- il y a 12 heures
Quantum 101, Part 1 The What, the How & the Who
Catégorie
🤖
TechnologieTranscription
00:04Welcome to Quantum 101. I want to start with a question. Who here thinks they understand quantum mechanics? A show
00:15of hands.
00:17One, two, two. Okay, well that makes this joke fall flat. The late, great Richard Feynman said,
00:24if you think you understand quantum mechanics, you don't understand quantum mechanics. So you lie in.
00:31Anyway, so today we're here to unpack quantum from the beginning up. We have two incredible panelists
00:38who actually understand quantum computing, the impact it's going to have on your business, your lives, your family,
00:46when that is going to happen. So without further ado, I would like to introduce yourselves,
00:53but to explain to the audience what first drew you into quantum computing. What was it about
00:59the quantum realm that first sparked your curiosity? Manjana first.
01:07Hi everybody. Thanks very much for having me here. Great to be here. So I'm Manjari Chandran Ramesh.
01:14I'm a partner with Amadeus Capital Partners, a global technology investor. And we've been investing
01:21in deep tech before it became cool to call it deep tech. And we've been investing in it for the
01:27last
01:2827 years, deployed over a billion into 190 companies, and we've exited 100 of those. And it's been in
01:36different waves of technology, including quantum. So I'm one of four partners in the early stage.
01:43And as part of our quantum portfolio, we've got companies like Riverlane, which is doing quantum error
01:49correction. New quantum, that's doing quantum networking. Photonic, that's doing silicon photonics
01:57based quantum computing. And actually one in Paris, CryptoSense, which got acquired by Sandbox AQ.
02:05And so to answer your question, what drew me to quantum computing? I invest in deep tech. So any
02:14technology that's going to have a huge impact in the population and the world is something that excites
02:22me. So I started investing in AI, not Gen AI, AI, again, before it was cool to invest in foundational
02:31models and other technologies that really make a huge impact. So we're not the investors that would invest
02:40in, say, your food delivery apps, because not that it doesn't impact, it's just not as transformational
02:47as some of the deep technologies that we do invest in. So when we started looking at the quantum field
02:55as
02:55a whole, we realized that this is a technology that has the power to now be able to do things
03:02that
03:03haven't been possible before, that haven't actually been possible by the classical computer.
03:08and that's hugely impactful. So that's what made me look at it.
03:13Fascinating. Thank you for that. Ashley.
03:16Hi, everyone. It's really good to be here. I'm Ashley Montanaro, co-founder and CEO of FaceCraft,
03:21the quantum algorithms company. And really, our goal as a company is to develop the algorithms and
03:26the software that will make quantum computing useful in the near term, you know, in just a few
03:30years rather than in 10, 20, 30 years time. And I've actually been working in the field of quantum
03:36computing for about 20 years now, primarily in academia, although a few years ago, together with my
03:42co-founders, John and Toby, we decided to make the shift into industry and set up a company to try
03:47and
03:47take this technology out of the lab and into the real world. And what originally got me into quantum
03:53computing, in fact, before that, I was a software engineer working on software for smartphones before they
03:58were called smartphones. So, you know, back in the dark ages. And what got me into it is I heard
04:03about this
04:04ridiculous idea of using quantum mechanics to make better computers. And I actually didn't believe it. I thought,
04:10okay, this, this doesn't seem real. Like, how can we possibly use this crazy physics idea that people don't seem
04:15to
04:16really understand to do something better than we can using normal computers? Then I started reading more and more about
04:21it.
