Tony Uttley - President, Honeywell Quantum Solutions

(Joel Beasley at 00:00:00) Hello, my friends. Today we are talking to Tony, the president of Honeywell Quantum Solutions, and we discuss what it took to build the world's fastest quantum computer, the present, near-term, and future of quantum, and how industries will use this technology in the future to solve problems they never even imagined. All of this right here, right now on the Modern CTO Podcast. Here we go. This is the Modern CTO Podcast.

(Joel Beasley at 00:00:37) It's Tony.

(Tony Uttley at 00:00:40) Good afternoon.

(Joel Beasley at 00:00:41) What's going on, buddy?

(Tony Uttley at 00:00:43) I'm doing well. How are you doing?

(Joel Beasley at 00:00:45) I know what that is on your background image.

(Tony Uttley at 00:00:48) I know, right?

(Joel Beasley at 00:00:50) Yeah, that is amazing. I've been brushing up on my linear algebra.

(Tony Uttley at 00:00:56) All right.

(Joel Beasley at 00:00:57) Yeah, I really wanted to try to wrap my mind around quantum computing, and so I put on my wish list for my team. I said, here are some people that I really want to talk to. I want to understand quantum computing. I want to understand the business side of it. I just want—it's the future. You know, I see these machines and it reminds me so much of the computer filling a 20-by-20-foot room, and it's like, this is it. We're watching it happen again.

(Tony Uttley at 00:01:24) Exactly. It's happening again the same way. It's hugely fun to be a part of it, I've got to tell you.

(Joel Beasley at 00:01:31) So we're just going to hang out and talk. We're just going to record the whole time. We can edit up what we like, what we don't like, and just have a good time talking about life, talk about quantum computing. Now, how long have you been into quantum computing?

(Tony Uttley at 00:01:46) I have been a part of this for almost ten years. I started my career at NASA, so I was at the Johnson Space Center for about a decade. I left that and went into consulting. I worked for the Boston Consulting Group for quite a long time, and then came into Honeywell—boy, almost ten years to the day. And when I came in, I led marketing and strategy for a very large business unit of Honeywell's, and I had a team of scientists that were reporting up into me. And I thought that was kind of interesting and tried to learn a little bit more about what they were doing, and had a chance to spend time down in the labs with people who were working on different kinds of technologies like our cryogenic systems and precision magnetic control, and they were doing lasers and photonics work. And these two scientists came up and said, "Hey, do you know what we could do if we put all these things together?" And I said, "No, no idea." And they said, "We could build a quantum computer." And so that started rolling—you know, a snowball that over the last decade has gone from some little milestones that we set for ourselves all the way up to now releasing the world's highest-performing quantum computer. So it has been incredible.

(Joel Beasley at 00:03:18) So all the Googles and the IBMs, all the big Tech Five are releasing their quantum stuff, and then out of nowhere, Honeywell is just like, "What up?"

(Tony Uttley at 00:03:27) It's like, "Hey, welcome, everybody." Yes. So great companies to be in kind of collaboration with and certainly in the ecosystem of trying to move quantum computing forward. You said it right, which is we're, one, at the beginning, and two, these systems are taking up a room again. And when everybody says, "Well, how fast are you going to shrink it down so it fits on a watch?" Not fast. These are going to get way bigger before they get smaller.

(Joel Beasley at 00:04:02) Maybe that's the style. Maybe these types of computational machines happen to live in these larger facilities in a cloud-style way.

(Tony Uttley at 00:04:13) And certainly, that's going to be the most appropriate way to utilize them. Yeah, it'll be from a consumer standpoint—it'll be just like you do anything. You'll be on your phone, you'll send in a request for something. It might be that you're searching, and you won't know that in the background, your search, when you press that button, it got split into an algorithm that had a hundred different components, and of the algorithm, two of those components got shuttled over to a quantum computer because that quantum computer could do those so much more efficiently, and it got recombined with a classical computer, and all you saw was the answer that came back up. And so you may never even know that it was quantum resources that were behind that job. But certainly, in this kind of first expanding era of quantum computing, that's how we expect the interaction to go.

(Joel Beasley at 00:05:06) Yeah, I actually made this little ten-minute "follow the tutorial" to do a quantum roll of the dice. I think it was just through this API called Forest or something. I was just trying to understand the landscape. I spent about twenty hours really deep in quantum computing trying to understand it because I really wanted to figure this out. And here's a couple places where I'm hung up. Maybe you can help me. Okay, so I was on the quest to find out what it's specifically useful for today, that today I could actually use a quantum computer to actually solve this specific problem. And I wasn't able—I saw people, they wrote abstractly, "Oh, it's examples. It's going to be good for GPS maps or, you know, being able to figure out the best distance between locations." I could—or maybe searching a database through Grover's algorithm or something. And I was like, I'm an engineer. I've been writing code for seventeen years. Show it to me. Let me see it happen. Give me one playground-style example where I can actually see something happen, and I couldn't get there. Is it not there? How do we get there?

