy dollars, for that matter, tend to flow to wherever the biggest science can be done, so I did see that as pretty urgent, to stabilize. As it happened, as the TMT receded, two things occurred. One is that the competition for dollars for building new things at Keck became a little easier. The other was the competition for intellectual talent became easier too. Because many of the scientists at Caltech and UC who are critical to building new instruments had started realizing that they probably were not going to ever get to use TMT in their working life, and they recommitted to becoming principal investigators for Keck. So we had a sort of sudden flourishing of talent again, which was great but unanticipated.

ZIERLER: Hilton, one project that actually came to fruition, of course, that we talked about earlier, was the James Webb. As director, did you have some role in that, or were you following closely these developments and how they might impact Keck operations?

LEWIS: Yeah. We had a tiny role, in that some of the algorithms that are used for the Webb primary mirror were tested at Keck, were developed at Keck. That's because some of the former Keck people went on to places like Ball Aerospace and ended up working on Webb. There's this fraternity of folks in the community, all connected at some level. So, there was an actual technical contribution, but it was small. I might say in passing that one of the technologies that was developed for one of the Webb instruments—a technology that Webb subsequently dropped—became the cornerstone of one of our instruments at Keck built by UCLA and Caltech. So, we benefited too from the work and investments that Webb did. The main thing that we were paying attention to is where Webb would basically completely dominate areas where we had been scientifically competitive before. In the mid infrared, some of the work we had done, it's no longer viable for us to try and compete with Webb on that front. So, we had to look at how we could continue to be complementary to Webb. So far, our instrument suite has enabled us to do that.

Remembering Jerry Nelson

ZIERLER: When Jerry Nelson passed away in June of 2017, I wonder what that meant for you personally, and what reflections there were at Keck on his legacy.

LEWIS: I personally was heartbroken, because I told you, he was the foremost of all my mentors, and I regarded him as a friend. The day before he passed away, he had been emailing me. He was still advising on a project that was delivered to Keck about six months after his death, so he and I had been communicating about that. It was poignant to have probably his very last email, about the telescope he loved, because he had built it, he had made it happen. Even at that time, there were far fewer people left at Keck who had interacted directly with Jerry, but of those of us who had, I think we all felt a huge loss. Jerry had been less involved in Keck over the years. He had ramped up his interest in CELT, the precursor to the TMT, and we didn't see him as often. I'd see him when I'd go out to UC Santa Cruz. Then he had a stroke, so he was significantly impaired physically—not mentally—by that. It was always hard to see someone who had been so vibrant and was such an extrovert, reduced by a physical ailment. But intellectually he was fully present—we used to communicate by email because speaking was so hard for him. So, it was very sad. You come across people like that very rarely in your career, one's career, I think. There are a few giants. There are many great people, but there are very few that tower above everyone else, and he was definitely a towering intellect. A renaissance man, for astronomy, for sure. [laughs]

ZIERLER: Your emphasis—science first—as director, what were some of the instruments that helped you achieve that vision, to maintain the primacy of science at Keck?

LEWIS: The original instrument suite was fantastic. There was LRIS, NIRC, and HIRES. LRIS was the machine that basically did the dark energy work. HIRES is the one that did all the exoplanet work. NIRC was our earliest one, and it was involved in a wide range of discovery. But we have always been blessed by incredible instrument builders at Caltech and at the UC campuses, UCLA and Santa Cruz principally. It has always struck me that it's not the individuals so much; it's the history of instrument building at those institutions that is the secret sauce here. Because people work in this, building instruments, for decades. Of course they raise new generations of postdocs and engineers who come through the process, and then those people become the PIs of the next generation. So, you feel like you're part of this continuum of instrument-building that has been going on for a hundred years, and it really is extraordinary. I know I periodically would get people in from outside the community—engineers, or program managers, who worked in industry—and they would be astounded at what they saw as the incredible inefficiency of how instruments are built at the universities. They were right, but the end product is still something extraordinary. It's this combination of a long history, a deep understanding of materials and properties of materials and optics, coupled with this immense dedication. When I think about instruments like DEIMOS that we have, and the PI, principal investigator, Sandy Faber, she must have given up eight or ten years of her life to that instrument. And she was not unique. This is kind of true of all the PIs we have had. It's an immense commitment. They probably do it initially because they want to use it themselves, but pretty soon, it becomes apparent that it is a real service to the community. And it's because we have those people that we have these incredible instruments that come out.

ZIERLER: Tell me about your commitment as director to public engagement, public outreach. How would you engage the public? What audiences did you want to reach? What were the most important themes for you to touch on?

