Dr. Julie Love - Senior Director of Quantum Computing at Microsoft

(Joel Beasley at 00:00:00) Hello, my friends. Today we are talking to Julie, the senior director of quantum computing at Microsoft. And we discuss the advancements that have been made in the world of quantum computing, how you can write code today using Azure Quantum and their QDK, and the security implications surrounding quantum and encryption. All of this right here, right now on the Modern CTO Podcast. Here we go.

(Joel Beasley at 00:00:26) This is the Modern CTO Podcast. Hello, Julie.

(Julie at 00:00:38) Hi, Joel.

(Joel Beasley at 00:00:40) I was reading all about you. I'm like, oh man, this person's so much smarter than me.

(Julie at 00:00:46) I'm sure that's not true.

(Joel Beasley at 00:00:48) You were a quantum physicist at Yale. How did you get interested in that?

(Julie at 00:00:53) You know, I fell in love with physics in high school. I had a really special high school experience. I went to a small all-girls Catholic high school in Nebraska, and I had this amazing physics teacher from my junior and senior year who taught physics in an AP physics class. And it just captured my imagination. The idea that you can have a framework and a set of equations that described and helped you understand the way that the world worked and that you could actually then go build stuff and measure it in the real world was super fascinating to me.

(Joel Beasley at 00:01:30) So a teacher had a large impact on your life.

(Julie at 00:01:33) Absolutely.

(Joel Beasley at 00:01:34) So then you go to Yale. What did you get to work on in the lab?

(Julie at 00:01:37) Well, I started out at MIT.

(Joel Beasley at 00:01:41) Oh, okay.

(Julie at 00:01:41) Right when I got there, I met someone who had just come back from a summer at CERN doing particle physics experiments. And that was the most exciting thing I could possibly think of to do. And so I said, okay, that's my goal. And so I went and joined the lab, and I started out in high-energy nuclear and particle physics, building detectors and designing particle detectors that would eventually end up on the space station.

(Julie at 00:02:06) And that was super fun, but by the end of my time at MIT, I had taken every class in quantum mechanics that they had. And they offered, for the first time in the graduate school, a seminar class on quantum computing, and it brought these two areas that I was interested in together. And I took that and that changed the path of my studies, and that's how I ended up at Yale and doing work in quantum computing and quantum physics.

(Joel Beasley at 00:02:33) And then you have this jump. I'm looking at your stuff, and you've got this jump of all this quantum computing at Yale and MIT, and then you go into corporate strategy. I was like, well, how did that happen? She must have seen everybody partying and making a ton of money. She's like, I'm gonna go over there because I know I can be really smart over there and make a ton of money.

(Julie at 00:02:54) Not exactly. But, you know, the way that jump happened, because it does look like a discontinuity in my resume and in my experience... You know, the way that that happened was I was doing my work at Yale, and at that point, quantum computing was super exciting. I'd seen the promise of what this technology could do and the impact that it could have on the planet. But at that time, when I was finishing my degree, it felt really, really far away.

(Julie at 00:03:19) And it was this amazing promise of, if we could build a quantum computer, we could do these amazing things, like solve some problems exponentially faster. But at that time, it felt really, really far away. We could make one and two qubit systems. We couldn't make them... You know, one of the key issues with qubits, or these building blocks for the quantum computer, is that they have any interaction with the environment, it measures the quantum state, and we lose the quantum information. So we couldn't get them to stay quantum very long.

(Julie at 00:03:48) You know, for me, it was hard to imagine this path to scalability in terms of having enough that you could do something interesting with it. And so at that point, quantum computing still felt really, really far away and was a big if. And I felt like I needed to get a broader set of skills, and I saw myself more on the productization side of things. And I felt like I needed to get a broader set of skills to do the type of work I wanted to do and have the impact I wanted to have. And I went to McKinsey and did work primarily in semiconductor manufacturing.

(Julie at 00:04:17) So I went from the academic clean room into industrial clean rooms and helping semiconductor companies with all the elements of their business. And so from that perspective, it was a huge jump going from an academic research lab into the heart of corporate America. But from a work perspective and a content perspective, it wasn't as radical as it seemed or it could seem from the outside.