(Tony Uttley at 00:06:22) It's there, but I think you have to, in addition to rewinding the time period and saying, "Okay, we're back where computers take up the size of a room again," you have to go and say, when they did, how did it work? And what you'd find is that you had people who, potentially with a piece of paper and a pencil, could go faster to compute than the computer could. We're back in that era again, which is—and I think the way I describe the entire quantum computing set of timelines are around three eras. So the era that we're in right now is this emergent era, and it is a situation where quantum computers didn't exist and now they exist. That's a pretty profound moment, right? So what do you do in an emergent era? You can take these quantum algorithms that have been contemplated, you can run them on real quantum computers. Now, does that mean that you can't also run that same algorithm on a classical computer? No, you can. Will you come up with the right answer on a classical computer? No, you will. But that's actually very important right now because—I'm going to jump forward and I'll come back. So in these three eras, you go from emergent era to—and I'm going to skip one—go to the third era, which is classically impossible. So that's the one that gets all the press attention. That is where you could run the highest-performing supercomputer in the world for a million years, and it wouldn't come up with the answer. That is still some distance away. There's a middle era, which is much more important, I think, which is this classically impractical era. So what is classically impractical? It means that you could run this computation on a classical computer, but you won't, and you don't today. And the reason you don't today is either it takes too long, it costs too much, or it consumes so much power to do that that you have decided just not to do that computation. Crossing that boundary into classically impractical—other people might call it going to quantum advantage—we believe is going to happen much sooner. And when you cross that boundary, so when you go from emergent era into classically impractical or into this quantum advantage area, you have now done something where if you haven't already built up the trust in your system that it's going to come up with the right answer every time, you have now pushed into a period where you can't fact-check that computer anymore. There isn't something else that you're going to be able to run the computation on and go, "Yep, okay, we did get the right answer." So this emergent time is actually very, very important, which is you take being able to find binding energies of a molecule and just use a really small molecule and demonstrate that you can do it, and you can do it over and over again repeatedly well. You take a search algorithm and you train it down to a very, very small number of variables that you're searching, or you take a machine learning example and, you know, something, again, that can be easily done on classical computers, but you are proving that you can also do it reliably and accurately using the quantum computer. And that is the time period that we're in right now, and it's important. It's just absolutely critical for the long-term success of quantum computing.

(Joel Beasley at 00:10:15) Yeah, you know, when I was trying to understand this, I actually went back and read history of computers because one of my thoughts when I had that gap, that gray area, I said, "Well, you know, I sit so high in the stack as a software engineer. There's so much that happens below that I'm unfamiliar with on a detailed level." And so I said, "I bet you some of this stuff that's going on now is really, really low-level programming before they had programming languages where you could communicate in English with the computer." And I said, "I think that's probably where we are now. That's why I'm having a harder time wrapping my head around it."

(Tony Uttley at 00:10:55) You're absolutely right. You know, so we are in the machine code language for quantum computers, right? And so let me—when I use the trapped ion as an example, our quantum bits, our qubits, are individual atoms of ytterbium, individual atoms that we have imparted quantum information on. And when we have multiple qubits that we're having do quantum operations, so getting entangled, we're physically moving two atoms together in the same potential well, hitting them with lasers at 90 degrees and entangling them. That is what's going on in the middle of—after you've sent your job, after you've sent your quantum algorithm to our system, we're manipulating atoms, shooting them with lasers, and reading out whether or not they're giving out photons, whether or not they're light or dark. And that is what is allowing us to say, "Here are the results back." And so the operations to make sure that you can scale this successfully over time, to make sure that you're able to get the correct answer on these smaller-scale problems, that is all of the—I was going to say the magic, but the intense amount of technical development that has had to go on to be able to get to that point.

(Joel Beasley at 00:12:15) And so this method's a little bit different than some of the other companies, right?

(Tony Uttley at 00:12:19) It is, it is.

(Joel Beasley at 00:12:20) So what's special about the trapped ion method?

(Tony Uttley at 00:12:24) So it starts with that idea that we're using an individual atom—again, of, in our case, ytterbium. And the reason that that's important is that every atom of ytterbium is identical. So Mother Nature did us a huge favor and said, "Our qubits start perfect." And when you have a qubit that starts perfect, then any error that's being created is because of something we did. As we built our trap, as we built the physics package—so this chamber that is surrounding that, which is kind of a basketball-size, ultra-high-vacuum stainless steel chamber that we then pump out particles so that we're something like five times, six times less particles than in outer space. We cryogenically cool that to somewhere between four and ten Kelvin, right? So between four and ten degrees of absolute zero. And we do all of that to be able to maintain these atoms in a position where we can put quantum information on it and then have that quantum information be retained for, you know, in a relative sense, for a very long period of time. So they call it coherence.

(Joel Beasley at 00:13:44) Yes, yes. But there's so much good stuff. I'm just going to interrupt you so I don't lose my—okay. So the actual process of writing information to the atom, can you explain that? Is that the laser? Is the laser writing? How does the atom get written to?

(Tony Uttley at 00:14:02) That's exactly right. So I'll start at the beginning. This trap, the reason it's called the trap, is that we create an electric field, electromagnetic field. And as we heat up a substrate of ytterbium, it atomizes. So it's basically starting to give off atoms. And we have a very small hole—you probably won't be able to see in there, but we have a very small hole. And as those atoms come in, a laser is just set already right over the hole to photoionize it. So the first thing we do when the atom comes through is strip off an electron. When we strip off an electron, it now is a charged particle. It is an ion. And because it has a charge, now I can control it with this electromagnetic field. And so think about a surfer. We make a wave, and that surfer just rides that wave to a part of our trap that we have already pre-designed to be able to do our quantum operations. So on different parts of this trap, we have lasers already preset. They're waiting for that ion to get there, and we hit it with a laser that imparts quantum information on it. It's called initialization. And when you initialize it, you can initialize it either as a zero or as a one. It is either light or it's dark. And then as soon as you have initialized that, you—again, you may have initialized multiple qubits in it—that's how you start. And then you go through a process where you are moving these ions, these qubits, closer to each other. You're entangling them, again, using lasers. So we use, I think, eight lasers eight different times.

