 Hello everyone. Quantum computing is one of the most hyped but maybe one of the most misunderstood technologies that are currently in development and I'm really delighted to have Ling G from Tencent here to talk us through it. I think maybe we should start with the basics, right? So what is quantum computing, Ling? Well for me, I think quantum computing offers us a window to look through the universe. What I mean by that is in the 1970s, 1980s, when Richard Feynman first proposed quantum computing in the modern day age, he mentioned that the nature is quantum. So at the sub-atomic level, we are all made of atoms, atoms are made from electrons or quarks or something even smaller and all of these things are governed by the quirky, superposition, entanglement, decoherence, etc. So all of these are all operating at a quantum mechanical level and that's how the world really works. Yeah, so I guess it's this idea that when you get down to the really small stuff in nature, as you get down to the sub-atomic level, you get these strange interactions because of these quantum properties of the universe and if we're going to simulate those systems, those natural systems, then we need the computer but we'll do that and that's, I guess, the crux of what quantum computing is and what it's for. But what are some of the applications of quantum that you're most excited about? I think I'm most excited about modelling the world at the sub-atomic level. There are some better publicised applications such as optimizations or doing quantum finance as well as quantum biology. But I think at the end of the day, it's how we model all the materials, properties at the sub-atomic level to look at how the electrons move, how do we describe the world in their true nature. I guess one of the other applications that gets talked about a lot is the cybersecurity implications of quantum computing. Do you think that's a real threat or is that something that gets a bit over-talked about? I think it's a double-sword edge. So quantum computing can make things super secure, unbreakable, but on the other hand, it can also break into systems. Everyone knows about the RSA encryption. So in quantum computing, there is a very famous algorithm called SUAR algorithm and that's to factorise very, very large numbers into prime numbers. So things don't sound that complicated, but to factor, say, a number of 10 billion digits might take the age of the universe to crack it. But once quantum computing is here, it takes a split of a second. On the other hand, a lot of research, for example, in the US, next day, they are setting out a competition, trying to set the standard, having various crypto schemes in order to make things unbreakable. So a lot of researchers are doing post-quantum in order to use classical computing to make the crypto super secure. I think that's a pattern that appears a lot in quantum computing. The headline is really shocking and scary. Quantum computing is going to break RSA encryption and open up all of our emails to everyone. But actually, when you dig a bit deeper, you find that, well, yes, quantum computing can break RSA encryption, but people are already working, as you said, on NIST, on quantum secure algorithms that won't be vulnerable to quantum computers. And I think one of the big misconceptions about quantum is that it's going to be the best at everything. And that's not actually true, right? Quantum computers will be very useful, but only for a subset of tasks. It's not going to be like a supercomputer ties a million, it's something slightly different. Yes, that's totally right. So as probably most of the people know, there is two sets of problems. One is the P that stands for polynomial and NP non-polynomial. And the quantum computing only solves a subset of the NP problems. So lots of the things, A, it's not exponentially faster, mainly in factoring into prime numbers. So that's exponentially faster. In most of the other cases, say, optimization problems, people normally use another famous quantum algorithm, the so-called Grover algorithm. So that's like searching a database, you can make it still dramatically faster, but not exponentially faster. And those are the two algorithms that get talked about, a lot, Shure's algorithm for factorizing and Grover's algorithm for search, as you said. But I mean, there have been a lot of hardware developments in quantum over the last few years. So Google very famously achieved or said it achieved quantum supremacy back in 2018, 2019. And recently IBM said it created a 128 qubit supercomputer. We'll get into what qubits are in them. But how far away are we realistically from something like Shure's algorithm being able to run on a quantum computer at a useful scale? I think while still, yes, all of these developments are super exciting, but I think people should remember building a quantum computer is extremely hard. So I made you mention about how do we build one. So let's go back to the fundamental building block that's a quantum bit, the so-called qubit, which is in analogy to the classical computing, the classical bits, 1 and 0. And in the quantum computer, when we use superposition, so we can have a qubit that's 0 or 1 or 0 and 1 at the same time. And the second important quantum property is entanglement. So that's the two