LEWIS: Well, I touched on right at the beginning how one patient engineer or scientist changed my life, and I was always aware of that, so I always felt that it was important for us to go out into the schools. I really wanted to make sure all of our public lectures were a once-a-month lecture series, where we engaged members of the public, I always thought those were really important to keep going. But more than that, I wanted to welcome people into Keck. Because the example for me was actually JPL up in La Cañada, and how when they have their open house, how the community floods in and really regards JPL as their research lab. I always wanted that for Keck in Hawaii. So, we'd hold open houses. We'd get the public in. Not on the scale that a JPL could do, obviously, but it seemed to me really important that people feel that this was their research facility, not just for the scientists who used it. By the way, I do think one of the amazing facts about Keck is that the astronomers who use Keck feel it's their telescope; we're not a research facility that they just use. After nights observing, people would walk into my office and either praise the staff, or they'd criticize something, but they were always super engaged. Astronomers never felt that there's Keck, and that they get time on it as a casual user. That's actually one of the super strengths, I think, of a model like the Keck Observatory. As opposed to, if you want to use James Webb, you submit your proposal, someone else carries out the observation, and in time you get your data back. It's not the same thing.

ZIERLER: There's a visceral closeness to land-based astronomy, you're saying?

LEWIS: Yeah. And I would get a call at 2:00 in the morning from someone who was upset that the operator had closed the dome, and why wouldn't I just overrule them? [laughs] I think that's really great. The sense of the community ownership I always felt was really important, but I feel even more now. Because I think that is how to protect and advance astronomy, in Hawaii.

ZIERLER: Moving the conversation closer to the present, of course when COVID hit, what were your immediate concerns? How did you keep the staff safe? How did you keep the science going?

LEWIS: We actually acted very quickly. We shut down before the state shut everything down, which in COVID terms was only a week, but that was a huge amount of time. I remember when we shut down, there were a lot of people who said we were crazy. The concept of exponential change hadn't quite hit any of us.

ZIERLER: Right!

LEWIS: I'm sure you remember how—

ZIERLER: All too well.

LEWIS: —literally, day by day [laughs], everything changed. We were, and I was absolutely clear, we were not going to risk people's health and lives, so we totally shut down. Then we regrouped to figure out how we would get back online. Actually we were able to reopen something like six weeks later. We lost something like six weeks. We had emergency crews go up to keep equipment safe, but I think it was after six weeks, we were fully back in operation. We just prioritized the science observing. Normally, there's a balance between observing, and doing new things, and maintenance, but early on I decided that what we really wanted to do was maintain our scientific productivity, and we did. At a cost. It meant all our future projects slipped, and we built a backlog of maintenance. Eventually we had to deal with all of that. But it did mean that for the initial COVID period, when none of us knew how things were going to go, Keck was actually back in production. We did develop something we called pajama mode observing, which was a mode that allowed you to sit in your bed with your laptop and run the mighty Keck Observatory. We knew that was a dangerous thing, because we knew that people would fall in love with it. And they did, and they do [laughs], and they still love it. And actually it has reduced the number of people that come to the Observatory, which is a bad thing, because it reduces the sense of ownership, interaction between the staff and the astronomers. And it's a good thing because it lets more people participate, especially in an era where budgets for travel are really tight.

ZIERLER: When Andrea Ghez won the Nobel Prize, did you, did Keck, feel included in the celebrations, given her longstanding partnership with Keck?

LEWIS: Yeah, we absolutely did. Andrea is one of the astronomers who are very close to a lot of the staff, and she repeatedly came to Hawaii during the observing time, but even later, when she had a big team of people. She has always been super connected with Hawaii, not just the Observatory but the place, and all the staff. She brings her team out once a year for a weeklong retreat. She is closely involved with the adaptive optics. She is just one of those people who have sought those connections, where the people are paramount. So we all felt very proud. In fact, when she did win the Prize, we even went so far as to print up a T-shirt for that, which staff liked wearing. Which they didn't for the dark energy discovery. [laughs]

ZIERLER: [laughs] To return right to the beginning of our conversation, you talked about your decision-making, the timing of the retirement at a time of stability. Was the stabilization of COVID also a part of it? Had things sort of felt back to normal, and did that influence the timing of your decision at all?

LEWIS: Yeah, I think there were really three factors, maybe four. One was the Hawaii situation, the Maunakea situation. The second one, we had just delivered our strategic plan for the next decade. Then, I finally felt that we had resolved COVID. There are still post-COVID shocks, but for the most part, we had come to an understanding of how we were going to be working going forward. We had, like many companies, come to some kind of decision around how much time people were going to be in the office, or remote working, all that kind of stuff. And then, the funding was in a very stable state. But I don't kid myself for a moment; these things are never stable for long. There's always—one of them will suddenly slip. [laughs]

Discovery Thanks to Adaptive Optics and Segmented Mirrors

ZIERLER: Now that we've worked right up to the present in our conversation, for the last part of our talk, if I may, I'd like to ask a few retrospective questions about your career, and then we'll end looking to the future. One is, as an engineer, but one that achieved the leadership position at Keck, do we see engineering insights or engineering innovations that have an afterlife beyond Keck or even beyond astronomy that were originated at Keck?