(Joel Beasley at 00:15:50) When you're initializing them, is it a literal zero or one, or is it the concept of the vector or the ket zero and one?

(Tony Uttley at 00:16:01) It is basically the concept of zero and one. So we have—

(Joel Beasley at 00:16:07) It is the quantum concept of zero and one. Yeah, so if we were drawing out mathematical notation, it would have the ket around it.

(Tony Uttley at 00:16:13) Correct.

(Joel Beasley at 00:16:14) Okay.

(Tony Uttley at 00:16:14) We are putting that atom, that ion of ytterbium, into a hyperfine state. And there's enough existing body of knowledge to say when it's in this state, it is acting as a one right now.

(Joel Beasley at 00:16:31) So you're putting it into superposition, or is that a separate concept?

(Tony Uttley at 00:16:34) That's a separate concept. So when we start, we actually initialize it intentionally as a zero or a one.

(Joel Beasley at 00:16:40) Okay.

(Tony Uttley at 00:16:41) Basically, so you know how you're starting. There is an operation—it's called a single-qubit operation, and you do that to be able to put it into a superposition. And so you put it into a superposition. You may do that with multiple of your qubits, put them into superposition, and then you bring them together and then you entangle them. And as you entangle them, now some of the other just beautiful parts of quantum mechanics come into play, which is as you move them apart, they will stay correlated, right? So I can now know something about one as I'm doing something to another one. And it goes back to kind of where we are in terms of coders being able to take advantage of quantum computing, right? The software. We're a little bit away because the people who know how to program a quantum computer right now are typically called theorists, and they're these hybrid physicist-mathematicians, and it is just a different level of thinking entirely of how you do this computation. The work is already being done to abstract that. So there are companies like Zapata Computing or Cambridge Quantum Computing that have built an abstraction layer of software that sits on top so that you don't have to be a quantum chemist. You could be a chemist, right?

(Tony Uttley at 00:18:10) So you can think like chemistry, but I don't have to know exactly what this means from a quantum mechanics standpoint. And that's gonna be important, right? As quantum computing continues to evolve, having those abstraction layers already built so that somebody who is more comfortable writing software in whatever language can do that and then have it automatically, basically, recompile to be able to run on the quantum computer.

(Joel Beasley at 00:18:38) It almost feels like it's like an add-on or like an upgrade.

(Tony Uttley at 00:18:45) It would be a part.

(Joel Beasley at 00:18:45) It would plug into your existing systems. It's not like you would say, oh, I'm gonna go get a quantum computer. It's like, we need quantum for this task. Let's, you know, boot up, spin up a quantum server and run our data over here with the quantum algorithm. Because, you know, you get all sorts of visions in your head as a person of what the future is going to look like in quantum computing, and it's gonna be in a watch. And, you know, it's like, why would I need a quantum watch? Like, I need a watch to tell time.

(Tony Uttley at 00:19:20) Exactly. So two things that I will say, because they are—it's, I think, easier to understand, but I will later probably get beat up by the folks. Yeah. So one is for trapped ion quantum computing, what I get a kick out of is that if you look at sci-fi movies from forty years ago or fifty years ago, you have these racks of just blinking lights. That is how our quantum computer works. That is what we're measuring is basically blinking lights. So I get a kick out of it. The second part is why do you care right now? What is quantum computing doing? And what does quantum do? So it's not gonna solve for the typical things that people think about from, like, multiplication and addition. It doesn't do that. In fact, it does that horribly. It takes so much effort to try to get a quantum computer to do a simple math problem you'd never do it. But what it does do really well is it finds, let's call it something like a global maximum or a global minimum, right, where you can look at a set of data and say, what is the lowest point in that, or what's the highest point in that? And you say, well, doesn't normal computing, doesn't classical computing do that today? Not well. And the reason it doesn't do it well is because of the way classical computers work, which is very sequentially. You may have, you know, a million bits or a billion bits, but all of them are either in one or zero. They're one value at the same time. They're always one value. That quantum computer can be in two values at the same time. And so as you add quantum bits, as you add qubits, the number of values it is considering as it's taking part of the computation increases exponentially. That's the super cool part of quantum computing is that, you know, this two to the n superpower, which is you have one qubit, that's two values. You have ten qubits, that's over a thousand values. You have fifty qubits, and now you're talking about one quadrillion values that can all be considered at the same time. And so what it looks at is it looks at all of those values at the same time, and it finds that point. And is that generally the answer? No, not right now. What you are taking that as in the rest of the computation is it's the starting point. So think about it as you use the quantum computer to give you the reference point to start the rest of the classical algorithm. And then you run it, and it comes back and says, okay, quantum computer, I want the next starting point, and you run that. And so these co-processing jobs, that's what the next three to five years of quantum computing is gonna look like, is finding those individual pieces of the puzzle that the quantum computer can do much more efficiently than a classical computer and basically leveraging that and then getting the rest of the computation done classically.

(Joel Beasley at 00:22:52) And so where do you see your role or the project you're working on? Like, are you guys an infrastructure play, or like, where do you see you fitting into the market?