(Joel Beasley at 00:04:43) Yeah. I jump around a lot. I got into software really young, and I started building software. And then the thing that fascinated me was getting to learn an entire business. I would learn the entire insurance business, build an insurance software, then I'd learn the entire fitness business. And that variety of learning the business really kept me energized and excited about what I was doing, which is super important.

(Joel Beasley at 00:05:08) You're getting to do quantum physics. I have so many questions for you. But first, can you tell me what do you do at Microsoft with Azure Quantum?

(Julie at 00:05:18) Well, so with Azure Quantum, we're really bringing this technology to developers. So we're hoping you get your hands on it. And so my role at Microsoft is I run program management for the quantum program, and this spans everything from the qubits that we're building to the control systems to the cloud platforms and everything in between. And so with Azure Quantum, we're really bringing this technology that we're building. We have an open source programming language called Q Sharp.

(Julie at 00:05:45) We have a quantum development kit, and we're bringing those together with the power of Azure to make this open, accessible ecosystem to start building and learning and playing with quantum technology. So you can use this dev kit and the language to run a simulator of a quantum computer, and you can just download and run that on your laptop. Or through Azure, you can run that on real quantum computers from partners of ours. So we have trapped ion systems in there from Honeywell and IonQ. We're bringing superconducting qubits to that.

(Julie at 00:06:18) And then a really exciting thing that we've discovered over the past few years on our team is that as we've learned about how quantum computers solve these interesting hard problems, we can use some of those techniques already on classical hardware. So the hardware that we already have in Azure, and we call those quantum-inspired techniques. And we can use some of those properties and principles of quantum computing and quantum physics to accelerate problems using the hardware that we already have. And so we have capabilities for those quantum-inspired tools already in Azure Quantum as well.

(Joel Beasley at 00:06:50) And so how did this all happen? Did Kevin Scott call you up and say, Julie, we need you. You're the quantum king. Come do this with us. How did you get to Microsoft?

(Julie at 00:07:01) So I came to Microsoft originally out of McKinsey. So I went from Yale to McKinsey. And I came to Microsoft and did a variety of mostly corporate strategy roles, as you pointed out earlier, working across all of Microsoft's businesses. And then I took a role at Adobe leading strategy for their creative business. But along the way, where I didn't have a discontinuity is I'd been helping friends and former colleagues from Yale start quantum computing companies.

(Julie at 00:07:28) And so I've been advising quantum startups, and I'd really kept up with that field. I have a lot of friends. It's a small field, and I have a lot of friends in the field. I stayed in that field, and I'd been looking to see when I would do quantum computing again as my day job instead of on the side of the other things that I was doing.

(Julie at 00:07:49) And there was a terrific opportunity at Microsoft to come back to Microsoft about, going on four years ago, to come back and lead business development for the quantum organization. And that was really, you know, business development in the most fundamental sense of thinking about what would a business around quantum computing even look like? You know, really understanding what customers needed from that and creating programs and products around that for enterprise customers, for startups, and for partners of ours to really bridge the gap between this new emerging technology that's coming online and really making it so that we understand that true customer need for the technology and guide our thinking and development of all of the pieces that need to come together to make this a reality.

(Joel Beasley at 00:08:37) Well, you sound like the perfect fit for the role. You've got this entrepreneurial business ninja experience, right, helping with all these startups. Meanwhile, you have the subject matter expertise of the quantum. So you can connect the applicable business use cases today to wherever the technology's at as it matures, because you gotta get that hook, that money so you can improve it and grow it. It's one of the hardest things to learn, but once you learn it, it's the most clear and obvious thing you have to do.

(Julie at 00:09:05) Well, thank you.

(Joel Beasley at 00:09:06) I wanna know. Alright, so we've got a lot of CTOs, VPs of engineering, geeks that listen to the podcast, right? And so I was curious because you're so deep in the quantum world. You've got other bright people out there that just don't spend time in it, right? They're in the computing world, but they're just not deep in the quantum. What's some of the misconceptions or the things that you would wanna put out there and let them know?

(Julie at 00:09:33) Yeah. That's a great question. You know, one of the things, I guess, is a common... There's a couple of common misconceptions. One is that, you know, I hear from some people that quantum computing can solve any problem faster. You know, that's an element where really understanding what quantum computers can do and what they will be able to do, it's a harder problem than that, and it's an oversimplification. So first of all, you know, we're computing with really different building blocks, and they work in really different ways, and we have to program them differently. And so they're not... You know, quantum computing is not gonna solve every problem faster.

