 My great pleasure to have esteemed colleagues in this panel. Let me first introduce the panelists, and then we'll have an opportunity to have a dialogue with each other. And then I'll also open it up for Q&A from all of you. So first we have Dave Bacon, who is a software engineer at Google, and a self-described quantum bridge builder, which is a very cool title, like it. Andrew Childs is a professor at the University of Maryland, and co-director of the Joint Center for Quantum Information and Quantum Computer Science. Professor Richard Joseph is a pioneer of quantum computing, of course, and a professor at the University of Cambridge. Also, Eleanor Greifel from NASA is a senior research scientist and the lead of the Quantum Artificial Intelligence Laboratory at NASA Ames. We have Mathias Stefan from IBM, who is an IBM fellow, our chief quantum architect, and Jun San Kim, a professor at Duke University and co-founder and CSO of IonQ. So the central theme of the conference that has been an active discussion on approximate quantum computing and what are we going to do with quantum computers before full tolerance. And of course, the field has advanced enough that now we have finally been able to realize working approximate quantum computers. So I wanted to start by asking each one of you, what do you see coming in the next five years? What is this horizon ahead of us? It seemed like that some of these classes of computers were always far off, and that horizon has been pulled in. But what's the next five years going to look like? So maybe, Dave, I'll start with you. OK. Yeah, so first of all, let me say I've been away for a quantum for seven years. It's nice to be back, but what the heck have you guys done with this field? It's crazy. But I was thinking about this question, and my role in the last six or seven years has been as a software engineer. And at Google, we have software engineers, and we have product managers, and that makes up a huge number of people at Google. And when you're a software engineer, you're working on some product, and you get sort of marking orders. We're going to build this product, and you're sort of down in the woods, coding, coding, coding, and then your product manager comes over and is like, so could you just change this a little bit this way? Can't you just do that? And of course, the answer is always yes, you can, because it's software and it's malleable. But as an engineer, you're kind of like, ah, why are you doing this to me? And one of the things that I think is going to happen in the next three to five years, and is already starting to happen, is a similar thing that you guys may have observed before, which is between the experimental side of the world and the theory side of the world. If you ever encounter an experimentalist out in the wild, and you're a theorist, what you do oftentimes is talk to them about their experiment and ask the following question, can you do this for me? Or how about that, this other thing? And it drives experimentalists absolutely bonkers, which is why they hide away, and we put them at the end of the building and keep them in basements. And I think that's about to shift, and the IBM quantum experience is the first example of a shift that's occurring in which there's now going to be an interface between this. As the quantum computers grow to the next size scale that everybody's talking about, 10s, now 100s of qubits with better and better fidelities, there's a nice decoupling that can occur where now all of those things that you wanted to ask of your experiment and the list, they are going to tell you the following. They're going to say, yes, run it on our machine. And I think that decoupling will allow the people who, especially in this conference, who are building the new algorithms to start doing exploration. And I think that's what is going to be a major shift in how the field approaches the problem. There's going to be lots of graduate students are hacking away at these machines, seeing what they can do. And fundamentally, that's still the question we have. In this conference, we've seen simulation has made tremendous progress in the last seven years. But, and, you know, Aaron, for instance, has heuristic algorithms that we want to run. But we don't know what's in that regime yet. And it's a discovery. Discovery is a pathless land. We don't know what's there. And so I think that decoupling is going to allow that process to occur. And that, prior to now, that's never been the way the world works. And I think that's really going to free up people in this audience to start actually doing really interesting heuristic algorithms to see what's in that space. And I'm looking forward to seeing what everybody builds. Yeah, very interesting. So you see a big expansion of the community of who's going to be able to touch these computers, experiment with them, and explore with them, and innovate with them. That's right. That's right. And even just the people who are, you know, it does take, there is an overhead to be a theorist and to understand the quantum circuit model, right? You know, if I have to teach it to my fellow software engineers at Google, it's a challenge, right? I think that bridge will be there. But I think that just getting more and more people in an exploratory space is going to, you know, if there is a there there, it's going to be people, you know, lots of people exploring that space. Now I understand better your title of bridge builder. That's right. The goal is to be in between, right? Like somehow. Great. So Andrew, what do you think for the next five years? Well, I think these are really exciting times. I mean, it's amazing the progress that's happened. And I think over these coming years, I hope we will see demonstrations of quantum computations that look like things that we don't think we can do with classical computers, so quantum computational supremacy experiments. I think we can see that coming. I hope that we'll see also better quantum simulations, so probably better analog quantum simulations as well as hopefully, as I was trying to discuss a bit in my talk here, maybe even early digital quantum simulations. Perhaps not quite yet at the point where there are things that we don't think we can do with classical computers, but hopefully moving in that direction. So I