(Tony Uttley at 00:23:03) So two places in particular. The first is Honeywell already has businesses that we believe are going to be profoundly impacted by quantum computing. We are in the chemicals business. We make molecules, and the development of those molecules can take three years, five years, ten years. We make materials that go into anywhere from bulletproof vests to coatings on engines. We make—we have an entire aerospace business that is built upon trying to have the right parts be at the right location to service the aircraft that are in existence. These types of problems lend themselves to quantum computing. So one of the things that we look at is we absolutely expect to be one of our own biggest customers. We have the quantum computer itself. We have the horsepower that we're going to be utilizing to do these kinds of hard problems. We also then have the ability to help our customers with those same kinds of problems. And if you go back to the classes of problems that lend themselves to quantum computing, and that can be optimization, it can be machine learning, it can be chemistry-type problems, those are extrapolatable problem sets that once we've proven out that you can generate this value from doing it, it then becomes a repeatable set of steps. And so how we look at it is something that I've talked about as this path to value creation. So we shrink the problem set down. We run it on this quantum volume that we have today. And then as you expand that quantum volume, you expand the problem size. At some point, we know we're going to cross that boundary to classical computing. It becomes classically impractical. You wouldn't have waited for that problem to be answered using classical resources because it might have taken you a month to get the answer, and the company is not gonna wait a month to get that answer. Are you willing to wait a day to get that answer or a few hours to get that answer? It may be yes. And if you are, then quantum computing becomes that option that never existed for you.

(Joel Beasley at 00:25:31) You're very smart, Tony, but you also speak well.

(Tony Uttley at 00:25:36) You're gonna ask me something right now. You're buttering me up for it.

(Joel Beasley at 00:25:42) No, I—yeah, I am. I am. I'm buttering you up for you to give us some really good advice about life here. Yes. Alright, so you're really smart, but you also speak well. And how did you do that?

(Tony Uttley at 00:25:57) So I think I had mentioned at the beginning, I had worked for NASA at the Johnson Space Center for a decade. And so I started off life as a mechanical engineer. So maybe you could argue I'm now a quantum mechanical engineer, but I think, again, most of my technical team would beat me up if I said that. I found that I had a bit of a knack for explaining what was going on in a way that maybe more of the population could understand. And as I did that, I started to find out that, hey, we had some really great ideas, and they just weren't being communicated in a way that if people understood what those great ideas were, we would take them forward. And so I did that at NASA. I did that when I was in consulting, and then I found this opportunity to do this inside of Honeywell, and it has been fantastic. It takes a commitment of an organization to say we're going to go do this over the course of ten years. We hit a milestone about five years ago that we had set for ourselves intentionally, forward-looking, and said if we ever got there, boy, not only could we be good at quantum computing, we might be able to be best in the world at quantum computing.

(Joel Beasley at 00:27:16) Absolutely.

(Tony Uttley at 00:27:17) When we did that, we came back together as a leadership team and said, okay, what is it gonna take for us to be successful here? And so the last five years have been this strategically laid out plan of dedicated resource to be able to get us from ideation all the way to releasing, you know, the world's highest performing quantum computer. So I didn't have the idea, and I think that's part of the—if there's a lesson learned in there somewhere, it is listen to other people who may have the great ideas. And then if you can go and surround them with the kind of necessary support to be able to make them happen, then you can make some very, very cool things happen.

(Joel Beasley at 00:28:09) Yeah, it's pretty interesting going on this journey of life and figuring out, like, where you're really good at and what you're really good at and what you enjoy. And it's a difficult thing. I mean, it's definitely a process. It's a journey. But I find some similarities to you where before the podcast, I was just very interested, and I would do consulting projects, building technology to solve problems. But I was always so interested in, like, going in—I'd sit in the business, understand the business, watch the problems happen, and then go back and build tools that would allow these people to not have these problems. And I, you know, I guess there was some part of me that was like, oh, you shouldn't be doing that because you're going from project to project. I was like, I just solved the problem. I solved the problem. The problem's solved. I go to the next problem. And so I gave myself the freedom to, you know, be in a position—so I realized I was in a position where I was doing a lot of contracts, building apps for, like, a decade, learning these problems, solving them, deploying stuff. And then I was like, how do I do that, you know, with the podcast? And so getting to, like, research these topics and understand them, it's like I'm just doing mini consulting, but, you know, trying to prepare for interviews and understand these concepts. Because that feeling that I want to talk about that I think you alluded to is, like, that feeling of crunching down these complex things and delivering it simply to somebody and watching them like, oh, I get it. Like, that right there is one of the most amazing feelings in the world. And I'm sure teachers get it, but they don't get paid enough.

(Tony Uttley at 00:29:44) Exactly. And you're right. It's, you have to find the joy in, I think, a couple of different ways. And one of them is seeing how amazing the technology really is. You know, so when I do this sometimes where if you hold—if you cup your hands together, I'm gonna force you to cup your hands together like that.

(Joel Beasley at 00:30:03) Okay, yep.

(Tony Uttley at 00:30:05) So within that space is a trillion trillion atoms.

(Joel Beasley at 00:30:07) Got it.

(Tony Uttley at 00:30:08) A trillion trillion. Imagine you just needed one. You have to grab one and get everything else out of the way, and then you grab another one. And that is what we are doing. We've, you know, the same scale of a baseball to the size of planet Earth is as a baseball to an individual atom. That is the scale difference that we're talking to be able to control. Now that absolutely gives me such a kick that we're able to go do that. Having said that, our customers don't care. They actually don't—they just want the answer. It's like, you guys do a bunch of good, cool, amazing technology stuff. I just want the answer. And so you have to take also some joy in the fact that, yes, you know, you were able to use all that, but the reality is we can perform a quantum computation on a real quantum computer. You know, the things that we were talking about as science fiction, you know, a decade ago is a reality. And as we're able to take and evolve this, it's going to open up an entire new space for the planet. And I, again, couldn't be more tickled in being a part of that.