(Julie at 00:10:22) One of our experts at Microsoft, and distinguished scientist is the title, he says, you know, on average, quantum computers will solve problems about a billion times slower for the average problem. And it's really finding these problems where you can take advantage of these quantum properties like superposition and entanglement and using the wavelike nature of these particles to find problems where you get an exponential or a polynomial speed-up that's greater than quadratic. And then getting into the details of how that algorithm works. You know, all the things that we know how to do classically. What's the IO? How many gate operations do we need? And getting into the details of that because there's lots of places along the way where you could lose the quantum speed-up that you thought you had. And so that's one of the problems.

(Julie at 00:10:59) And we see clear areas where quantum computing is gonna have an impact, areas like chemistry or material science, where you're high up on the periodic table or you have a lot of what's called quantum correlations and molecules, or these problems are really, really big compute problems. And so that's the first place. It's gonna be... We're focused on areas where it's big compute because that's where quantum can make a big difference. And so we hear a lot of problems that are really high-performance computing problems, big data problems, and that's not where we're gonna see the impact. And so I'd say, you know, the first misconception is this is gonna solve everything faster.

(Julie at 00:11:38) And the second misconception is really getting into unpacking this word "faster" because that's also, you know, the devil's in the details on how we make these things work and where those improvements are gonna come from. So it's not just that we can solve... It's not that we can solve every chemistry problem faster. It's where we have real bottlenecks in classical compute that send these run times into billions of years with all the compute power that we have. And it's, you know, in the details of that. So if you think about problems in, say, drug discovery, we talk to a lot of pharmaceutical companies and, you know, for organic compounds where they can do those calculations using high-performance computing, they're looking for higher throughput typically on those simulations.

(Julie at 00:12:28) And where quantum computers are really gonna help is where we can get the energy accuracy to where we need it to design new compounds. And so if you think about the average simulation for drug discovery, they're looking for increase in throughput. And what we often hear is they would trade off energy accuracy for increased throughput. And so that's not specifically where quantum computing can help. But an area where it could help is demonstrated in the work that we published last summer on, you know, thinking about designing new catalysts.

(Julie at 00:13:01) And one that's really exciting and top of mind for us at Microsoft related to the pledges that we've made in carbon is there have been proposals for ruthenium-based catalysts for removing carbon dioxide from the atmosphere. But to design a catalyst that would be really efficient, you need to understand the energy levels of all of those intermediate reactions. So all that process where you have carbon dioxide interact with the catalyst, and it undergoes this chemical reaction where you spit out the catalyst again and you have something valuable like methanol come out and you consume the carbon dioxide. But to have efficiency in that catalyst, you need to understand those energy levels really, really accurately. So that's an example where quantum computing can help.

(Julie at 00:13:48) And it takes that calculation from maybe hundreds of millions or billions of years on the computers that we have today into what we believe can be done on a quantum computer in about a month.

(Joel Beasley at 00:13:59) That's pretty cool. I was just listening to Elon Musk talk about the... He's just announced a $100 million prize for...

(Julie at 00:14:08) Have you heard about this? I haven't, actually.

(Joel Beasley at 00:14:10) Oh, okay. So they're looking for new ways to extract carbon from the atmosphere. And because the current ways that we know about, there's enough problems with them or they're not efficient enough. And he was describing how, you know, carbon, when it's in the atmosphere, it's in this low-energy state and trying to get it to rebind to something physical takes a lot of energy. And that's the problem that needs to be solved, which I think is what you were just discussing.

(Julie at 00:14:34) Exactly.

(Joel Beasley at 00:14:35) Yeah. But he's got this contest going for $100 million, and we should form a team and you should do all of the work. No, I'm just kidding. Feels like school again.

(Joel Beasley at 00:14:45) But yeah, I think that's fascinating that he's out there advancing it and he's not waiting for this to happen. He's just saying, alright, here's $100 million. Let's find some better ways. I really like when him and Bill Gates and when these wealthy people who are, you know, titans of their industry are doing things like this. They're helping humanity move forward.

(Julie at 00:15:07) It's so important, and we think about this every day, you know, here at Microsoft and the work that we're doing saying, you know, we're building this incredible compute capability. Let's go solve the world's most pressing problems with this.