think those kinds of sort of working towards applications and showing we can actually use quantum computers in a useful way is an exciting potential development. I think at the same time, over the next five years, I guess if that's the horizon we're talking about, I think it's important not to lose sight of the really long road ahead, because I think getting to quantum computers that can really do useful things and that can really explore quantum mechanics and a high complexity regime, it's a long game. And so we're going to have to work, I think, for more than those five years before we have such really useful devices that explore the full potential of quantum computers. So hopefully over the coming years, we also can take a lot of time to do maybe less, so things maybe where there's a bit less glory, but maybe things that help to pave the way for what we might do in the future where we characterize the era models of our machines and develop better algorithms and develop better fault tolerance protocols. So we really are set to have sort of a long-term path to really realizing scalable quantum computation. Yeah, and we do have some precedence in the past in the computer industry, famously in the semiconductor space, like the sustained roadmaps where you have to work on multiple generations of technology at the same time and a combination of what's two years, five years, and beyond. And now later I would like to explore with the panelists like this sort of balance between science and engineering and where do you see that horizon moving. So Richard, your perspective of what's ahead of us. Well, I guess my perspective on the whole subject is rather different from that of this conference. I mean, I have no real experience with industrial or applications, useful applications-oriented things. And my perspective is academic, fundamental science, and just ivory tower of blue skies research. And well, no, but I think it's important because I think it's getting forgotten a bit. The imminent appearance of this fabled quantum computer has taken the whole limelight of the subject, and you can be forgiven for thinking that there's nothing else in quantum computing now except building this machine and getting it to work. And I think there's a lot more in the subject. There are pure theory sort of aspects like complexity questions, like P versus NP for quantum complexity class type things. And these are really important and just as exciting. Sure. So what's ahead in the theory side for us? Yeah, OK. So no, but one point I really wanted to make was I think for me, the really most exciting possible thing you can do with your machine has not even been mentioned in this conference, I don't think. And that is to verify or refute quantum theory. Now, I don't know how many people here believe that quantum theory is actually ultimately correct. And it will be the final physical theory. Now, I don't believe that. And this machine is the perfect tool for exploring issues of that sort. And I guess the obvious target is like the measurement versus unitary evolution business. And maybe even with a 50 qubit cat state might already reveal peculiarities of superposition versus collapse. So and interestingly, the goalposts and the main thrusts of this point of view are curiously almost opposite to supremacy goalposts. I should say also another thing I dislike about this focus on supremacy is obviously important for many things, but it has no scientific significance. Because the strength of our current classical computing is essentially arbitrary. This year's supremacy is next year's classically doable and you build a better computer. But so for testing quantum mechanics, what we'd be interested in is not surpassing the classical simulatable barrier, but trying to implement very large scale quantum processes which are still classically simulatable, exact opposite, right? Because there's no conflict there with testing quantum mechanics. There's no kind of sense that because you're classically simulatable, you're not probing quantum theory as well. So I think that would be not only exciting, that would just be unimaginably awesome if this machine could precipitate the next great paradigm shift in physics. No, that's a great perspective on that. Yeah, that's right. And I guess another point I want to make a different point is that the whole conference has been talking about using this computer for essentially standalone computation, like demonstrating supremacy. It's all about computing power and versus classical power, which is obviously important and especially safe for IBM and commercial companies. But there are other quantum benefits too. So if you think of, say, BB84 or teleportation, the Elson-Bob's computational ability is very minimal. You can classically simulate it on a scrap of paper, right? Or even in your head, practically. Yet it's still important, it's still valuable to implement it because you gain something else. There are further quantum features of state disturbance when you try and get information or this kind of thing, which make that valuable. So I wonder if there are communication or cryptographic tasks which require Elson-Bob to have, say, 30 or 40 qubits in order to implement them. For example, you might want to do a collective attack on a BB84 transmission or maybe some kind of secret sharing state thing where you have lots of parties. So you've got a fairly decent sized quantum state you need to process or something like that. So there's more to quantum life than computational power that can be useful in applications as well. So I think that's another point that seems to have not really been thought about very much. Yeah, no wonderful perspectives. That's very helpful. So Eleanor, I'll ask you the same question now from NASA point of view and your own experience and your own center. What are you most excited and what are the areas where you're going to be pushing for the next five years or so? Yeah, so what I'm most excited about is actually being able to run things on machines and being able to test algorithms that we can't test quickly and that the regime that I think starts to be interesting for this is not the numbers that Ashley gave at the end, but where you can ask what can you do in a microsecond that might take a day because you can't do fast iterations