PsiQuantum Co-Founder, Jeremy O'Brien: Quantum Is Sooner Than You Think

0:00Jeremy O'Brien:Quantum computing could have a really powerful impact, whether it's next generation solar cells, hydrogen production, ammonia production, even steel and cement productions. We can't actually start to do chemistry until we have a quantum computer.

0:41Today on the Giant Ideas podcast, we welcome Jeremy O 'Brien, co-founder and CEO of SciQuantum, one of the world's leading quantum computing companies, to tackle some of the world's greatest challenges. Jeremy has dedicated 25 years to this mission. He was previously a professor at Stanford as well as Bristol in the UK. This is an episode for anyone curious to peer into the fascinating world of quantum computing and to understand the radical changes it will have on the world if it scales. There are huge opportunities to transform healthcare, climate, finance, transportation, security, and much more.

1:16In this episode, we talk about what quantum computing is, how soon it's coming, and what will actually happen when it scales. We're really excited about this one. I think in some ways, Jeremy has the potential to be the person we've had on the show who most profoundly changes the world. Jeremy, welcome to Giant Ideas.

1:33Jeremy O'Brien:Thank you very much. Perhaps no one in the world is a better place to explain quantum than you, having been in it for over 25 years as a professor. How would you try and explain it to your dinner party guests? Yeah, I guess I would just say that quantum mechanics is a physics theory that's about 100 years old and it is in fact the most successful theory that humans have ever had in terms of its you know accuracy and precision of making predictions and the degree to which it's been tested and it tells us some really counterintuitive things about how the world actually is at the microscopic scale so it tells us that you know a particle can be somehow in you know more than one place simultaneously it tells us that two particles can be inextricably linked with one another such that doing something to one of them somehow instantaneously affects the other without you know no matter how far apart they are you know there's no transistor without quantum mechanics you know there's no semiconductors without quantum mechanics so you need a quantum mechanical description at the very bottom of a computer at the transistor level.

2:47Jeremy O'Brien:But then immediately above that, the quantum mechanics gets abstracted away and you don't have to worry about any anymore. It's just a conventional, what physicists call a classical system with binary logic, zeros and ones, and so on and so forth. And again, not a deeply satisfying explanation, but I think nevertheless starts to get into it. A quantum computer is one where that is not the case, where from the equivalent of the transistors all the way up to the applications, quantum mechanics is threading through that system, and it's harnessing these very surprising counterintuitive effects of quantum mechanics to do some things much, much faster than any conventional computer ever could, and in some cases exponentially faster, and things that are that are effectively impossible for conventional computers to do suddenly become impossible if you have a large-scale quantum computer.

3:48Jeremy O'Brien:And many of those things we now understand are very important to the world. Tell us a bit about what quantum computers can do. Why does this matter for the world and for our listeners? Yeah, so in the abstract, without quantum computers, we are constrained to that small subspace of what is possible in chemistry, maths, physics, materials, and so on. So we're fundamentally constrained without it. And with it, we open up a whole new sort of green, you know, virgin territory of new things that we can do. You know, it's interesting, I think governments are actually, in my experience, and, you know, we have engaged very closely with DARPA over the last several years as part of their quantum benchmarking initiative, which, you know, has involved them sort of testing our technical things.

4:43Jeremy O'Brien:We have a, you know, a sort of billion dollar scale partnerships in Australia and in Chicago to build these first million qubit scale systems. And we've engaged with governments around the world beyond that. And it's interesting to me that principally governments see this as something like rocket fuel to the economy, right? Like that's where their interests principally lie. And that's because the impact of the technology is going to be very deep and wide, so very far reaching in its impact. So, you know, to give you some examples, as an organization, we work with, you know, aerospace, automotive, finance, pharmaceutical, chemical, semiconductor, materials companies.

5:37Jeremy O'Brien:So, you know, broad spectrum of companies that we work with. What are you most excited about, Jeremy? Out of all those examples you've given, what are the most exciting use cases and the ones that can come first? So I'm very excited about energy, sustainability, healthcare and agriculture, which all kind of coincide a bit together. And I'll pick one of my examples that I like because it matters and I think it's interesting and it's concrete. So with Böhringer Ingelheim, a big German pharmaceutical company, we've been working on the problem of P450. And I'm sure many of your listeners like me prior to this work had never heard of P450.