(Joel Beasley at 00:15:20) When am I gonna be able to go to Best Buy or pull up my Amazon app and buy a quantum computer?

(Julie at 00:15:27) Well, you should pull up your Microsoft Azure account.

(Joel Beasley at 00:15:31) Oh, just kidding.

(Julie at 00:15:34) And I mean, you can use this today. So you can use this and start to do small experiments with Azure Quantum today already. So I think in terms of buying a quantum computer, I think that'll be a little bit further out. The quantum computers that we're building have to be run at millikelvin temperatures. So maybe not everyone's calibrated in millikelvin, but it's a few hundred times colder than deep space.

(Julie at 00:16:00) So it's not something that most of us have lying around the house. So maybe a little while before you're buying that. But the more relevant thing is, think about what you're accessing through the apps on your phone. For a lot of this, you have some front end interface on your phone, but you're accessing some massive cloud resource in the background, whether it's something that's done with machine learning or your movie repositories, all of that streaming to you from the cloud. And that's, for the foreseeable future, the natural way to access these quantum computers will be in the cloud. And this is gonna evolve over time as we move from these kind of small demonstration systems where we can learn and play and start to build and understand how to program these things.

(Julie at 00:16:45) Programming looks really different when your bits are in superposition of zero and one, and you have to use things like quantum entanglement to get results. And those systems become much more powerful over time, up to the point where we have systems where we can reliably control the qubits, keep them quantum long enough to do interesting things, and have a system that scales up into the millions of qubits that we need to solve these problems, like designing new catalysts for carbon fixation.

(Joel Beasley at 00:17:17) I like that. Yes. Definitely go to Azure if you wanna play with the quantum playground thing that they have. I still want one in my office. I would just like one right here. Maybe I buy a used one on eBay or something, and it may not even turn on. But I just think it's so cool because it is the future. I mean, this is stuff that was being theorized twenty, thirty years ago, and now we're seeing that it's reality and it's fantastic. Right?

(Julie at 00:17:44) It is. It's really, really exciting. And I think back, you talked about the just seeming discontinuity in my career of going from this place of being in a university lab thinking, if we could ever build a quantum computer, to now saying we're looking at when will we be able to solve these hard problems, and what are the end pieces of engineering work that need to be done to be able to get there and put this into your hands to go solve these planetary scale problems.

(Joel Beasley at 00:18:16) When you're out and about, right, like at a restaurant or with family and you're meeting new people, how do you explain to them what you do?

(Julie at 00:18:24) Well, when we used to do that, go out and about.

(Joel Beasley at 00:18:28) Yeah. Back in the day.

(Julie at 00:18:29) Back in the day. Yeah. It's a good question. I'm trying to think. I get a lot of questions from my kids too about, you know, what is quantum computing and what do we do. It really depends on who I'm talking to, certainly how I describe it.

(Joel Beasley at 00:18:46) Or people wanna talk.

(Julie at 00:18:47) Yeah. But it's, you know, at the heart of it, it's figuring out new ways to solve these really hard problems. And that's been one of the things that's been continuous along my whole career is really tackling hard problems. And it's, for me, tied deeply into scientific method and my training as a scientist. But figuring out how to solve these new problems, because all of the technology, while it's super interesting and there's tons of really fascinating physics in every aspect of what we're doing, the only real value of any of this technology is what you can do with it.

(Julie at 00:19:26) And so really understanding and thinking about what I do and our team here at Microsoft is really understanding what are these hard problems to go solve, and then how do we assemble all of the building blocks and put that together into an integrated system to solve those real problems. And there's lots of elements of that. It's really deeply understanding the problem, matchmaking that with the technology and understanding, can our capabilities help, improving our understanding of what those problems are, and then building to really deliver the system. And the underlying quantum physics and the qubits or the building blocks of these, those are just, that's one small part of building this integrated system.

(Julie at 00:20:09) We have to control it. We need programming languages. We need new types of compilers, runtime, building the cloud ecosystem, building the services to make it accessible by the people who understand these deep problem domains. And so it's really figuring out all of those pieces and putting them together in the right way to get to those solutions. And so that's I guess that's how I describe what I do.