04:21I understood a bit more, realized that actually, this was a really serious thing. And we were going to be
04:26able to
04:27do things like model quantum mechanical systems occurring in nature exponentially better than we can using our
04:33standard computers. We could break cryptographic codes, which are totally secure against classical computing's attacks,
04:40and even solve hard optimization problems better than we can with our normal computers. And so I realized that this
04:46was
04:46actually going to be a really transformative technology, as Manjari said. And I thought I really wanted to work in
04:51this
04:51field. And I've been working in it ever since. Transformative is the word, I think, I want to stay with
04:57you,
04:58Ashley, and I'm sorry, but you have the world's most difficult question. So all of these good people can go
05:06home tonight and explain to their family and their friends what quantum computing actually is, in a way that
05:12they will understand. Could you, comparing it to classical computing, give us a, an outline of what we're
05:20actually looking at here? Sure. So it's important to say that a quantum computer isn't just a faster
05:27standard computer. It isn't just a smaller standard computer or, you know, a more efficient one. It's a
05:33computer that operates in a totally different way. It's a computer that uses the features of quantum
05:37mechanics to do things that are just not possible to do if you just use boring old classical physics
05:43that's been known for centuries. So an analogy which I quite like is it's like the difference between
05:48getting to your destination using a car or using a plane. It operates in this sort of different
05:54dimension, this different space, which classical physics doesn't have access to. And in particular,
05:59there are some key features of quantum mechanics which quantum computers use, such as quantum
06:04superposition, which means the ability of a quantum system to be in effectively more than one state at the
06:09same time. So it's like a form of parallelism, you know, we might feel natural from classical computing.
06:16And also quantum entanglement, which is the fact that if you have a quantum system made up of many
06:21parts, you can't understand it just in terms of its separate parts. You have to think about the whole
06:26thing together. And this means that to represent the state of a quantum system with a classical computer,
06:31you actually need an exponential complexity to do this exactly. So if you have a system with a hundred
06:37qubits, you actually, quantum bits, you actually need two to the one hundred numbers to represent
06:43that with a classical computer. So it's somehow this exponential blow up in complexity, which is one
06:48of the key reasons that quantum computers get their power. Okay, I think I got some of that. Manjari,
06:55as an investor, you perhaps have a different way of explaining to your potential clients and potential
07:01investors, what quantum computing is, how do you think about and define it in those situations?
07:11So as much as we invest in deep tech, it comes back to who's going to pay to use these
07:16systems,
07:18right? And so for me, it's all about what can these systems do in the real world? What is the
07:24use case?
07:25Is that going to be financial? Is that going to be in pharma? Is it going to be in new
07:31materials? How is this
07:33actually going to be transformative? So as much as the technology excites me to my underlying investors,
07:40that is almost a second layer. The first is what use case is actually going to happen? And what's it
07:49going to enable once quantum computers are actually here? And that's how I explain it.
07:54Can you speak on any particular projects that you're currently funding or working on, where the impact
08:02might be? Yes, absolutely. So, you know, one of the companies that I'm very closely involved with,
08:09New Quantum, which I mentioned, is doing quantum networking. Now, it's easy to, you know, sort of assume
08:16that like classical computers, the genesis would be, we'd get the, you know, vacuum tubes,
08:23you'll get the quantum computer, you'll then look at how you connect a lot of the quantum computers.
08:29And interestingly, the way the industry is going is actually, that's not the steps that it's going
08:36to take. The networking has become a lot more important a lot earlier. And what that means
08:43is it's looking at a possibility of not just connecting the quantum computer to a classical
08:49computer, but also connecting the different modalities that are within the quantum computer
08:54to each other. So it's really opened up a huge space.
09:01Ashley, there's so many questions I could ask you about the technical side of this, but
09:06what do people get wrong about quantum computing? What do you see as the biggest
09:14thinking error, mental model disaster that people have?