(Joel Beasley at 00:31:28) I'm energized because for a long time, I'd see—I'm 32. So I'd see these, you know, companies growing. And I was like, man, I was, you know, 16 or 17 when that thing came out. And I was like, if I had been 30, I would have had the resources to do that. You know, I would have known. And so now I'm just hungry. I'm, like, looking in the market, trying to figure out what's going on. I'm saying, okay, over here you got this quantum stuff. And the market is looking at it like, oh, all you can do is run a math problem. And that is the same exact thing that happened when they had a big computer and, like, all you could do is just crunch the census data or just do some basic math problem. And it's like there's—it reminds me, I heard, um, I think I heard some motivational speaker, some life coach talking about this, and he was saying something along the lines of, don't try to convince, like, the naysayers because that's what they do. They say nay. Like, there will always be the people out there, like, oh, I don't think quantum computers are gonna be useful. I don't think this. Because they were the same people that existed back when the computer was happening. They were saying the same exact thing. And yet there are still people, I read as early back to, like, the thirties trying to fail and create computers with, like, steam pumps and cranks and stuff. And so the—to me, it's beautiful because we're sitting here and we're watching it. And as fast as things seem, it just—it's a part of my life where I can anchor in this watching the progress of quantum computing and space travel and feel like time is slow because everywhere else in life, it feels like it's going by way too fast.

(Tony Uttley at 00:33:02) Yeah. And you're not wrong at all. You know, both in terms of there are a lot of naysayers out there, and you just—you do. You have to say, okay. Well, if you're gonna use the analogy to classical computing and what happened, then you have to just at least replay the same analogy, right, which is it takes people. It takes them existing, quantum computers existing, and then it takes a bunch of smart people to say, okay, now that they do, try this, and let me try this, and let me try this. And then is this working better than that? And, you know, it's—I get this question all the time in terms of, you know, is our technology better than somebody else's, or is our technology gonna win? And, you know, it's like, that's not even how I think about it right now. There are opportunities where we may find that a certain type of quantum computer works best for a certain type of problem, and a different type of quantum computer works best for a different type of problem. And if we haven't done the exploration to say, I'm going to try those permutations, then shame on us, right? And so can I conceive of an algorithm that—again, I know people don't like to talk in algorithms. Somebody wants to solve a problem, and they hit enter to go have my problem get solved. And somewhere, this problem is distributed into multiple places, and part of it gets done by one type of quantum computer, part of it's done by another, part of it's done by some graphical processing units compiled together with a classical set of resources. I can absolutely believe that that's a thing that can happen here. And we just have to be open to saying we have to try those out.

(Joel Beasley at 00:34:44) Yeah. When I was reading and I saw—you know, I've read a lot. I think I read every piece of available information on this technology and not—and I'm not an expert, but one of the things that struck me was I liked your play with the ion because the reason why is it was an older idea. I think the University of Maryland or one of the universities had really thought it out and had, like, a lot of experts in it. And I kept noticing the one of the largest problems in the quantum computing is, uh, persistence or coherence or decoherence, one of those words.

(Tony Uttley at 00:35:27) Coherence. Yeah.

(Joel Beasley at 00:35:27) Basically, uh, and tell me if I'm wrong. It's how long the atom can survive.

(Tony Uttley at 00:35:31) It's how long can you keep the quantum information in the qubit? Yes.

(Joel Beasley at 00:35:31) Okay. How long you can keep the quantum information in the qubit? Is that because the atom degrades? Or

(Tony Uttley at 00:35:36) So again, this is where the different technologies matter. So let's take an example where some companies out there are using a technology called superconducting circuits. These are qubits that have been manufactured in an array on a chip, and the only way to even form the qubit, the only way you can even make the qubit happen, is in the millikelvin range. That means it's almost at absolute zero to be able to go do it. And then when you impart quantum information on it, it only lasts in the microseconds range.

(Tony Uttley at 00:36:15) That's the whole time that quantum information will exist, in the microseconds range. We, like I said earlier, have taken an atom. We've put it into a hyperfine state when it's in this ion configuration, and we can keep the quantum information in there in the seconds range, which, again, doesn't sound like a long time, but the difference between those two is something like 200,000 times. 200,000 times longer that we can keep the information coherent and in that quantum state. And in a lot of ways, that is like a lifetime to be able to do some of these quantum computations.

(Joel Beasley at 00:36:55) Okay. So help me get my mind around this one. I've got a series of questions. It just popped into my head. So when you write—I want to use the right word—when you write to the, so first you initialize it. Right? So it becomes initialized, and it initializes into a state of zero or one. Right? And when you want to put information, how do you manipulate information on it? Or how do you write to it? A laser?

(Tony Uttley at 00:37:24) Yep. Lasers.

(Joel Beasley at 00:37:25) Okay.

(Tony Uttley at 00:37:26) Yep.

(Joel Beasley at 00:37:26) And in the most childlike, explaining it to a three-year-old way, what is the information that is being written? How can I imagine it?