qubits, they can be linked with each other no matter how long apart, how far apart they are. So I can be on the moon and a minute you can be on the earth. Our two qubits can still be entangled together. And then there is decoherence. So when in a quantum system, say when we have entangled qubits, but they are in the real world, they are very messy. They get lots of interference from the environment, from the noise, and the system decoherence. So they are no longer in an entangled state. That's why quantum computing breaks down in a way. So in short, it's still a long way away from a quantum computer that can properly use short algorithms. I think the biggest number of factors so far a while ago was 21. But we are also entering a near-term phase that's called a NISC era, which stands for near-term intermediate-scale quantum computing. So we're already starting to see some of these initial applications, which Google and IBM, they are showing and working with enterprise customers. I think the idea of qubits is quite difficult to get your head around if you don't have a background in physics or a PhD in quantum computing as you do. The analogy that I like, and this is a very dumbed-down analogy, is if a bit is like a coin and it can be the heads or tails, then a qubit is like a spinning coin. And it's either or both, or not at the same time. And that is quite an oversimplification, but it just helped me to kind of understand what I was thinking of. And it also helps to explain why quantum computers are so difficult to build, right? The hardware required to build a qubit is really, really difficult to get right, right? Yes. So most of the times, all these qubits, the fundamental building blocks, they could be electrons, they could be photons in the superconducting approach. Google and IBM are leading the superconducting approach. So they're using something called Josephson junctions. It's like one superconductor, another superconductor connected or separated with a nanoscale gap. So all of these things are very difficult to control. And most of the times, they have to be at extreme temperature, almost absolute zero temperature in a superconducting case. Yeah. And you might have seen those fridges that they have, that kind of cool, cool these tiny little chips down. So if you see these kind of chandeliers of a quantum computer, and I think you can see something like that at the IQM stand where they've kind of created a sort of model of one of these cryostats. And that's all for cooling this quantum chip down to near to zero degrees to keep it cool. So the qubits kind of stay in the state that you want them to stay in. And as you said, there are a bunch of other different techniques. The superconducting qubits, there's photon based photonics and there's quantum computing and a bunch of other stuff. But the real challenge is how you scale up from dozens of qubits to thousands or millions. We might need millions of qubits to actually run Shaw's algorithm because of the error correction that's required and things like that. Yes. And also hardware isn't the only hurdle, right? There are other hurdles that need to be overcome for this bit to become a reality. Yes. So the longer term goal is to build a fault tolerant quantum computer. So Amit, you just mentioned we need millions of qubits. So all of these are physical qubits. And then when you read academic papers, most of the times they are talking about this concept of logical qubit, which is fault tolerant in a quantum computing scheme. You can think about a bulletproof qubit in a way. But to get to that stage, there is still a long way to go. So say almost if we want to achieve one logical qubit, we might need thousands of messy physical qubits to surround it. So the 127 IBM qubit quantum computer that's referring to 127 physical qubits. And the world so far doesn't even have one single logical qubit achieved. But a month and a half ago, in Chris Monroe's group, they have achieved some milestones in this space. So I think there will be some exciting progress coming. That's why the qubit number can be a bit misleading. And I think why you need to take stories about we've achieved an X number quantum, we've built a quantum computer with X number of qubits. You have to take that with a bit of a pinch of salt because the number of qubits in itself doesn't really tell you a whole lot about how useful that quantum device is. And I think some companies have developed their own terms, you know, like quantum volume or things like that, quantum advantage to tell you how useful the quantum computer is. Because it's not just the qubits. It's the way they're wired up. It's how isolated they are. It's how fault-corrected they are that really adds up to whether or not this hardware is going to do what it needs to do. But hardware, there's also kind of the software that needs to be written, right? So you mentioned Shure's algorithm and Grover's algorithm, but there's a whole kind of missing layer to this that we're going to need if we're going to, these machines are going to be useful, and that's the software and the skills, right? Yes. So in a software space, there is some exciting development recently. For example, Cambridge Quantum Computing and Honeywell, they just announced yesterday they have formed world's largest quantum unit. And