if each time you run something takes a day. But you may still be able to, that doesn't quite get us to the supremacy area, but it enables us to experiment much more effectively with algorithms. And if you look at what's out there classically, most algorithms that are run on supercomputers, you don't have a proof that that's the best algorithm out there, you don't even have a mathematical proof that it's better than last year's algorithm, you try it out and see what happens and there's constant competitions to find better algorithms. And that's only something that the quantum community can do now. And the way I like to put it is that this should end up having a huge blossoming of algorithms in quantum computing because what's the chance that any place where you have a quantum speedup is a place where you can prove by pencil and paper that you have a quantum speedup? In the classical case, this is rare. And so we know that there are speedups in the quantum case and we would expect that to expand out. Now what does that mean for what we should be doing? I think one of the scariest things is that means that we should have a certain number of us or new people come in and do quite different work where instead of trying to prove things, we sit there and try to come up with reasonable hunches as to what would be a nice thing to try. And I think that this can tie in actually with the fundamental work that Richard was mentioning before, which is you really want to figure out what are pieces of quantum mechanics which might give you a quantum mechanical advantage. And even if you won't be able to take advantage of them to solve a practical problem of interest in the near term, you can run a proof of concept experiment, see if you can take advantage of that piece of quantum mechanics and have a new algorithm that maybe you don't have a proof that it could do something better, but you have some evidence that it can do something better and that that will really broaden what we can do with these machines. Yeah, so you're very excited about experimental, heuristical approaches to be able to exploit what the machines could do and see how our intuition combined with that axis maybe give us an interesting direction. Yeah, and one other thing that I think we should really concentrate on is hybrid quantum classical algorithms. Yes, I agree, yep, yep, yeah. So Matias, you know, I'm gonna pretend I don't know your views at all. Yeah, you probably heard us wrong before. So time is what's ahead. So needless to say that the hardware needs to improve from where it is currently, I mean hardware is something I've been doing for many, many years. We know the implementations that have been done so far, they're errors and we're not simulating some of the quantum chemistry results that match exactly the prediction, so we need to improve the hardware, there's no doubt about that. For look at Andrew's numbers he talked about yesterday, so even if you add a few nines of fidelity to our gates, we still need hundreds of millions, billions of gates, that's pretty depressing. But there have in the past been many improvements in the fall-tall and threshold over the years that kept coming down, that gate count will come down, and maybe there's something in error mitigation in Ed Farry's talk, he mentioned that the point is maybe some of the errors don't matter, depending on what the observable it is that you're trying to measure, maybe some of the errors just don't matter. So some ideas about error mitigation, how to deal with these errors, even if they may be not fall-tall, that's fine, so as long as you're still computing something meaningful. So those are two almost complementary directions because the errors in the hardware will dictate a little bit the direction of where error mitigation is going to head. And then it's benchmarking, Micah mentioned this yesterday on our hardware, you have the spectator errors, if you do something on two qubits, what happens to the other n minus one qubits? How do you do this? And then Dave mentioned this, there's a bridge between theory and experiments. So if I wanted to run some of these things, like three qubit randomized benchmarking that Micah also mentioned yesterday, I don't want to go in and just code all my arbitrary waveform generators and microwave generators, I just want to have something where I can go into Qiskit and say, I want to run randomized benchmarkings on qubit 314 and 19, run, give me the result. And that's great for an experimentalist to have, I can just go in and do that. And meanwhile, the theorists can then look at the results and hopefully infer meaningful information on what to do in the future. So having said all that, and this is what something you touched upon is benchmarking. So what do we mean by something ran better on a quantum computer? Do we mean the answer is more accurate than we could solve classically? Is it faster, but maybe not quite as accurate? And I guess it depends very much on the application you're trying to solve. They can very well be applications where you want incredibly accurate answers. And I don't mind so much if I wait weeks for the answer to come back. In other cases, I want something in seconds. And even if it's just mediocre, I just want it fast. So what do we mean by a quantum computer ran better? And what are the appropriate metrics? And that's going to be a challenging topic to address, I think, but it's something that's gonna have to happen moving along the future. So I think those are my two cents. Okay, very good. So, Jung Sang. All right, so I would like to, well, Dave mentioned about this experiment to leave me in fear, is that I'm actually, I just try to put both my heads on and talking about academics meeting industry in this field, which has not really, I think a lot of the R&D that's still going on is very much in the realm of R&D rather than an actual product, just because nobody knows how to make real commercial value out of this. But I think there are pretty interesting prospects of what might happen. I think one is Dave and I talked about being fools over lunch, because each and every one of us who are in the field have to believe that there's something we can do. Exciting and fantastic. And you have to be convinced that what you're doing is really the best approach than others. And