(Joel Beasley at 00:20:35) Yeah. Organizing people to achieve outcomes. Right? Now here's what I wanna know because I'm an entrepreneur. I've owned several businesses. I currently own a business. And so I'm always interested and fascinated to talk to other people. You're essentially acting as a business owner because you have this business unit. Right? And so I'm always fascinated by team structures and things of that nature. And what I'm hearing you say is that you have to have a deep understanding of these domain specific languages, essentially. And you have to be able to think that there might be a potential. I talked to a customer. Right? You said that earlier, and they were talking about throughput versus, I believe, energy density. And then you figured out that it wasn't a good applicable use. Right? So how are you structured? Do you have teams of people that are searching for this? Is this on a much smaller scale where you're just looking for one really large market? So it's essentially like you and one other person going out and talking to people. How do you actually do this as a corporate strategy? How do you do it?

(Julie at 00:21:40) Yeah. So I'd say how we think about the work that needs to be done, we're really focused on the impact of the technology. What problems can we go solve? But then, as you mentioned, we have to really understand those specific problems and verify that quantum computing will be a good solution to that. And so we have experts on the team who are inventing and coming up with these quantum algorithms.

(Julie at 00:22:06) But in order to really understand the speed up, we need to write those algorithms down in code. So we need to write out the program. It's really hard to know if your program's gonna run when you if you were to write your code down on paper. It's hard to debug. It's hard to know if you get the right answer. And so we need programming framework to be able to verify the speed up. And so as I mentioned earlier, it's the inputs and outputs because quantum computers are gonna be slower. The clock rate is gonna be slower. It's all of the gate operations that are happening. And so doing that translation of, let's say, an academic algorithm that someone's come up with and said, you know, this algorithm has an exponential speed up.

(Julie at 00:22:47) But that algorithm, that exponential, there's this academic notation with this curly O in front of it, which means order of. It's order of exponential speed up, but that can have a large constant prefactor. So you may think you have something that would run in an hour or a month, but if the constant prefactor is a million, a billion, it's gonna have a really long run time. So we have that work that's done to write that algorithm down in code and know the run time. And so we've developed tools and capabilities. We've developed the language. We've developed tools on resource estimation. Just because, if you think back to when the first programs were written, we didn't have those tools and capabilities and understanding of how code works. And so today, we have tools where we can know how long a run time will be for an algorithm that will run on a quantum computer that we don't have yet today, that we might have in five, ten years from now. But we can do that coding work now, which is really, really impactful.

(Julie at 00:23:47) So then that tells us, okay, for a specific, let's just stick with this chemistry algorithm that I talked about for carbon fixation. So we say, okay, we've got that run time measured. But our first cut at that run time might say, oh, it'll run in four hundred years, which is a great improvement from the classical algorithm that maybe was billions of years. So we've gotten it down to hundreds of years, but most of us want results faster than that.

(Julie at 00:24:14) And then we can do that computing work. So it's a set of skills on writing good fast code, high performance computing, where can we parallelize, really understanding the details of those algorithms. And now we've gotten that down to just saying, okay, now we know that we can solve that in one month if we had a quantum computer that looked like this. So that understanding of the problem, the run time of the algorithm, that gives us insights into what hardware that we need. Now we say, okay, we need qubits that stay quantum long enough. We need gate speeds that are fast enough. We need to have something that's reasonably sized that we can actually manufacture. So if we're doing this at super cold temperatures, so if we had a qubit technology with a physical dimension such that with the millions of qubits that we need is the size of a football field, that's gonna be really hard to make cold enough and have the engineering infrastructure around that.

(Julie at 00:25:13) So that really informs the decisions that we've made for the qubit technology that we're working with. And so we were taking a really different approach to that fundamental building block, the qubit, and we're using what's called a topological approach that protects the quantum information. And so we've got our team, going back to your question, we've got our team structured along all of these elements. So there are people who have deep understanding of the physics of quantum systems and condensed matter systems that are designing these qubits. We have teams that are growing the materials and fabricating the devices. It's a strong feedback loop as we validate what we're building in the lab with the theory and simulation. We've got teams that are building these measurement and control capabilities. So one of the announcements that we've been talking about in the last year was a team in Sydney, Australia that's designed a cryogenic CMOS chip that can control up to 50,000 of these qubits with just three wires running up to room temperature. So that's an important element because information carries energy. And how do you control this from a hot noisy data center? And how do you get information in and out of that without disturbing the qubits? And then we've got folks thinking about all of the software stack. And so runtime, compiler, language, and we've got teams working on, okay, if we want to make this accessible to customers, it's gotta be a cloud service. And what are the tools and capabilities that we can put into developers' hands to build that?