then there's a lot of quantum operating system, the new quantum languages, all the IBM, Google, Microsoft, they all have a new set of, so not C++ language, there is a whole new set of new quantum languages to learn. And my Oxford classmate, he's starting a very interesting startup in the software space in Singapore. And that can translate any type of classical computing programs into quantum computers directly, which can be then run on quantum computers. So in the near term, I would see a hybrid model, so say we can have some of the quantum computers embedded in the high-performance computing centers, and then we can have a combination when we run a whole suite of classical computing programs, but with certain parts where quantum computing can bring the most acceleration, now we just use that particular part in quantum, the rest of the computing programs can be still in classical. And I think that can still bring a lot of near-term advantage. What does the future look like, do you think? I mean, because I guess we're talking about these machines as if it's something that I'm going to have in my study, but that's not the case, right? These are going to be in research labs, and if I as a founder or as a programmer access them, it's going to be remotely via the internet. So is that kind of how you see it? Is that how companies will access these machines in the future? I think most of the providers these days, yeah, so in short, I think it will be accessible via cloud. So for example, IBM probably, they started offering these services a few years ago. Google, Amazon, Web Services, Rigeti, IonQ, they are all offering these kind of services via some of the cloud providers. IQM in Finland here, they have a unique selling point which is kind of over at on-premises, but I guess in summary, one thing I can guarantee you, we wouldn't have a quantum computer in our study or bedroom like how we play mobile phone or tablet. Yeah, it's more like a supercomputer, right? It's something you access when you've got a specific requirement. I guess you might not even know that you're using a quantum computer. I think the way if you talk to Microsoft, if you talk to IBM, the way they're selling it is that a company will come to them with a problem that needs solving, an optimization problem, say a delivery company needs to find the best route for them to send its driver around to deliver thousands of deliveries, right? So they might approach a company like IBM or Microsoft or something like that with that problem and then that company will then go away and decide which combination of supercomputer, quantum computer, traditional computing is the best approach to tackle that problem and then they'll give the answer at the end of it. And actually, do I need to know as a startup founder whether that's been done on a quantum computer or a supercomputer? Maybe I don't. Yeah, I think you are totally right. I don't think you need to. So all of these things can be orchestrated at the back end. So quantum computer, hardware, software providers, they can work in tandem and then provide a whole solution and then choose which bit needs to use quantum. The rest of the parts can be classical. So you mentioned a bunch of different startups and there's a lot of the ecosystem is growing. Do you think from your position at Tencent, do you think this is a good time to invest in the quantum ecosystem? Is this something that you'd be recommending to other people in the space that they pay a lot of attention to this? Or is it too soon? Yeah, I think it depends on your risk appetite. Yeah, so we can see there are so many quantum startups in recent years and lots and lots more government and private investment going into all these companies which is supporting great research and advancement for the whole human race in a way, which is fantastic. But how do we, how do investors know quantum computers in this black box? How do you know the startup you are looking at is truly at the forefront of quantum innovation? It's just very difficult to know. And at Tencent we have, we call this investment challenge black box paradox. And ourselves we have developed three lines of defense to look at at relevant stages, say from seed stage to series A. We are looking at certain set of achievements or criteria or milestones and series B to C, that's the second stage, and the series D to pre-IPO, then there are some other criteria we're looking at. But out of all the quantum startups there at the moment, even if it's the most advanced one with the current technology, it's still not possible to tell if it's quantum computer who's actually calculating behind the black box or if it's classical computing because at the moment with all the classical computing power, we can more or less simulate around 50 qubits or so. I guess once people go beyond, say, 200 qubits, then that's well at a stage where you can use techniques such as blind quantum computing or other quantum crypto methods to evaluate if the claim is true quantum or if it's just classical computing pretending as a quantum computer. I'm interested in your thoughts on what the exit strategy for these kind of early stage hardware and software startups looks like, considering the fact that we could be 10 or 20 or maybe even 50 years away from practical quantum