I think that's, we need a lot of those very intelligent yet fools who will go in and put your time and effort and put things on the line to get it done. And I think that's important, but it's also very important to accept the fact that maybe there are things that I don't know that maybe Mathias knows much more about and so on and so forth. And then, and I think that really allows us to really make a collective progress, even if we're in a position where our approaches are different, maybe, maybe competitive, but I think there's a lot of issues we can collaborate to learn, especially this benchmarking techniques and how do we characterize the PowerPoint computers and so on. Now, on the flip side, I think there is quite a bit of a very heightened expectation that we may be able to do something really important and useful. And yet, I don't think we, any of us really know what that thing is, right? As Richard mentioned about. So there, I think there will be probably some as the rubber meets the road, meaning as people bring up real systems where we can, some of these real interesting ideas can start to be tested, we may be surprised or maybe some complexity problems so that the quantum mechanics may not scale as we expected or maybe some approaches we felt was very solid may have blindsided by some problems we didn't know about today. So I think there will be a lot of this shaking out of what you believe in has worked out and what hasn't. And I think that's a good thing. Some will be disappointed than others, but I think all of that reckoning will actually give us a better insight of where to go in the future. And I think that type of new discovery in this complexity area of quantum, I think it's going to be some of the really interesting discoveries that's yet to come. And on the flip side, and in order to keep that going, I think it's really, really important that our field continues to maintain the excitement so that the young brightest minds feel like jumping into this field is more exciting and interesting than going out and doing something else. And I think that talent, continued attraction of good talent is really the field that will keep us going and look for these breakthroughs that will really make some of these effort really useful. So I think it's really important that we all keep up that momentum and keep attracting the brightest minds to come and work with us. So let me segue a little bit and perhaps on your theme that you brought up and it's been brought up David, you were talking before about also like community and other stakeholders that we want to bring along. If I could ask all of you to reflect if there was another community or a different scientific discipline that you think would be important where we are right now for the next coming years to sort of start bringing on board because they either bring a skill set or a perspective or a requirement that you think is important to consider as we continue to advance, what would that be? For your own work, what kind of collaborators or colleagues would you seek in the coming years that you think will be particularly productive? We've talked about software engineers at one extreme of the application where you're trying to apply it more but you were talking about perhaps that we can use it as a scientific instrument in its own right to test theories. So who would you like to reach out to? I could take a good funny first yes which is that at Google everything is solved by machine intelligence. So obviously we just need more people in machine intelligence to optimize our problems. But more seriously I do think the community will start to, as a lot of groups are starting to go with the transition from physicist to engineer, and there is definite difference having made that transition myself into become a software engineer of the engineering mentality of how you structure, how you're going to have a team that builds something large. And so I think pulling in people who are coming from engineering departments or coming from engineering in industry from other parts of the company, pulling those in is going to be a huge change in a lot of the tone. That's already occurred to different extents in different groups, right? But I write a lot of code every day and if I write code that I see written by a physicist I have to tell you I sometimes pull my hair out. And I think that that's going to change a lot of the, there'll be a maturity of the tooling that you're building. Of course at the same time you're doing the hardest experiments in the world to build the most important thing in the world. And so there's a tension there because you want to move fast as well. But I think we need to bring in great engineering practices or if you've read The Soul of the New Machine you'll end up going nowhere. Okay, so. To follow up. Who else? Also related to that I think we want to bring in domain experts but and people who are really doing the top of the line algorithms for applications that we're interested in but we need to find the right people because they're going to be the people who want this to solve their problem better tomorrow and they won't have the patience but finding the people who find this area exciting would like to collaborate and are willing to look the long distance. I think will just be a huge help to this field. And some of the greatest pleasure of working at NASA is to have found a couple of those people to work with and we're, I think the whole field will benefit from more such people coming in. Okay, no, that's very, yeah. I think that that's right, right? Because we're going to have to be able to bridge these communities together and see what is that continuum of what you can do right with different approaches and techniques, yeah. Other community? Maybe I can make a kind of related point maybe not exactly directly answering your question but that I really think that our area is already a pretty big tent. I mean, quantum computation is about sort of using physics to do computation. So it kind of naturally brings together physics and computation. And through meetings like this one, we have people from both sides, physics and computation and also other areas of science who are all interested in these questions. And now because this area has been going for so long, I mean, it's kind of