(Julie at 00:26:47) And so that's really how we have our team structured across all of these elements of the stack that are required. And then, of course, because we're doing all of this in parallel, that gives us rapid feedback mechanism so that we are really efficiently designing and building these systems.

(Joel Beasley at 00:27:04) I was talking with John. He's the founder of Blackpoint. And what they do is pretty cool. They do like cybersecurity, but like nation state grade, like NSA guy. Very, very cool stuff. Right? Had to cut a lot of interesting information out of that podcast. But there's a lot of talk about like quantum and security, specifically with encryption around Bitcoin. And I was curious, like, when will I be able to use a quantum computer to steal all of the Bitcoin?

(Julie at 00:27:37) So, you know, that's, to abstract your question a little bit. So that's top of mind for a lot of people. Top of mind for us at Microsoft is that, you talked about quantum and encryption, and one of the early algorithms that was designed for a quantum computer is Shor's algorithm that does large prime factoring very, very fast. And so that's an example of this algorithmic approach. Classically, factoring large prime numbers is really hard for classical computers to do, or figuring out the two large prime factors that go into an RSA key or elliptic curve algorithm.

(Julie at 00:28:15) And that's something that quantum computers are really, really good at, it turns out, exponential improvement in speed. But this goes back to something that we talked about earlier, which is quantum computers are not gonna be good at everything, which is, you could say, well, that might be disappointing for some, but it's actually good news for encryption. If you think about these systems and the way that we communicate in the world, it's a good thing because we can pick another, encryption protocols are based on hard math problems and the hardness of that math problem for all of the computing technology that we have. And so this is an area, well, you may be excited to crack Bitcoin, but we actually, we wanna live in a world that's safe and secure, and we can encrypt our information.

(Julie at 00:28:58) And so there are, we and many teams around the world are working on what we call post quantum cryptography. So quantum cryptography that will be robust to quantum attacks from these systems. And so that's where I feel reassured that we have viable solutions in place already today to protect us from quantum attacks. And these are classical algorithms. So that's another, it's part of this relief that quantum computers will not be good at everything. We can use different classical protocols that we believe are secure from quantum attacks to encode our information.

(Joel Beasley at 00:29:36) I was hoping you'd tell me, give me a different answer.

(Julie at 00:29:37) I know you were.

(Joel Beasley at 00:29:38) No. I talked to the creator of Ripple, which is one of the popular cryptocurrencies, XRP. And I asked him about this. That guy, I'm telling you what, he was ridiculously smart. He was explaining to me that the current, he has a ten year horizon, so he doesn't think it's a problem within the next nine to ten years. So they're not allocating resources currently to be solving this problem. He also explained that there is algorithms that are very susceptible to being cracked by quantum computing. And then there's also quantum computing resistant algorithms. So he explained to me that there's these differences out there in the marketplace, and he says at any point in time, if something happened, we could just switch it. The problem is the quantum resistant algorithms are slower. So for transaction speed and things of that nature on the network, it's just a slower process. So they're using the current traditional one. So that really took this big cloud mystified thing that I had researched for hours online, seen all these different opinions, and you just drill down, you get to the expert, and they just explain it to you like you're a child. And I was just like, thank you so much. But I was hoping Julie would tell me, we've got something secret at Microsoft, Joel, and all the Bitcoin is yours.

(Julie at 00:30:54) No. He's right about that. And there's these different cryptographic protocols, and they have different trade offs. It's absolutely right. Some of them are slower. They might take more compute resources. The key size changes. And the thing to, we don't know. I think that ten year time frame sounds about right. You need about 4,000 really high quality qubits.

(Julie at 00:31:19) And so with every qubit technology, you have to do some level of error correction. So the number of physical qubits you need is orders of magnitude larger than that. So this is going to be a really big system that's needed for that. But the issue with—I think we don't want to wait until somebody has that system in place for a couple of reasons. One is that anything you're encrypting now, you have to think about what's the security lifetime of the information that you're transmitting.