LEWIS: I think the most obvious one is around adaptive optics. Of course, we didn't invent adaptive optics, but we certainly advanced the state of the art significantly. Adaptive optics also has use in other areas. One of the really interesting ones is in vision science. If you want to study a retina, you can only really do it when the person is dead, [laughs] unless you use adaptive optics, because there are all these fluids in the eye. So, adaptive optics is used for vision science and for diagnostic purposes, so I think that's a spinoff that potentially will have an impact on millions and millions of people. We've had our little part of advancing that science. I think in the other areas, they tend to be pretty esoteric. Segmented mirrors are definitely the way of all future large telescopes, of course adopted by the Thirty- and Thirty-Nine Meter, James Webb, and the next generation telescopes in space. I think that's a technology that was not only developed for Keck and by Keck, but it was really put on the map by Keck. Those are the ones I see with immediate impact. There's also lots of technological spinoffs, like the way that mirrors were shaped for Keck, it was a technique called stress mirror polishing. Basically you bend the mirror into a particular shape, you polish an easy curve on it, and then it relaxes into the shape you want, the complex shape you want. Those are some technological processes that have an impact. I think the instruments themselves, there's lots of work on things like coatings that can affect communications technologies, and how fibers are used. But those are not developed at Keck, specifically, but in astronomy, whether it's Caltech, UC or JPL or some other institution somewhere. It's a steady stream of technology progress that comes out of observatories, but Keck itself I think will be remembered for AO and segmented mirrors.

ZIERLER: What you're saying, it sounds like, is even though astronomy is a purely fundamental discipline, so many technologies that come out of astronomy really do have benefit to society in a wide range of applications.

LEWIS: Right, absolutely. But I think it's wrong to focus just on the technology development, because the real story of astronomy is the advance of science, physics. And with its 100-year lead times. You may have heard about GPS systems and how they connected with the predictions of general relativity. Of course, there's the world-changing and life-changing one of the discovery of life elsewhere in the universe; it's hard to imagine that will not have an impact. Although, given the political discourse today, maybe it won't have any impact at all. [laughs]

ZIERLER: That's right. [laughs] You preceded me on my next question, which gets to the science. For all of the discovery that you've been a part of, that you've helped to make possible, what has been most exciting to you? What has given you most personal satisfaction?

LEWIS: I'd like to say there's one thing, but I'd have to say it's the general progress. I alluded to this earlier—that when I was studying general relatively, there was not really strong experimental evidence for black holes. There was Cygnus X-1, we knew about, but it was really the work of the last couple of decades that—oh!—that culminated in an actual image of a black hole, which is now possible. There's another example—my favorite image, that I used to keep at Keck, of a system called HR 8799, which is a star with planets around it. Bruce Macintosh did a lot of that work. He's now the UC Observatories director. That was a discovery that, until it was made, we just knew could not be done from Earth. It was impossible. And yet, one day, there it was. It's that sense of breaking through barriers that just could not be broken through. You could not make an image of a black hole. You could not image planets around a star. And yet, all of those things happened. And then, there was everything around exoplanets. Going from a field where to be studying exoplanets was a career-ending move to being probably the most dynamic part of astronomy today is extraordinary! Like, how do these things come out of nowhere and suddenly dominate everything? Of course exoplanets, or the discovery of exoplanets, are our key to finding life elsewhere, whether intelligent or unintelligent. So, what was really total speculation back in the early 1990s is absolutely hard science today. That's what is key for my sense of wonder about astronomy. So when you ask me, "What's the one thing?" I would say the one thing is this incredible, unending progress we've had, over the last few decades. I used to joke that it was a golden year for astronomy, but the golden year implies it's a short time; it seems to have gone on for a long, long time. And it keeps going.

Exoplanet Excitement

ZIERLER: Finally, Hilton, some questions looking to the future. You mentioned exoplanets. Do you see a role for Keck in that foundational discovery, if it happens, when it happens?

LEWIS: Oh, yes. Definitely. Interestingly enough, some of the most important progress is coming from much smaller telescopes. Because—this comes back to a point I made earlier—you can divert a small telescope to a single project; you cannot take your most precious facility, resource, and divert it to only one thing. So, what will happen is when the next generation of telescopes come up, then we can devote Keck to much more of those niche explorations. We could take one of the Kecks and make it purely an exoplanet machine, or a spectroscopy machine, which would make no sense today because you'd be cutting off too much other science. I do see a role for Keck in that future. I also see that we've been developing a suite of exoplanet instruments, very specific ones, and they've been coming online the last few years. There's a new one that Dimitri Mawet at Caltech is building called HISPEC. There's one at Santa Cruz called SCALES being built by Andy Skemer. We just got the Keck Planet Finder, which was led by Andrew Howard. These are all instruments I helped make possible at Keck, also by helping to find money for them. I'm excited to see these instruments come online. I think the next decade, before the TMT and the ELT really get going, we'll see huge progress in this area of exoplanet discovery, and Keck will be right at the forefront of that story.