09:19So I think it's connected to a point that I mentioned about this, not just being like a better
09:24standard for classical computers. So it's not the case, for example, that you have, if you have a
09:30gigantic data set and you, you know, you want to do something to it, the quantum computing will
09:34necessarily help you do that. In fact, in general, quantum computing is helpful for problems where
09:39you have a very challenging problem to solve. But it's not that you have, the reason it's challenging
09:45is that you have a huge amount of data. It's just sort of fundamentally challenging for some
09:48sort of real sort of mathematical theoretical reason. So for example, if you want to model a
09:55battery cathode, then to do this accurately, you need to represent the quantum mechanics within that
10:00battery cathode. And that's why standard computers can't do it. And it's why quantum computers can do it
10:05very well. But if, for example, you have a gigantic amount of invoices in your accounting company and
10:13you want to process them a little bit more efficiently, then quantum computing is not
10:16going to help you with that. So I think it's just a question of understanding like where the
10:20applications are. And some of these applications are incredibly important, where they are and where
10:24they are not. Okay, thank you. And our next talk, by the way, quantum 101 part two, we're going to
10:30be
10:30doing a deep dive on the impact where it's being felt the most. So stick around for that one.
10:36Ashley, as CEOs of PhaseCraft, I think most people are familiar with the term qubits.
10:45Where are we in terms of the number of those? Is that even important? How many we have? Where are
10:50PhaseCraft? And are we ahead of schedule, behind schedule? What's the timeline looking like?
10:57Right, so this is a pretty big question. So I'll, you know, lots of questions.
11:02Yeah, so really, there are a couple of key parameters that go into the specification of a
11:08quantum computer that tell you how powerful it is. And one of them is the number of qubits,
11:12which is basically the analog of the number of bits in your classical memory. And the other is the
11:18two qubit gate fidelity, which is basically the number of operations that the quantum computer
11:24can perform before it gets overcome by noise and errors, and the computation no longer has any kind
11:30of quantum power associated with it. And firstly, on the the qubits point, the best quantum computers
11:37that are available now that you can use on yourselves in the cloud, they might have a few hundred qubits,
11:43that's, you know, possibly getting up into the sort of high hundreds now in some cases.
11:47So if you think about it, this is, you know, like having a classical computer that has a few
11:52hundred bits in its memory. So it's clear we're at the very sort of the equivalent of the very early
11:57stages of classical computing. This is like the 1950s or something like this. So it's kind of amazing
12:02that these machines can already outpace the world's best supercomputers for some tasks. And on the other
12:07hand, this two qubit gate fidelity metric, the best quantum computers we have now can do maybe one or
12:13two thousand operations before they, you know, can't compute anymore because of noise and errors.
12:19So this is already a stage where we can solve problems where, like, the world's best supercomputers
12:27might take hundreds of years or thousands of years to solve that problem. But these problems are ones
12:31which are generally chosen to be really easy for the quantum computer and really hard for the classical
12:37computer. So there's a really key next milestone for the field of quantum computing, which is getting
12:42to this point of so-called practical quantum advantage, which is where quantum computing beats
12:47classical computing for a problem which we genuinely care about, not just to make the quantum computer
12:52look good. And that's what we at FaceCraft are working towards because we're developing the algorithms
12:56that will push down the complexity of these problems. Okay, thank you. Is it true to say that
13:04no two quantum computers are the same? Are there different modalities, different ways of solving
13:12the quantum problem? I mean, is that a yes or is that a no, Ashley? There are a number of
13:21different
13:21technologies out there and it's kind of amazing that they are very different. It's not like
13:25classical computing where now basically every computer is based around large-scale integrated
13:30CMOS circuits. In the quantum computing world, there are things like superconducting qubits,
13:36there are iron traps, there are neutral atoms, there's photonics, and there are others as well.
13:40And it's kind of amazing that in fact the number of technologies that are used to build quantum
13:45computers hasn't gone down over the last years as it's been sort of there's been a selection process,
13:49but in fact it's gone up. There are technologies now which are really at the forefront which we're not even
13:54really considered to be contenders a few years ago. So I think it's an amazingly diverse field and it
14:01may even be that there is no one winner and that we end up with different hardware platforms used for
14:05different problems. Which plays nicely into the hands of investors and founders, does it not? How does this
14:13variety of quantum computers affect your Amadeus investing and the fundraising aspect of this?