routine at these kinds of meetings that you hear physicists saying, certainly we can't solve that problem because it's NP hard, right? So somehow this concept from computer science has just been brought into the language of physics and the other way around, we see people using ideas from physics to think about computation. So I think that's a fantastic thing that to a great extent has already happened and I think we can kind of hope to bring more people into our community and bring new kinds of expertise. But I think it really is an area that in some ways is ideal for bringing together people from different disciplines. Yeah, no, you're right. And it is already a big tent, right? But I think we're gonna see a very significant expansion in the coming years. And it'll be interesting to also to see how the conferences and that dialogue evolves and the language that gets used associated with that. And I think just to the point of the domain experts, I think one thing to be very, what we're really looking for also are people who are skeptical in some sense, but skeptical, but movable, right? Because I think there is a danger that when you bring in somebody to work in a domain that and merge quantum computing, all they do is think about how can quantum computers solve my problem? And you need a mentality that is, I want to do the best in the world for that domain, what is that? And so that's a different type of, I can be a domain expert in an area, but not necessarily a domain expert who will also approach it in sort of a red team way, which I think is a very important distinction. So, oh, please, please. And having a sort of white hat, black hat for the algorithms so that you have people trying to push the quantum algorithms forward, but also trying to push the classical algorithms forward. And everybody views that as a team effort. We don't hate each other. We want to build off of each other. And at the very least, we'll come up with interesting, we'll continue to come up with very interesting classical algorithms as well as quantum. Yeah, let me just say, I mean that that's actually one practice that within IBM, we believe very strongly, right? Of continuing to push the boundaries of what you can do with simulation, with classical computer simulation and absolute limit because as you were saying, that's a moving frontier, right? It keeps getting better and better and we're going to get better and better at it. And we're going to push that to the absolute limits and forever. And then, but we're also doing the other one in parallel. But it's not a nasty competition. It's like that's how computing advances and moves forward, right? And the benefit of the two worlds make everybody better. But with classical computers, when they first emerged, I mean, no one had any idea what to do with them, right? That's true. It was meant to be just for solving partial differential equations by numerical methods and things. And no one thought about computer games or internet or email or word processing. And now everyone does it, right? So do you think when we have this quantum computer, then a similar kind of creative explosion will occur? And I think, I mean, and this has been touched a little bit about it. I think that that's what's going to happen also in the coming years that we're going to expose this new tool and new capability. And people are going to approach it creatively too in ways that we don't anticipate and it won't be driven by proofs and so on. And people will just try things out and see what's interesting. So let me ask then all of you, I mean, since it was alluded to about like some of mistakes or pitfalls that we could encounter, I'd love to hear your perspective of what should we avoid? I mean, what are things that you may be concerned about right in the coming years that you would say, let's not do that. Let's not make that mistake. What would you say? Well, I think Junk Singh mentioned kind of hype and high expectations. And I think there's a lot of concern, potential concern about kind of over promising and under delivering. So I think, if we sort of set to near a horizon for when we say we'll be able to deliver large machines that we can use to do practical computations and then it doesn't happen, people will be disappointed. And so I think that's, I think we have to be cautious and sort of, as I was saying before, kind of realize that we have to go the long haul. Okay, so I can follow up in what Richard said. You know, I think the kind of engineers or algorithm people that we're trying to push this forward are not necessarily the people that we see or that exists today. Just because the frontier of what's exciting and out there is just so ill-defined. So I think it's very important for us to keep in mind that whatever your short, near, or far term goals that you set, this is going to be a moving target. It is actually going to evolve based on the time effort and the productivity that we as a community generate, right? So it's not, we're not walking on a well-set path. We're trying to create and understand what that path forward is. So I think the successful talent that we need to pull in is going to have to have an open mind. People like Dave Bacon has said, okay, I'm gonna leave my field and go and become a software engineer for a while. As we said, it's a foolish thing to do, leaving all the things behind. But that is exactly the kind of mindset that we need because whatever we're going to have to figure out how to do is not going to come from, there isn't somebody out there who'll be able to tell us exactly what we should do. I think it's going to, we will need each other's help and the expertise, but we also have to approach it with a very open mind because it's a big challenge ahead of us. So I think making sure that we have that open mind, accepting the fact that yes, I really believe in what I'm doing, but also accept the fact that I may be missing something, I may need help from others to make sure that we continue to explore the realistic and practical opportunities that's ahead of us. I think having that mindset is very, very important. Yeah, I think also moving forward, having the future generation of quantum scientists is incredibly important. Ones that already speak the language of P versus NP and at the same time being familiar with the