(Tony Uttley at 00:37:39) So it starts when you initialize. In the order of operation, imagine that the atom came through this hole. We hit it with a laser, one kind of laser, stripped off an electron, and now it's a charged particle. Now it's an ion. Then we move it. We move it to a space that we have another laser already set up, and that laser hits it, and it initializes it as a one or a zero. At that point in time, you have put quantum information into it right then. Okay. Now I can go do a series of computations. I can take that initialized qubit, and I can hit it with another laser that puts it into a superposition. I can bring two of them together and entangle them, and then I physically move them around and manipulate them and put them next to other quantum bits, and I entangle them. And I keep on doing it basically until I run out of coherence time. So as long as my computation takes less time than I have for coherence, then I can have what in quantum computation is really long depth circuits. So quantum algorithms that can take a long time to run, as opposed to when your coherence time is really short, everything has to happen almost at the exact same time. Otherwise, you're just not going to be able to do the computation.

(Joel Beasley at 00:39:14) So we're talking about total time on the trapped ions, the longer version of it, and what's the total time that that thing would run?

(Tony Uttley at 00:39:24) So an entire circuit, an entire quantum algorithm, might run in the span of, like, say, a second. And then you would run a thousand of the exact same circuit to be able to build up statistics to say, okay, what was the right answer? What had come out to be the right answer? And then you may put that into another co-processing job so that it is taking that answer, finding out what the ultimate statistical answer is, and then sending it to a classical computer to say, okay, based upon that answer, go run something. And so that's where you get some of these back-and-forth computations that it may take you minutes or even an hour to go and perform. But that's where we are in history. Right? We're at that time period where you can start to do these kinds of computations, show that you get the right answer, show that you can do the co-processing, and then everybody's waiting for the one next thing, which is, okay, somebody start scaling. So somebody start making bigger and bigger quantum computers that have more quantum volume so I can go do larger and larger computations. Because right now, as we talked about, yep, you can take and perform the same thing on a classical computer and get the right answer. At some point, you will cross that boundary. And when you do, it's outstanding. There's no going back.

(Joel Beasley at 00:40:54) I think the analogies are useful, but I also think they're hurtful because it's harder—because I want to draw an analogy to everything. You don't know where it's imperfect, you know? And trying to, really, whenever I get lost like this, I go first principles. Like, what is the purpose of the quantum computer and how can I best serve its own purpose? And so I keep pulling my mind back from when's it going to get to having a GUI interface to it's going to be useful to perform these functions, and people are going to go here to get that. So it might not ever look like a normal computer. It just—I mean, like, we don't think a water pump should look a specific way. It looks the way based off of the form and function it requires to solve the problem. It looks that way because it has to look that way. Right. And so that's just some of the things I'm struggling with in my head.

(Tony Uttley at 00:41:49) You're totally right. And in fact, there's an even bigger part of that that's playing out right now, and that is we talked about this two to the n because a qubit can be both a zero and a one at the same time, but that forces a mental model where it is still either a one or a zero. But that's not how quantum mechanics work. A superposition means it can be anything in between a zero and a one. And how does—

(Joel Beasley at 00:42:24) Have you come across Three Blue One Brown or One Blue Three Brown, whatever that site is? Okay. So this guy, I heard about it from Tim O'Reilly. You know O'Reilly, like, the book publishing company? So I talked to him last week and heard about it through him. And so I went and found this article too that was called The Quantum Country. It's like this paper that tries to help people understand quantum and even has these breaks in the paper where you can take a little quiz. It's really actually quite interesting. But you've seen it then?

(Tony Uttley at 00:42:57) I've just, I'm envisioning taking that quiz and failing.

(Joel Beasley at 00:43:02) I think I'm going to send this to you after the show. You're going to look at it and you're going to be like, wow, I could probably forward this link to anybody. But they reference the Three Blue, One Brown or whatever the YouTube channel is too. And so I was like, whoa. Two people in one week referencing that. Let me go check it out. They had this playlist and this teacher is the most fantastic explainer in the world because they have programming skills. They're incredibly brilliant at quantum mechanics and linear algebra. And the playlist is 10 videos and it's like understanding linear algebra. And if you understand that, you understand these core concepts and they're 10 minute videos and they're just the most beautiful, simplistic explanations that you could just tell a three-year-old and they would understand linear algebra, and it makes me so frustrated. I'm like, I went through school and it was just garbage. And so I got this understanding. I'm halfway through right now. So I have this basic understanding of the issue when people are talking about zeros and ones versus what it actually looks like in a graph, linear algebra way, and how you can look at it in three different ways, both as a mathematician, a physicist, and a third way I can't remember right now. But they show the three perspectives of this linear algebra concept, and it really broke it down to visualize and understand. And as I'm going through it, my mind's building these structures of how to think about the quantum computing, but it almost seems like it boots up. It goes to these infinite variations, and it comes out with some very specific result. So, like, you have to know what you're going into it with, the magic happens, and then it comes out with a specific result. Is that kind of how it happens?

(Tony Uttley at 00:44:47) I think about it the same way. I do. And I would love to, I'd actually love to get the link because I'd love to take a look at that. There are, I know, we have a set of brilliant theorists within our organization right now who have helped get to where we are. There are those kinds of folks in the, not many, they're not thousands, they're hundreds around the world, and it's going to take that level of thinking to try to get us to basically adopt that we don't have to think of it as a binary zero and one. We can think of it as this continuous distribution. And I do think once we can get through that construct, we're going to open up a set of possibilities within quantum computing that we just haven't considered yet. And so the tool has to come first. Right? You create the tool, and with the hammer can come a very simple structure, and with the hammer can come just a beautiful, you know, skyscraper. And you say, okay. Well, that was pretty cool. And we, again, just have to remember the time frame and say where we are is at the beginning stages of this.