(Julie at 00:31:46) And so, you know, if you're thinking about your Bitcoins, your encryption around your Bitcoin or your passwords, those are relatively easy to change, but we don't want those out in the open. If somebody had this technology and could decrypt them tomorrow, we want something that lasts longer than that. But if you had really secret information, like health information or, you talked about nation state security, you know, things like this, you think about a security lifetime of that data being much, much longer than that. And so if it's longer than when you think this information could be cracked, you actually need to start thinking about that now. It could be because you could think of someone who you wouldn't want to have this data fall into their hands just recording it and saving it so they could crack it later.

(Julie at 00:32:34) So it is, you know, it is important to think about it now. And then the other factor—so the first one is that lifetime of the data that you want to protect and how long you want to protect it for. And then the second is how long will it actually take you to switch? And so, I think about it as we talk to CTOs and leaders who are concerned about security, do you have the cryptographic agility to make that switch overnight? And as we think about building these systems, building them in with that cryptographic agility to be able to change quickly is important.

(Julie at 00:33:09) And then, of course, I guess maybe I'll add a third one, which is you don't want to—you want to have time to test, of course, any system that you're putting in place.

(Joel Beasley at 00:33:17) There we go. I was talking with Tony—I think his last name is Utley—Tony Utley from Honeywell, I think, like, last summer. But I saw that you had a partnership with them on the website.

(Joel Beasley at 00:33:28) So I went and then looked at the free trial and everything. And then at the bottom, you had your partners and you had Tony. You also had Trimble on there. You had a bunch of partners. I kind of went into a rabbit hole. Like, oh, I know all these people.

(Joel Beasley at 00:33:40) It's just so cool just to get to see them and to see everyone working together. It's weird because it's such a large world. It's such a large community, but it's also very, very small. I think earlier you were saying quantum computing is like that too.

(Joel Beasley at 00:33:54) How did this partnership happen, and what is this partnership with Honeywell?

(Julie at 00:33:59) So the partnership with Honeywell is all around Azure Quantum, and it's putting these tools and capabilities in people's hands. And so this came from—we've been in conversation with Tony and the team for a long time. And we're building our end-to-end quantum computer with this really different hardware and the programming framework, coding language. Of course, we have the Azure infrastructure. And in conversations with our customers and partners like Honeywell, we really believe that it's important to put all this technology in the hands of developers and researchers worldwide, and it felt really natural for us to work together on this.

(Julie at 00:34:41) So, you know, combining the quantum technology that Honeywell is building with the programming models and frameworks and cloud infrastructure that Microsoft is building. So we're super excited about the partnership with Tony and his team and combining these capabilities into this open ecosystem. And then this all goes back to the impact that we want to have with this technology. You think about problems here in chemistry, material science, hard optimization problems. It really takes building a global ecosystem of different players in the industry to not only build the technology that's required to solve these problems, but to understand all of these use cases and problem areas.

(Julie at 00:35:22) And we see that as super, super important. We don't know where these developments are going to come from. And so it's really important to put these tools and capability into the hands of people like you and developers and researchers all over the world and put that together with the quantum hardware that's available today. And the beauty of that is—we talked about resource estimation and being able to understand the run time of a quantum computer that we don't have yet. Well, we've built a high-level programming language. So you can do that programming.

(Julie at 00:35:56) It's abstracted away from the hardware, and what that gives you is a programming language that can run on the quantum computers that we have today from players like Honeywell and IonQ. They're going to continue to evolve those systems and grow. And you can run that code across multiple different systems and multiple generations of, say, the Honeywell machine. And so they've got their H0 machine in Azure Quantum. They've introduced the H1 machine, which is the next evolution of that hardware. And you can have that single Q# program that gives you the durability of your code, and you can run that on larger quantum systems, higher quality systems as those become available.

(Julie at 00:36:38) So we see that as really, really exciting, bringing different partners in the industry together with our tools and capabilities and making that available to the brilliant minds around the world that are going to go figure out how to solve these tough, interesting problems with this technology.

(Joel Beasley at 00:36:56) So we can go on there right now and download this programming language or do it in the cloud and actually write some code, run it on a quantum computer. I could select if I want to run it on Honeywell systems or this IonQ, I think you said. I can choose. And from what I understood when I talked to Tony is that these different technologies—because there's a couple different companies out there, they're doing quantum computing, but they're using different underlying technologies to do it. And they have little pros and cons and differences.