14:22So I'll first cover how it affects the the whole landscape as a whole, right? So if you look at
14:27what
14:28investors have done is some of them have decided they don't want to pick a horse at all. So they
14:35will go for
14:35everything that is agnostic of the qubit modality and that's all they'll do. So they they won't they
14:43won't decide whether it's an ion trap or a superconducting etc. There are some who've decided
14:47that they will bet only on one qubit modality for whatever reasons, you know, whatever has got them
14:54excited. So you'll find some investor portfolios literally having only the ion trap or only the silicon
15:00play. And a few of us, which I include Amadeus in, who have taken a few different bets. So if
15:09you look
15:09at the Amadeus portfolio, our general hypothesis is that there are going to be different modalities.
15:18The current research shows that it's going to be different modalities for different applications.
15:24So we have gone with silicon photonics because silicon really scales the number of qubits.
15:30With what, sorry?
15:31Silicon photonics.
15:32Silicon photonics?
15:33Yes, on the quantum computing side. But we've also invested in, as Ashley mentioned,
15:39quantum error correction. We've invested in Riverlane. We've invested in new quantum,
15:44which is quantum networking. And these guys work with all of the different modalities.
15:48So there is a sort of a bet across the different modalities through the fact that we have gone
15:56for the agnostic plays as well.
16:00Okay. Back on investing for you, Ashley, how have you found the landscape? Has it been easier than
16:07you expected? More challenging than you expected? Or are people throwing money at you?
16:14So I was coming from academia when I co-founded Facecraft. So obviously, it was a bit of a culture
16:20shift by its very nature going into this sort of startup world. And I think, you know, within that,
16:26raising funds for a quantum startup company is particularly sort of interesting challenge. And
16:32I think we, as a company, were very clear that we wanted to look for investors who really were trying
16:39to understand the space and took it very seriously. And it's certainly the case that a few years ago,
16:44there were investors out there who are not perhaps traditional deep tech investors who,
16:48you know, had heard about this quantum word and thought, okay, this is pretty cool. I think we want
16:52to get involved. But they may have been perhaps not really appreciating the deep tech elements of it
16:57and maybe thinking, okay, this is maybe more like a traditional sort of SaaS company where we can be
17:02sort of in and out within, you know, six months or so. So I think that there, definitely at some
17:08point,
17:08there was a bit of a sort of educational piece for the sort of investment community and sort of
17:13understanding how quantum works and how it does fit in within this deep tech community. But I think
17:19for us, you know, we have found that there are great investors out there who do understand quantum
17:25and are, you know, very enthusiastic about it and see its potential. So it's all about finding that match
17:30up really. Can I? Yes, of course, go ahead. So I think it's an interesting point that you make,
17:36right? So there are investors who absolutely understand the space and understand it's going
17:42to take the time that it takes. And with those investors, while, you know, I think companies,
17:49our portfolio, but otherwise as well, if you spend the time with them, yes, they may have the conviction,
17:56they may get the conviction, and then they are your supporters. They're absolutely your supporters.
18:01And it might be that you don't get the conviction with them. But that's fine, right? Because they
18:06do understand it. And their conviction lies elsewhere. I think the dangerous bit is going
18:11for the investors who are just in it because it sounds cool, because they're not coming at it from
18:17a conviction play. And they're going to be the first ones to pull when the thing, when the going gets
18:23tough. So part of our job on the board as board directors is actually working with the management
18:30teams to work out which type of investor we want to build that syndicate with. Because these things
18:36are hard. You know, the field is hard. Deep tech is hard. Quantum is hard. You really want people who
18:43understand it, who can support it. Quantum is hard. Yes. There's a lot of people in here who are perhaps
18:54thinking that quantum is still science fiction, that they don't need to be worried about this. It's
19:00more like perhaps even a novelty technology. Where, as to leave this, could you paint us a picture of the
19:08immediate future? So the next couple of years. And a more long-term, like 25, 35 years out future?
19:16And should the same people in a few years have the same concerns with the people in 20 years?