physics. So I think there's great value in academic institutions in generating the next leadership generation for quantum computing. While at the same time also being able to explore completely crazy new ideas as they apply to quantum information and quantum computing. I mean, we often or sometimes don't see what's behind the exit sign. We keep plugging along the problems that we encounter every day, but there may be other completely novel ideas on how to solve and tackle some of these issues in a much better and different way. So I do believe that fundamentally academic institutions play an actually incredibly important role. This is not just something that should be shipped out to industry, we're done, okay, fine. No, I mean, I think universities really play an important role for future hires as well as, yeah, coming up with the next idea. Following on from that, partly because quantum computing seemed so far out for a while, this is a very open field with everybody talking to everybody. And I've started to have a little bit of a fear of a change in culture because of the market forces and the hype. And I think we all need to really push against that, that the challenges are still so significant that we all need each other and we all need to work together. And that not just the academic work, but the open work is where we need to stay for a long time still. So let's all push against any forces in our opposite direction. Good advice, that's great. Good advice. So I'm gonna open it up in a minute for questions. So start thinking about things that you'd like to ask the panel. But one last question I'd like to ask you before I open it up. So now to the public, right? If there was one thing that you would like the public to know about quantum computing and not have a misconception, that's something that you say, this would be a helpful, either a concept or a way to understand it, that would ground it, right? Against the hype, as an example. What would you share? What would you tell them as a mechanism to understand what's happening? This is your opportunity to tell them how to be grounded. All right, what would you say? Let me actually start. So I was having a conversation with one of my colleagues who is a software engineer all his life. And he basically said he had a dad who was a computer scientist and his dad started teaching him how to code at the age of five before he could add numbers, right? So, and if you think about that, okay, that's weird. But in this day and age, that starts to become a very important skill set that we do want our kids to start thinking about it that way, maybe not all of them, but that aspect starts to become important. And if you go back and think about, you know, how we teach every first graders to add numbers, I'm sure a thousand years ago that was now we taught our kids, right? And I think the way that we approach new science or framework at a very young age, I think it's totally possible. And I think those are the things we have to think about. I think we have to break the myth that quantum mechanics is something that, you know, is crazy and mysterious and can't understand. I'm sure adding numbers a thousand years ago to a general public was a mysterious thing, but we don't think that way anymore because we train people at a very young age that this is how the war works. And therefore if you have those skill sets, then you can actually, you know, contribute in a more meaningful way. So I think the notion that quantum mechanics is mysterious. It is a matter of training. Nobody's born how to ride bicycles. But once you train them, it becomes very natural to them. So I think those are the things that I think we should think about more. Yeah, Charlie's a passionate advocate at that very point that you're making, right? I'll try to make it less mysterious and that when people say, this is something I cannot understand. He said, yes, you can, right? And let's make an effort. It's a great point. So what else, what would you share? I'm sort of the opposite side. Some young people here that quantum computing is about to be there and solves all problems. And one problem with that is that then they think there isn't anything for them to do. And so we should make sure that such people understand that there's a huge amount to do there and that we'd love to have them come in. So there still should be some mystery. All right, so we're balancing things, that's great. What else, what would you educate them on? I mean, certainly I think that there's a misconception about sort of where the power of quantum computers comes from and that somehow, I mean you see this sometimes in the popular press that somehow quantum computers explore all possibilities in superposition and somehow by doing that can sort of solve really hard search problems in polynomial time that basically they can pick out the needle from the haystack. And we know that that's not the case, that we don't expect to be able to solve, let's say, incomplete problems in polynomial time. We need very specific structure to be able to solve problems efficiently with quantum computers. So I think it's something that it's good to get across is that somehow quantum computing is not a magic bullet that solves all problems and it's related to the point that you're making that somehow, and in many ways that makes it more interesting because there's more to do. We need to try to understand what is the kind of structure that we can exploit if we wanna be able to find new quantum algorithms, understand what problems are easy for quantum computers but also what problems are hard for quantum computers. Okay, any other thoughts? Well, I guess for the non-scientific public, right? Right, yes, yeah. I wouldn't attempt to, there's a lot of this temptation to kind of make quantum mechanics really bizarre and paradoxical and so on. And that just confuses people, right? I think, so I would simply say it's just a powerful new resource like atomic energy. It relies on physics beyond, it's real physics beyond our day-to-day experiences but it can be understood if you take the trouble of learning some maths and so on. And it can be used for tremendous benefit but also great harm and that's up to the whim of the user and yeah, and that's all it is. Okay. Yeah, I was just gonna say I think there's an important