6:23Jeremy O'Brien:That's an enzyme in our bodies that's responsible for metabolizing something like 75 % of all therapeutic drugs. And we don't know how it works. So you can imagine that that's pretty problematic for the drug discovery process to not understand how the enzyme that in all likelihood is going to metabolize that drug is going to function, if it's going to function safely, what the byproducts are and so forth. And a big part of the reason that we don't understand how it works is we can't simulate it on any conventional computer that we have today, nor on any conventional computer that we could ever build.

7:03Jeremy O'Brien:And some of my colleagues generated some numbers that would take something like 10 to the power of 100 years on a conventional GPU system to solve that simulation problem, whereas it would take minutes on a quantum computer to solve. So that's to the point of taking something that's absolutely impossible. So many, many, many ages of the universe is impossible into minutes and you can immediately start to infer, well, that's just one very specific problem across a multitude of problems just within the drug discovery process where you'd have a huge impact and similar applications. We're working with Lockheed Martin in aerospace applications similarly with Airbus on computational fluid dynamics, Mercedes-Benz on battery chemistry, So new, better batteries with Mitsubishi Chemical on new energy efficient materials, JP Morgan on financial service use cases and so on.

8:05Jeremy O'Brien:So you can see it's very broad. And then within that breadth, you know, if you take a particular industry, it's also often very deep. So that's a great segue, I think, to your vision for sci-quantum. But just before we dive into that, maybe to summarize, classical computers, as we know them, rely on classical physics and have been a fantastic tool, obviously, for the world. What you're trying to build, a quantum computer goes a bit of a level deeper and simulates quantum physics, which is a world where, I guess, multiple states can be true at the same time. So just a couple of quick things on your summary there, which was perfect.

8:44Jeremy O'Brien:Classical is just an annoying word, but that's what physicists mean. Classical physics means physics of more than 125 years ago, essentially, which is Newtonian physics. The physics that describes the world that we experience, essentially, we call classical physics strictly. General relativity is also classical theory, but that doesn't matter too much. And I prefer the word conventional. So a conventional computer is just a computer as we know them to be. Physicists will call it a classical computer, conventional computer. And then you mentioned that a quantum computer can simulate quantum physics.

9:22Jeremy O'Brien:And indeed, it is very good at simulating quantum physics. And indeed, chemistry in many ways is applied quantum physics. And so that's an area of chemistry and materials and so on. But there's also more to it than that. And the way to describe it, I would say, is that it is understood that of all the computational problems that are out there, of all the things that you could think about trying to compute, a subset of them can be efficiently computed on a conventional computer. And a subset can be another different subset can be efficiently computed on a quantum computer. and that second set subsumes the first one so there's a bigger set of computational problems that you can uh solve on a on a uh on a quantum computer than you can on a conventional computer and indeed you know sorry for the for the uh delay here but to to to segue into your your question uh and sciquantum's approach is very much built on that understanding and then a very hard look at what would it take to actually solve problems that are really important with a quantum computer?

10:34Jeremy O'Brien:A quantum computer is, to some extent, it's necessarily prone to errors in ways that a conventional computer is not. So one of the miracles of a conventional computer is that it is digital. And that means there are just two states that each of those transistors are encoding. and um so if if there's you know if if this if the transistor wanders a bit away from uh its low voltage or its high voltage that encodes the two uh states say then you can easily you know latch it back to the right the right state so um you know big feature of the power of of conventional computing and its pervasiveness in the world is that it's that it's digital um such that the error rate in our laptops and phones and so on is incredibly, incredibly low.

11:30Jeremy O'Brien:And you don't need to worry much about error correction. That's not true in some memory cases, but for CPUs and GPUs. You've got quite a unique approach to trying to solve this fault tolerance, which is leveraging photonics, which you feel quite strongly is the only plausible route to building a million quibbit error-corrected systems. Maybe explain a bit why you chose that approach over all the other potential approaches to quantum. Yeah. So the first thing is that a quantum computer, unlike a conventional computer, is in some sense it's a hybrid digital and analog system in the sense that when you measure a qubit, you measure a zero or one.