19:22What does the future look like, Manjani?
19:27So I think, you know, I started first investing in quantum in 2017.
19:31And, you know, my first investment memo to my colleagues was very much, this is a moonshot.
19:37And I think I used the word moonshot so many times. The first thing everybody went was, we get it.
19:42It is a moonshot. You've made your point, right? So, you know, we got to the moon.
19:47Yes, exactly. So that was that was the plan. But I think the landscape has really come up
19:54much more quickly than what, you know, a lot of us who got involved in 2017, 2018 thought it would.
20:00So it's fabulous for the field, right? As Ashley mentioned, you know, there are now systems where
20:06you can show, you know, the quantum advantage for particular problems. So I think in the nearer term,
20:13that's really going to progress. I think we are going to see a practical quantum advantage
20:18sooner than what we originally thought we might see, because everything is progressing towards that.
20:24The whole landscape and the ecosystem is working towards that. In the longer term, I mean,
20:30I would hope that it genuinely has the impact that it does, right? You want to be able to do
20:37the
20:37interactable problems. You want to be able to do the new materials. You want to be able to help with
20:42drug discovery. There's so much personalized medicine already. You know, if you could actually help with
20:47that, then that would just be so phenomenal. So I would hope it lives up to that vision.
20:53You're nodding your head there, Ashley. Yeah. So I can only agree with everything Manjari said,
21:01but just to add a bit more, I mean, at FaceCraft, we're really working towards getting to these
21:06practical applications of quantum computing in the next few years. And we see the first such
21:11applications as probably being in the world of modeling new materials for maybe battery cathodes,
21:18photovoltaics and solar cells, maybe things like high temperature superconductors, and also solving
21:25hard optimization problems, like maybe optimizing the energy network in any given country, or maybe
21:32modeling and optimizing a financial portfolio. And we think that the first such applications will
21:38be realizable within a few years time. And also that, you know, within a longer timeframe,
21:44the applications are only going to increase. I mean, in the future, we do think that quantum
21:49computing will be an indispensable tool for scientific and technological discovery,
21:53and people will be modeling next generation drugs, next generation battery storage technologies,
22:01and in sort of many other such applications. So, you know, we think really the future is
22:08bright in the near term, but also very much so in the longer term.
22:12Any closing thoughts, Manjari?
22:17The only thing I'd say is, if you can get involved in this landscape, please do. It's
22:20it's doesn't get more exciting than this.
22:23As someone who's got a bit of a background in crypto, final question, I have to ask,
22:28is my Bitcoin safe? And for how long?
22:33Right, so it's a good question. So, okay, so as maybe many of you are aware...
22:37I don't have much.
22:38Yeah, as maybe many of you are aware, one of the famous applications of quantum computing is that
22:45quantum computers could ultimately break the RSA crypto system. And interestingly, some estimates are
22:53that maybe up to 25% of Bitcoins might be vulnerable to this sort of attack, this cryptographic attack.
23:00But it's worth stressing that quantum computing is still, you know, a developing technology. And
23:06in particular, the algorithm that you use to break the RSA crypto system called Shor's algorithm
23:12is a relatively hard one to run on a quantum computer. And sort of current expectations are
23:18it's probably 10 years plus before you can actually do that. So cryptography is always a sort of, you know,
23:25developing field and people are even now developing post-quantum crypto systems that are
23:30able to be secure against quantum attack. And, you know, these are starting to be rolled out already.
23:35So I would say, you know, your Bitcoin's safe for now.
23:40For now. For now. I'm sure you've got a lot of questions. I was going to open the floor,
23:46but we've run out of time for this. If you have any further questions, then please find our lovely
23:52guests out in the corridor. They'll be happy to answer any questions. And if you stick around,
23:57we've got some some new big guests coming on next to talk about the real world impact of quantum
24:02system today. So stick around for that. I hope it's a little bit clearer for you now. Thank you.
24:09Thank you.
Commentaires