follow-on to what Andrew was saying that I think it's extremely important to also admit that we don't understand a lot of things, right? So we don't really understand if quantum mechanics breaks down and we don't understand the power of quantum computation, right? Like if I ask you where does it come from it's still hard to pin that down, right? I think if we do that in our public, like oftentimes when the journalist comes to you you say something the journalist thinks you're saying it definitively. We should try to be humble in saying that we don't know and that will give people who are on the outside this idea that there's much to explore and learn which is I think what we all know but that doesn't sometimes come across when we talk externally because it comes across as, you know, this is a solved problem and I think we need to be very careful about that in our discussions with the public. Yeah, these 30 seconds on bites can be very difficult. Explain quantum computing in 30 seconds. McQuade, people have been doing this for decades and they're still struggling. Just walk away. I see, very good. Okay, let's separate it up for the audience. Questions that you have for the panelists and we have David has some microphones I can bring them up. Antonio, come on, some questions. You must have some. Oh, one second. You could wait just a second for the microphone. Sure, you can't stand up. Just stop. I guess you guys have a quantum computing systems that you wanna build in your mind, each of you and maybe I have two questions. So the first question is, do you think that road to get there will be more incremental, adiabatic process or do you think that we'll need a few breakthroughs to get there? And if you think that we need a few breakthroughs to get there. Looks like it's for you and me. That's right, and hardware builders. There you go. Yeah, so progress is very rarely linear. I mean, it goes in incremental steps. You might be stuck for quite a while and you probably remember some of these early days from one of our programs where we had these hot cubits and we just sat around that for two years until we finally figured something out. That was a big step function that allowed us to go to the next level. And I anticipate this happening for future devices as well. And some of these things happened by accident. I think that that particular invention was because somebody wanted to prove the opposite. And so that's, you know, on that particular part it's gonna be very, very step function wise. And when the next breakthrough comes it's impossible to predict. I mean, it happens when it happens. Well, working on it and working towards it. In terms of the system, I don't know. I mean, for Super Nighting Cubits, I guess in the early days, people said there's no way you'll get a microsecond and people figured out how to get a microsecond. Then it was, oh, well, you know, you can't get to 10 microseconds and people got to 10 microseconds. So you hear the same thing in iron traps as well. Well, you have these big iron traps. How are you gonna do anything with it? Well, people figured out, well, we can build micro traps. Well, yeah, there you have problems too. Okay, fine. And when people work on it, they're smart scientists that work out problem after problem and progress will continue. I think the, ultimately the answer will not be what we think it is at the moment. It's pretty much safe to say. But do you believe, Matias, that the rate of progress is gonna continue? I mean, we've seen in the hardware space tremendous difference what has happened in the last four or five years than what had happened in the last 15, as an example. Do you think the pace is gonna continue kind of as it is for the last four or five years, or is it gonna be continue and accelerate? I hope it'll accelerate, especially with the availability of software and getting smart people, hopefully around the world to plug into your devices to help figure out what the problems are and how to make things better without. But I mean, that's a big deal because if you see what happened in the last five years and you're saying five years from now when we all sit here and talk about it, right, that we're gonna see a progress that is much more dramatic in the last five, that's a big deal. I mean, just think about how different the world will be, right, if we have achieved that. So, I think, you know, so people say you tend to overestimate what you're gonna do in two years, but you tend to underestimate what happens in 10 years from now. I mean, I think that's because usually there is a bifurcation point. I think that bifurcation point is somewhere in the next five years for us. The bifurcation point is where all the effort that we're making in terms of building hardware and all the effort we're making in trying to use it for something useful and interesting and important. And, you know, there will always be scientific value. So I'm not discounting that new science will always come out of this. But there is also a tremendous economic point. If you look at what happened to silicon computers and why they were able to keep this exponential pace for a very long time, it's because it can actually do useful things that generate small value. And then there's more investment. It becomes a very exciting field, although intelligence comes in. And then you can actually take a really, really nonlinear increase. Whereas if that, so right now, I think there's a lot of investment based on expectation that this can do something real. And if that expectation is realized, I think this will just continue and accelerate it faster. But if people started to realize that the expectation is not being met, then I think things can fall the other way around. So I think it's really, really interesting. And I think the whole point of this conference, and again, I'm not surprised that we don't have, we haven't yet pinned down exactly what that application and the point is. But I think that's really approaching. That opportunity is approaching. I think that that will happen within the next three or four years, that first discovery of an application that has some amount of quantum advantage. You know, even that is a nonlinear process. So I think it depends on how we approach it. For example, if you think about how all these apps on the