12:09Jeremy O'Brien:So that's just familiar from bits and transistors which are digital. but it's also analog in the sense that that qubit prior to its measurement can take a continuum of values something like a continuum of values between zero and one using this language and so on so that analog feature makes errors uh you know intrinsically a challenge for quantum computing and the idea is exactly the same as uh let's you know radiation hardening uh conventional computer chip, which is that you use redundancy. And it turns out that in a quantum computer, for any type of quantum computer that you build, you need a lot of redundancy.

12:50Jeremy O'Brien:You need of order 10 ,000 physical qubits to encode each logical error corrected and therefore useful qubit. So 10 ,000, 10 ,000 to one. And so if you do some quick arithmetic, you quickly convince yourself that you need about 100 logical error corrected qubits and therefore you need about a million physical qubits before you can do anything of real value and that's what we founded PsyQuantum on was that understanding that yeah you really need to reach to a million qubits before you can really solve important problems and so we sort of work backwards from there so So if real value comes when you have million plus qubit systems, well, the only way that we could see, and I mean for 25 plus years, the only way that we could see that that was going to be achievable was by leveraging the semiconductor industry.

13:52Jeremy O'Brien:You know, the industry that back to the start of our conversation, yeah, routinely makes billions of chips per year with billions of components on each chip, right? Like that manufacturing capability is really second to none. And I sometimes provoke people by saying it makes every other human activity look a bit like scrabbling around in the mud with sticks and stones. I mean, it's truly technical people should be wowing themselves on a periodic basis again, reminding themselves how miraculous it is that humans traveled up an exponential curve for the better part of a century to create this incredible capability in manufacturing that we have taken for granted for 50 some years.

14:38Jeremy O'Brien:and of course it took you know by some count trillions of dollars and the better part of a century and some of the smartest minds of the planet to achieve that capability so for us that was step zero right like you have to you have to be leveraging that and not just the chip part of it but the whole associated um supply chain and contract manufacturers and so on the same some organizations that turn, you know, NVIDIA GPUs into 100 ,000 cluster supercomputers today, right? Like all of that, you know, wiring the whole thing up and so on. And so, you know, we were at that for a very long time. And then we believed, you know, around a decade ago that we had uncovered a path where that would be possible using photonics.

15:25I know you're very excited about how quantum can be a silver bullet in the fight against climate change. Why is it potentially so powerful on climate?

15:36Jeremy O'Brien:Yeah, well, firstly, a quick sort of caveat and caution. I'm very careful to explain that, you know, I don't want people to think that there are silver bullets for solving climate change. I want people to think that there's, you know, we've got to do everything all at the same time starting yesterday, because that's what we need to do. And I also think it's important that people understand that we've had the technology to solve climate change for decades. What we've lacked is the, you know, the political, social, economic will to implement those changes because they're not, you know, they're not easy and they're not simple and so on.

16:08Jeremy O'Brien:And so the role of technology in general is to lessen that, if not turn it around, right, like to make it more economically favourable. So that's just sort of the generic role of technology. And then, yeah, quantum computing could have a really powerful impact across, you know, I mentioned that battery chemistry, but, you know, whether it's next generation solar cells based on perovskite materials that are, you know, cheaper and more efficient, hydrogen production, ammonia production, even steel and cement production, so on. I mean, this is like most of these challenges that we face in, you know, in energy transition and sustainability in fact in our human activity on the planet is a result of not really being able to do that chemistry and materials that is the underpinning foundations of it all i'm i'm in the habit of provoking uh you know professors of chemistry and anyone else who wants to listen with the claim that we can't actually start to do chemistry until we have a quantum computer it's designed to be provocative but also true in the sense that if you can't simulate the building blocks of the world you find around you, like P450, nevermind the building blocks of the world that you want to engineer, like perovskite solar cells or catalysts for every industrial process that we currently pursue, perovskite solar cells, you name it.

17:38Jeremy O'Brien:What are you doing? I think that's an important insight for people because many people think that AI is going to do something similar for what you just described for chemistry, whether it's for drug discovery or novel materials. We've backed a company here in the UK called Cusp AI, very exciting doing AI for novel materials. But you would argue that AI sort of approximates chemistry, whereas the quantum computers simulate it at a fundamental level. And therefore, you're going to be able to have a type of discovery that AI will not be able to, at some point, it will run against a wall. Some people would disagree with that and say that AI is improving exponentially.