(Joel Beasley at 00:46:02) So can that—

(Tony Uttley at 00:46:03) I saw the purple unicorn behind you, and I was thinking that is kind of where we are. You have to be purple unicorn. And when you do, you can showcase what the capabilities of these kinds of quantum devices really, what they can really do.

(Joel Beasley at 00:46:20) So where we're at today with the quantum computing, like, I know that number of bits, very hot topic. People are talking, oh, I've got 30 bits or 50 bits or whatever it is. But what I really want to get down, if I walk away with anything today, it's to understand this issue in my head of people and articles and things I'm reading saying, you know, it can do a lot of things that classical computers do, or it can do a lot of these, it could process formulas. It could do all that. But I'm having trouble seeing a very specific example of it. Is it because there's not enough bits yet? Like, is there a quantum playground? Is there something that I could actually see that would explain it to me? Like, okay. Here is a—like, I would be happy with this. I'll talk, let's talk about what I'd be happy with. I'd be happy with here is a linear algebra problem, and you take this problem, and you put it into the quantum computer, and you're expected to get this result. And so if you put it in this way, you're going to get that result. It's a formula. Right? So I could change the variables, but I'm going to get an unknown result. And then I could see it. I could run it on the quantum computer, and I could see it all happen. I could say, oh my god. There it is. I just did something. I pretend that this algebra thing is a problem I actually have in my real life, but I just saw it process and manipulate data. How do I get there?

(Tony Uttley at 00:47:58) That's a great question and one that you'd probably be surprised that the answer is already out there. And this is where I have some maybe some disagreements or at least different perspectives than some of the academicians, which is we absolutely do academic quality work, and the results are something that you can publish, and anybody would be able to peer review them and say, wow. You did exceptional scientific work there. But sometimes, those results are not then easily able to digest or understand. Even though if you were to simplify it down to what did this really do? So as an example, we've already run some algorithms that were designed to be able to train a machine learning algorithm to be able to say, you know, what configuration is this? And you can think about it as kind of shapes and say, okay. Is it this shape or is it this shape? Is it, you know? And then you just, you train it. You run it through a classical algorithm. You come back and you train it again. That is already happening. Now the result is, okay. Well, it's doing something. It is. It is doing something. It is working. It is training. It's just, yes, you can also use classical resources to do the exact same thing. But the important piece is it is working. It is actually using a quantum computer to do that. And, again, we're in this era where you have to build up trust. There's a short window of time that we have as a community to establish trust because at some point, again, you get past the ability to fact check it any other way. The quantum computer will be the only answer you know. And so if you haven't done all this prework to say, okay. I did this machine learning. How does it compare to doing an N classical? Oh, it gets the same answer. Now I've increased that problem set. I'm doing more complex shapes, I'm doing more complex images, I'm doing Monte Carlo type simulations that, again, I have enough resources to be able to compare the answer and say, okay. That is coming up with the right answer. And this is not a long period of time. This is something like the next, it's called eighteen to twenty four months. You break into a scenario where you have had to have already created that trust regime because now that's going to be all you know.

(Joel Beasley at 00:50:36) So you guys have, you made a video of that? Like, have you made a video of you guys doing that?

(Tony Uttley at 00:50:42) One of our partners has, and that's part of what we've been working on with a group of partners right now is to try to make, let's call it, marketing collateral that actually explains how this works simply. Because you're right. It makes most people's brains hurt when it's described in terms of its either scientific or mathematical properties. But once you see it happening, then you go, oh, I get it. This is a video of this machine being trained, and then the outcome is the answer I was looking for. So, yes, there is a video that has been created. We're expecting to be able to release that fairly shortly here.

(Joel Beasley at 00:51:22) Oh, nice. Well, we'll update the show notes when we do have it because these podcast episodes live forever. So—

(Tony Uttley at 00:51:27) Excellent.

(Joel Beasley at 00:51:28) Yeah. That'll be, I'm really excited to see that. I really, because here's the thing. There's so, everyone's on your side. I feel like the naysayers, like, point 0.1%. Like, I feel that there is so much good energy and momentum in this movement. Like, every geek you talk to, every person you talk would be like, yes. I am pro quantum computing because it sounds cool and it's the future, and we need to do it. It's just we, so we get that initial buy in, but there's so much waste and loss when we try to be like, oh, okay. Well, everything else that's complicated in life, I can understand simply because somebody's done the reduce, refine, repeat process to it. Where's the reduction for me? And walk me through it step by step in three minutes so I can actually watch it happen. And better yet, put an interface online so I can run the code myself on the quantum computer.

(Tony Uttley at 00:52:18) Yeah. You said something earlier that I wanted to come back to, which is trapped ion as an approach to quantum computing, it's been around for a long time. Right? And it has a huge benefit because these technologies absolutely take time to be able to get the physics worked out because these are massive physics challenges. So both why did it take so long and why do we go down an approach that was already existing for twenty years? Because when people say, well, what you're doing is so complicated, I go, yes. Thank goodness it is. That's what makes it so beautiful is that if you figure out a way to do something really, really complex in a repeatable, controlled way, then you have absolutely created some incredibly valuable intellectual property. And so that's what we did.

(Tony Uttley at 00:53:20) We absolutely stood on the shoulders of giants who had already figured out how do you form a qubit out of an atom? How do you impart quantum information on it? How do you make it coherent longer? How do you entangle these things? And then we said, hey, you know what Honeywell does really well? We do some of the most complex system integration in the world. We develop some of the control systems that can control anything else in terms of real-time precision at time scales that are hard to believe. We're going to layer that on top of this particular technology, and that's what we did. And so now we're at a place where we go, okay, December when we went into our beta phase for our quantum computer, that was the moment. That was this big deal because we were able to have taken all of it and proven out that we can get it to work. Everything that has come as a result since is an adaptation of that. It's scaling up what we had already done.