phones came about, right? People had the infrastructure, but the approach there is let's give them some API, make it really easy for anybody who doesn't know how to build an iPhone to come in and build an application on top of it. And then you really lower the barrier for a very large number of people to come in and innovate, creative young people. If you let that happen, I think instead of like 20 academic groups who are very smart focusing on it versus letting 20 million people think about what the application is, I think depending on how we search for it, the probability of hitting this is going to be much higher if we let it happen across a much more open environment. So I think it depends on how we approach this. We'll really increase the probability of finding that point, but once we find that point, and then I think it really bifurcates, right? So I think this really, at least in my mind, I think that's one of the greatest contributions that I attribute to your team at IBM is bring this online so people can start working with it. Let that framework work out, let a whole bunch of other people allow that to happen so we can actually provide better performance to a wider range of users in a more competitive environment because I think that will really open up the probability of us finding some real useful thing. So I think that should be some of the very important strategy that we all collaborate with to try to open up. Because if that doesn't happen, this is going to go down. I wouldn't be surprised. Okay, it's a question, then. Antoni, right there, next to Lior. Hello, yes? Is it on? Sorry. Yes, okay. I want to ask a bit of a hippie question and it's not totally serious. So do you think the quantum computer will make the world a better place? And I'm obviously not, it's obviously ill-defined, but I'm a bit interested or curious in your reactions to that question. Let's put it like that. Well, we just heard our comparison and you'll hear our weapons, so. No, not weapons, energy, I'm sorry. But okay, let's hear it. Yeah, what do you think? I mean, do you just see it as just like any other technology? Depends on how you use it or do you have a perspective of? I mean, there is, you're probably, I mean, the biggest thing to talk about right is cryptography, right? And large-scale quantum computers will do, you know, radically define, redefine the security of the internet as it's currently exist, right? I mean, I believe that the other outcomes, when people talk about simulation and what it can potentially do for industries, I think in the end, you bet that yes, it's gonna make the world better. But I do worry a lot about that next world and in particular worry about post-quantum cryptography because just think about the number of people you know who are working on quantum attacks on post-quantum cryptography. It's not a large number. There's like, Sean's here and you have effectively worked, you know, lots of you have actually pushed in that direction but it's not a large group of people who have validated that and so that is a scary thought, right? What happens if it is also true that those are not secure? You know, as somebody who has some idea about the amount of stuff that goes around in our own internal network at Google, that is a very scary prospect. And so we should be honest about that. We also can say that we have at least a little bit of time but I do think we should pour significantly more resources in making sure that those next crypto systems are secure and I strongly advocate that nation-state should be doing a lot in that effort. So. Well, let me try to pose the question back to you, right? So every major technology that's really changed the way that we do things in life, like internet. Has internet made the world a better place? You can always ask that question. I think it has enabled a lot of convenience but all technologies do come with a backlash of problems. And I think it's really a matter of if new capabilities are made available to us. I think, as Richard said, it's really up to us to try to figure out how to use it in a more beneficial way than the other. So I think as technologists, it will enable new capabilities and I sure hope that there are more people who figure out how to use it to our advantage and make it a better place than those who try to make it take advantages of it to make it a worse place. But there is a lot of value that goes into making that decision. Okay, do we have any one last question? We talk a lot about quantitative speedups, quantitative gains, et cetera, even how we define ourselves, I guess, in this area. Are there near-term applications even in a noisy regime where we will find qualitative difference? We've touched on new types of physics that we may be able to explore, et cetera. But are there things that really may be practical applications in a near-term, medium-term, noisy regime that will have real practical applications that aren't just speedups or not? The answer is we don't know. Yeah, I think a lot has been discussed along those lines for the past three days and without any rigorous answers at the moment. This is mentioning this to Andrew. Sometimes it is depressing to wake up and look at some of the numbers that you might ultimately need versus where you are, but things keep improving, people will innovate. So I think something will come around that is very useful, but I can't really put my hands on what this exactly is at the moment. So I'm actually absolutely with Matias on that because I'm, in a way, we're both fools believing that this will lead to something very useful. Otherwise, you know, we will go find something else to do that will actually make it lead to more useful things. And at the same time, you know, I fully accept the probability that I'm completely wrong. But I think until we give it the best shot we have and put us in that place, we just won't know. So is it worthwhile trying to make that effort and fools like us believe that it is and that's why we're doing this. That's a good role for research and taking risks. So I would like to thank all the panelists for taking the time to share your thoughts and your knowledge with all of you. And so we're gonna have an opportunity to close the conference and really thank everybody for your participation. So thank you for the panel.