18:14Jeremy O'Brien:No, no, I totally agree with that. Yeah, I totally agree with your description there. And it's important. It's a really important one, I think, and probably of interest to a bunch of your listeners, I'm sure, is to understand this AI and quantum computing, and what's that all about. And to elaborate a little bit on what you said, AI in general is a tool that takes vast quantities of data and infers and makes predictions on the basis of that data by, you know, seeing patterns that no human could hope to see or et cetera, right? And those, and the output is an approximation that depends on the, you know, on the exact details of the data.

18:57Jeremy O'Brien:So if you're a drug discovery company, you might be getting some good early wins with AI now by taking the vast quantities of data that you have on molecules that you've synthesized and measured properties of predicting new molecules. But because the answer is only an approximation, you can't be completely confident that that predicted molecule will really have the properties that the AI is telling you that it will have. And so ultimately, you'll have to go and synthesize that molecule and measure it, which is time-consuming and expensive. On the other hand, And a quantum computer, as you said, in some sense does the opposite, not that there's really an opposite of that, but in some sense does the opposite, which is it takes no data.

19:45Jeremy O'Brien:It does an exact first principles calculation of, for example, that prediction of the AI. So, you know, let's take that molecule and do an exact first principles calculation of some property of it and confirm or otherwise that prediction. And I can have extreme confidence in that because it is an exact first principles calculation. It's sort of generating, you know, effectively perfect data in silico. And so you can already start to see how these things will work together, because obviously now I can start to think, OK, well, the experimental data that I've got, I can test it, check with it with a quantum computer that it's useful.

20:27Jeremy O'Brien:I can start to generate data that spans a much, much bigger space of molecules that we've never synthesized before, we've never even had the chance to, can start to restrict things by containing particular elements or whatever it is and so on. And so you can see how these things are incredibly powerful independently and then even more powerful in operating in tandem with one another. Everyone's very excited about the impact of AI on the world. Do you think quantum will have a bigger or smaller impact on the economy and on humanity than AI? Quantum computing has some features and some differences, I would say, similarities with AI where it's also very pervasive, right?

21:11Jeremy O'Brien:Like it will also be very pervasive, so very broad and very deep. Predominantly, it's going to be in the back end of a little bit like, you know, semiconductor chip manufacturing, right? Where not many of us are making semiconductor chips, but boy, are our lives affected by that. Whereas AI, you know, it's affecting things in the back end, but we're also engaging with it as individuals as well. I think that ultimately quantum computing is going to have a bigger impact because of that fundamental fact that without it, you know, I sometimes quote my colleague, Mike Nielsen, who is formerly a professor of physics and wrote the textbook and the reference book all in one on quantum computing, who says, if we were to ever meet aliens, no comment about whether they're out there to be met or if they are, what the probability that we'd meet them is.

22:14Jeremy O'Brien:If we were to ever meet them, we already know something about them, which is that they're using quantum computers. now this is a very you know measured individual not prone to hyperbole and etc so why would he make such a crazy claim and the reason that he he does is because there's a sense in which humans without quantum computers are stuck in this small corner of what the universe or the laws of physics however you want to think about it enables across mathematics physics chemistry materials and so on So ultimately, that unlock is truly profound. And we already see how that's, you know, I've given you just a few examples of how we see that unfolding in some really important domains like the energy transition, like, you know, our own health.

23:04Jeremy O'Brien:And so it goes. Yeah. And Jeremy, we had one guest on Giant Ideas, Mike Maples, the very successful venture capitalist. And his advice to founders and to investors, actually, was you've got to be living in the future. You've been living in the future for 25 years building quantum. And I guess the future you were living in has now become the present. But you do have an inside view onto the future of the world with quantum, which I think is going to profoundly change things. And I don't think people are watching it probably closely enough because everyone's attention is on AI. But we've got this quantum boom coming.

23:35We've got a boom we're living through already with AI. Do you think we're about to enter a world of unimaginable abundance and wealth and societal pioneering developments? How kind of optimistic does this make you?

23:51Jeremy O'Brien:I'm extremely optimistic, partly by nature and partly by rational analysis of things. And I think you're right that, you know, AI is unfolding right now. And I think the world is probably underprepared for the imminent arrival of utility scale quantum computing. So these million qubit scale systems. And as you know, there's a lot of organizations, you know, nations, et cetera, out there who felt like AI took them a bit by surprise. Now, you and many of your listeners will have been engaging with AI for a decade or two ahead of ChatGPT. And I think that's often the feature, again, of emerging technologies is that they're emerging and they're emerging and they're emerging.