(Joel Beasley at 00:37:25) So this type of quantum computer might be better in these areas. Is that correct or is that not correct?

(Julie at 00:37:31) That's roughly right. And so there's a set of requirements that we've known for a long time actually. What it takes to have a—to build quantum bits, let's say, or have a quantum computer. You need to have some quantum two-level system that you map to your zeros and ones. You have to be able to control it.

(Julie at 00:37:49) It has to stay quantum long enough to do something interesting. So we have a general idea, but there's different underlying building blocks that people are using to build this technology. And we're building topological-based quantum bits. People are building superconducting qubits. Honeywell's building trapped ion systems.

(Julie at 00:38:08) And so there's different flavors of these, and they have different characteristics. There are different error rates of the qubits. So how fast they lose their quantum information—it's kind of a shorthand way to think about it. What's the gate speed? Tony's system is controlled by lasers.

(Julie at 00:38:21) Some are controlled by microwave pulses, and so there's these different pieces. And so your control systems look different and how you engineer these systems. And it's super exciting to be able to then put these different types of technologies together, just in your Azure portal for developers to then go try and use and test and learn from.

(Joel Beasley at 00:38:45) Excellent. Where do they go to do that?

(Julie at 00:38:47) Azure.com/quantum.

(Joel Beasley at 00:38:50) There you go. You nailed it too. I was like, oh, no. I just put her on the spot.

(Julie at 00:38:55) No. I know. It's exactly where you go if you get Azure to do it.

(Joel Beasley at 00:38:59) Yeah. That is actually really easy to remember too. Julie, this is great. We made a podcast. How do you feel?

(Julie at 00:39:05) I feel great. This is super fun. It's—you've really dug in deep to this technology. It's exciting.

(Joel Beasley at 00:39:12) I'm a super nerd. So I started taking, what was it, linear algebra courses to better understand because I wanted to understand—last year, I was trying to figure out where is this, because I'm an entrepreneur. Right? So I was like, where is this technology? I want to analogize it back to the past, you know, in the sixties when computers were the size of rooms, so I can figure out where the market's going.

(Joel Beasley at 00:39:37) Right? And when I can get involved, when does it come out of the nerd's hands and get into, like, the—I could start selling it. And in that, there was no clear information online about where it was because I'm a software engineer, so I was trying to figure out all of these questions. And so I said, well, I've got this podcast. Let's just go get—I got Robert Sutor from IBM, and I got Tony from Honeywell, and then I got Julie from Microsoft.

(Joel Beasley at 00:40:00) I was like, let's just go get these smart people and figure out where this technology is at so I can better understand it. And it is fascinating.

(Julie at 00:40:08) It's super fun. And if you want to keep on your learning journey, we've put together in the past year a set of learning resources. And so now we have a learning website. It's just coming online where you can go and get tutorials. We have the MS Learn resources just from Microsoft, and we've put quantum modules into this MS Learn resource.

(Julie at 00:40:32) You can just—you can go take quantum classes from Microsoft. And then as you're going on your coding journey and getting hands on with Azure Quantum, there's a set of tutorials we call the Quantum Katas. And so it's modeled after karate. These methods where you start relatively basic and then you get more and more complex over time. So you can start with our Quantum Katas, start writing quantum code.

(Julie at 00:40:58) You can run it on quantum hardware through Azure Quantum and grow your capabilities in a really hands-on practical way.

(Joel Beasley at 00:41:06) I spelled it right. I made up the word katas. I don't know what it means. But I spelled it right. I'm on the documentation page right here, learn by doing.

(Joel Beasley at 00:41:14) I'm going to play with this a little bit. Oh, look. Kata. Linear algebra.

(Julie at 00:41:18) I know. I was going to say your linear algebra was the exact right starting point. So all the math for quantum computing is just linear algebra. And you can think about it as matrix multiplication. And why that gets really hard and why we need quantum computers is those matrices get too big for classical computers to do.

(Julie at 00:41:36) And so that's where you can start to use those quantum resources. But linear algebra is a perfect starting point.

(Joel Beasley at 00:41:45) 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'd 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.