ZIERLER: As Keck is currently in an interim leadership position, whoever the next leader is, what are the most important things for him or her to consider as they think about their tenure as leader of Keck?

LEWIS: One thing is—there's no question—the whole issue around astronomy in Hawaii has to be front and center, because that's existential, as overworked a term as that is. Because no lease, no telescope; that's all there is to it. So they will need to understand both the—it goes much deeper than concerns—the anger, the hurt, the long history, and what to do about it. I think sensitivity to issues around Maunakea and Hawaii are really critical. The second thing I think is, when I grew up in Keck, we just had to be there, to be successful. We were the biggest kid on the block by a huge amount. Eventually, other formidable competitors grew up. Back then, whatever we did, we'd win. That era is long, long past. And so, the next director will have to fight that natural inclination to complacency. Which I actually saw as important for me, too. It's hard, when you're still enormously productive, to say, "This doesn't last forever. We've got to do things today that will actually impact our productivity to stay relevant in the future." I think that's very important. Exactly how to navigate that, where science goes, what the different capabilities are, how to build the partnerships, whether it's with ALMA, or Rubin, or JWST, I think that's really very important to get right. Then the last part is I think you have to stay sensitive to the roots of what Keck's success is, and it really has been this incredible collaboration between Caltech and UC. NASA has been a critically important part of it, too, but at root, it was the owner-partners, who came together to build Keck, and it has been such a productive relationship that it's worth stewarding for the future. That's something which means you have to figure out how to balance the competing needs of those two great universities, and you have to keep NASA—even as a one-sixth partner you have to keep NASA in the picture, because it's a very strategic part.

ZIERLER: Hilton, the fire, the horrific fire that unfolded in Maui a few weeks ago, it's one island; obviously the Big Island is not immune to these climatic changes either. Are you concerned? Is there a strategic response for Keck to consider in this new climate reality?

LEWIS: We had some very big fires, an enormous fire, two years ago, on the Big Island, where Keck is based. Our response to that was, "Community first." Our staff opened their hearts and their homes to people who had been displaced. I'm sure that is happening right now with the Lahaina fires. I personally have inquired of friends I have in Maui to offer them a place to stay if they need that. I think Keck, as one of the bigger employers on the island, has a responsibility to help when these tragedies strike. We don't have a direct role in the emergency responses, but what we can do is our staff can certainly be engaged with the community decisions around this issue, around development, around what kinds of development go forward, and what strategies people use for where they can build and how they can build. I think it's important for us to be an active part of the community. Long term, it's all of our responsibility to do what we can about climate change. But this is a pretty stark reminder of how serious the problem is, and how serious the challenges are.

ZIERLER: Finally, Hilton, for you personally, I know you're on a pause in your career right now, and you're not sure what comes next. If you want to get involved anew in astronomy, what might that look like? Where would your services be best put to use?

LEWIS: Well, it's time for someone else to lead Keck. Ten years is a long time, and if you take it seriously, it's very intense. [laughs] It's time for new ideas to emerge, and so I'm very happy for someone else to lead Keck. I mean, I love the place, I feel like my DNA is in it, and I really have committed the vast part of my career to this one institution. But there are other things I would love to do is—I have more than a passing interest in LIGO. I'm on one of their review committees. I'd love to help enable the next generation telescope that are being conceived of now, like Cosmic Explorer, the successor to LIGO. I'm also very interested in ways that we could bring a more entrepreneurial approach to instrument building. The beating heart of an observatory is the instruments. You build the telescope once, but you keep replacing the instruments. Now they take us decades to build, from concept to delivery, and they cost us tens of millions of dollars, anything from $25 million to $100 million nowadays. It seems like an incredibly inefficient process. I'd love to find ways to work with groups on building instruments much more effectively, much more quickly. We have very limited resources, and I'd love to use them in a better way. So, something along those lines. I've always enjoyed working with my counterparts at Caltech and UC and NASA. If there's some way I can help bring those teams together, that would be a lot of fun. What I don't particularly want to do is I don't really want to be hiring and firing people anymore, and mediating disputes amongst people. That's—a chance for someone else! [laughs]

ZIERLER: Only the fun stuff, from here on out, for you.

LEWIS: Yes. I think so. [laughs]

ZIERLER: Hilton, this has been a wonderful conversation. I want to thank you so much for doing this. I really appreciate it.

[END]