24:48Jeremy O'Brien:And then suddenly they appear. And I feel like that's sort of what's happening. That's what happened with AI. And I think that's what's happening right now with quantum computing is that, yeah, I mean, I've been, you know, at it for, you know, a quarter of a century or more. And, you know, people have been hearing about it and so on. And then suddenly it happens. And so I think that preparation thing is really important. And I think it's important because the impact is so, you know, it's like the problems that we can tackle are so important across energy, health care, agriculture, you name it, right?

25:24Jeremy O'Brien:Like we need to urgently do things differently. And so I think getting prepared for that is firstly, there's some technical work to do. There's obviously also some supply chain and business and so on. And there's an urgency to do it. And then sort of, you know, to get at your question in a bit more detail, like if you want to imagine a post-scarcity world, so a world where, you know, our energy and food and materials construction wasn't a big part of human activity, right? Like at the moment, that's the main part of our economy, obviously. In a world where that was a small part of what humans did, that world, it's a bit like the aliens with quantum computers.

26:10Jeremy O'Brien:It's kind of hard to imagine that world without a quantum computer. And it becomes pretty easy to imagine with a quantum computer where, you know, suddenly you have, let's say, mastery over physics and chemistry and materials and so on. where and by the way I always appeal to mother nature or uh whatever uh however you like to think about the natural world where you know we humans use the same chemistry set as mother nature and we sometimes fall into the mistake of thinking that we you know we're very advanced and we have all these great things that we can do I don't know like when I look out the window and I see the natural world I see structures growing from small seeds that drop from the neighboring structure where all the material in it is harvested locally, where all the energy is harvested locally, where when bits fall off it, they grow.

27:00Jeremy O'Brien:You know what I mean? Like this should be inspirational for us to think about how the world could be if we were even a fraction as adept as Mother Nature is at doing chemistry and all of the associated things. And in fact, we should be able to imagine doing better. That's an exciting future to imagine. Can we get specific on timelines? Because Quantum, a bit like Fusion, is always famously 10 years away. And it's having a real moment. Right now, companies like Rigetti and the public stock markets, huge valuations. But what is your best guess on when this future will actually arrive for people?

27:38Jeremy O'Brien:Yeah, so I can speak only for our organization and I can say that we're bringing online those million qubit scale systems in just a handful of years. And that's why I really see the urgency. And indeed, that is driving a lot of our work with our customers and partners to really do the hard work that needs to be done to utilize those things for all of these applications that we've talked about. And what's the blocker right now? If you could raise$100 billion tomorrow from nation states, would this supercomputer that you're building be live and ready to go? or what's blocking the future being here today?

28:24Jeremy O'Brien:So for us at SciQuantum, there essentially isn't a blocker in the sense that we spent a lot of time on the scientific part before we founded the company. I give you the short history of SciQuantum, 20 years of university research to figure out a path whereby we could leverage the semiconductor industry to make a quantum computer. And then we spent several years and several hundred million dollars getting into that industry because tier one semiconductor manufacturing fabs aren't in the business of doing small research projects with little startups or so on. And then we spend a whole bunch more time and money.

29:04Jeremy O'Brien:So the balance of, you know, a decade and a billion dollars, let's say, to be at the point where we're mass manufacturing the chips that go into data center like facilities that we're building in Australia and in Chicago. go. And as I said, their data center like facilities, so you know, to first order, they look a bit, you know, big, big building with a bunch of cabinets in that are a bit like the 19 inch racks, all of them are filled with silicon chips and so on. That takes time to build. And so that's where we're at now. You know, we're building those things. We broke ground at that site in Chicago in September.

29:49Jeremy O'Brien:And so, you know, work is underway building them. So yeah, it's very exciting. And we're building the infrastructure at the same time as finalizing the technology that goes into them that, you know, everything arrives. And at the same time as we're working with all of these customers and partners to identify the highest impact problems that will run on those very first systems. So it's very much exciting times. So we like to end the Giant Ideas segment of the podcast with a little bit of future gazing. So what is something that you believe strongly about the future that almost no one agrees with you on?