computing at scale. I think we've already seen some good exits in the field. For example, INQ has just become the first publicly listed quantum computing company. I think the market will tell how the world is responding to this inspiring deep tech company, but perhaps very early stage commercialization and how the market will respond to it. I think we see more and more IPOs of even the very early commercial stage with really solid technology, but they IPO'd successfully. I'd love to know a bit more about Tencent's own investments in the space you mentioned. Can you tell us a bit about IQM and what made them attractive to you at the time with the investment? I met the IQM founders about five years ago. I came here a few months before I came here for Slash for the first time. What struck me best is I think you have visited the Google quantum computer and we know how difficult to call the system to the almost absolute zero, to be you wrote in your quantum book, it took two days to cool it down and one week to bring it back up to the temperature, room temperature. IQM has developed this really innovative on-chip cooling technology, so that would really help with the scalability. Some of the other advantages are they have very fast, for example, gate reset time and operation time, so those are the advantages. They also have a new concept that's called co-design, so say if you are enterprise customer, like leading OEM, they can work with the customer and develop a quantum chip that's especially suited with that particular customer. I think on-chip cooling is a really interesting example of the kinds of technology that we're going to need if we are going to scale these things up, and it might not necessarily be but super conducting qubits with on-chip cooling are the end technology that ends up being. Because there's a bunch of different competing technologies, do you think that we've reached a winner in super conducting qubits and those are going to be the ones that are going to see us through to full-scale quantum or do you think there's room for another kind of paradigm shift? Because 10 or 20 years ago, super conducting qubits were like a fringe idea that no one really thought was viable and then Google came along and built their quantum computer and now it's like the leading thing and it's used by Google and IBM and to a certain extent by Microsoft as well. What do you think is going to be the kind of winner in this race? I think it's too early to tell. So we normally, when we look at investments so early stage, we would use a portfolio approach. There are just different approaches to have that individual advantages and disadvantages. Superconducting approach, it's the most far ahead at this point, but some might go this curve and some might go slightly more steady at the beginning and very sharp growth towards the middle. So iron traps is another. So iron cue is using iron traps. That's also the qubit fidelity is probably the best amongst all the approaches and they can operate at much more friendly temperature, but it's very hard to miniaturize the device, but maybe that's not so much a problem. And there is topological qubit. Microsoft has been pursuing theoretically it would be much better than iron traps or superconducting, but it's the fundamental problem is the particle itself is so elusive no one has observed it properly in the experiments yet. And there is also Psi quantum in California. They were originally leading professors in the UK from Bristol and the Imperial College and they have silicon photonics approach. The advantage is they can use traditional silicon fab and they can operate at room temperature and they're already targeting millions of qubits, so one million qubits in the next few years. So given the dozens of competing technologies at vastly different scales with different potentials to either saw or completely crash and burn, how can businesses prepare right now and investors, how can they, what should they do? What advice would you give them at this point? I think maybe two things. One is to study the field quite carefully, not just use cases, but especially for investments. So at the first stage maybe even look at the, find the main experts and to look at scientific publications, no publishing, nature science or physical reviews and to look at if the theoretical method is solid. And then within each company I think find out what kind of use cases are there. I know Goldman Sachs, they have a research team looking at how quantum computing can link with black shores and speed up the financial or risk calculations and some of the logistics companies are already starting to use the optimization part and to make, yes, so only the traveling salesman problem solved more easily. So and I'm sure there are lots of other use cases in making communication super secure. Yeah, yeah, absolutely. It's going to be one of those really interesting fields to keep an eye on over the next 10 or 20 years as we see whether or not quantum computing actually comes to fruition and delivers all these potential applications that have been talked about across everything from banking to chemistry to health to cyber security. We are just about out of time, so thank you so much, Link, for that. That was fascinating and thank you for watching, guys. Cheers. Thank you.