(Tony Uttley at 00:54:24) And so what we're excited about is not just releasing this quantum computer that we're working with partners and customers on. It's the fact that we know because it's the exact same system that we can scale fast. And that's, again, why people are excited to go see what we can go do in the very near future here.

(Joel Beasley at 00:54:48) Are you guys going to be opening up the ability to actually use your quantum computing through APIs?

(Tony Uttley at 00:54:55) Oh, yeah. So we already have a full stack that we've built. We've had customers and partners from around the world use it already. We have three different mechanisms for how we're interacting with groups. One is where people come just direct to us, and it starts with people who know how to program a quantum computer, so people who are theorists who know how to write a quantum circuit. And when they do, they just need the API token, and they can log in, run their jobs, and get the results back. There are some organizations that need some additional support. They know what problem they're trying to solve, but they don't know how to turn that into a math model that can then be turned into a quantum circuit. And for those types of customers or end users, there are companies like Zapata Computing and Cambridge Quantum Computing who we work with to go help that translation.

(Tony Uttley at 00:55:48) And then we announced a partnership last year where we're working with Microsoft as a part of their Azure Quantum offering. And so customers can come in directly through Azure and use Microsoft's programming language, which is Q Sharp, that sits on top of—that's a bit of an abstraction layer that can sit on top of our software stack—that allows people to be able to program and use both Azure classical resources and ours at the same time. So we've done that to be able to make sure that it is accessible to the world.

(Joel Beasley at 00:56:21) That's a smart go-to-market strategy. Just be the underlying infrastructure and let the biggest companies in the world sell it.

(Tony Uttley at 00:56:28) Well, we'll see.

(Joel Beasley at 00:56:30) Let's see. Oh, man. Tony, dude, this is amazing. What do you say for CTOs, CIOs that are, okay, this is quantum, this is new, they're kind of learning about it right now. Quantum strategy, quantum road map, should companies have these? Should they not have them?

(Tony Uttley at 00:56:51) Yeah. And I say this in sincerity, or with sincerity—I think it's the right way to say it—which is there are already companies in almost every industry out there that have leaned forward and have teams of people who are planning for how to integrate quantum computing into their organization. And so I say this with the utmost respect: if you're in an organization and you're the CTO or CIO and you have not done that, you're late. And the reason why—in many things, being a fast follower works out pretty well. Let somebody else go test it. Let them have all the hiccups. There is a scarce resource, and that scarce resource are the quantum computers. There just aren't many of them on the planet. And so the capacity of those quantum computers is what is going to be the limiting factor, and so the companies that have already leaned forward that we're working with have absorbed that capacity.

(Tony Uttley at 00:57:57) And you say, well, yeah, can't you just go build more? You can, but what we have done is put in place a road map where we are constantly replacing our own next best offering. And so there is a race. Quantum computing is a race, and it's a race between companies. It's a race between countries. And in order to succeed in that race, you have to be planning for not just what you're releasing today, but what does two years from now, what does five years from now, what does ten years from now look like. And, again, working with those organizations who have already thought that out along with us to be able to be prepared to insert it right into their own operations. It's not that you can't catch up. You certainly can. It's just it's going to take some energy. It's going to really take having a strategy. It's going to take getting some of your own team up to speed. You know? So go and watch the podcast and go read some of the books. And then it's going to start with, okay, let's take some of these business problems that we know we have and shrink them down to where we can work them on the quantum computers that exist today and get ready to scale along with the system. So there's a fun time to be a part of this, but it's certainly not in the you're way too early. Quite the opposite. You're late.

(Tony Uttley at 00:59:30) You're probably not too late yet, but that isn't very far away either. When we cross that threshold of that value creation threshold, that becomes a great position for these early adopters to be able to go leverage quickly and get the value out, and then it's going to take some time for others to be able to catch up.

(Joel Beasley at 00:59:53) Yeah. Well, the people who are in there playing the game are the ones that get the results. I mean, they get the biggest rewards. And, yeah, for me, it's like, if you're a larger company, I mean, you at least put together an exploratory team to see is quantum computing useful for any of the problems we're facing. Just to at least check it off the box. I mean, that's a pretty simple procedure to go through. This is great, man. I'm excited, Tony.

(Tony Uttley at 01:00:21) I appreciate the chance to talk. This is fun.

(Joel Beasley at 01:00:25) Hey, you are welcome back anytime you'd like to come back. I mean, I'm excited. I'm going to go learn more. I'm going to get questions. I'm going to talk to people, listen to the episode, get some feedback, post some of the links to the resources, and then I'll let you know what we hear back over the next couple months. And then we can have you come answer some more questions and hang out.

(Tony Uttley at 01:00:46) I'd love to. That'd be great.

(Joel Beasley at 01:00:48) Awesome. Well, thank you so much, Tony. You have a fantastic day.

(Tony Uttley at 01:00:51) Okay. Appreciate it.

(Joel Beasley at 01:00:52) All right. Talk soon. Thank you so much for listening. And if you found this episode useful, please share it with a friend or colleague who you think would get value from it. And if you have topics that you would like to hear discussed on the podcast, either add me on LinkedIn or send me an email [email protected]. Every time I get an email or LinkedIn message, it absolutely makes my day and inspires me to keep going.