From weather forecasting to artificial intelligence, chemical engineering, and drug design, quantum computers have the potential to completely change the way we think about solving complex problems. But how far away are we from seeing quantum computers used to tackle the world’s biggest problems?
Our guests in this episode:
- Anna Phan - IBM Q.
- Alberto Peruzzo - RMIT Quantum Photonics Lab
- Will Zeng - Rigetti
Research for this episode by Mahalia Carter.
Our theme music is by Breakmaster Cylinder.
And our cover artwork is by Andrew Millist.
Reporter: I was going to ask you to explain quantum computing, but... [crowd chuckles] Um, when do you expect Canada's ISIL mission to begin again, and are we not doing anything in the interim while we prepare?
Justin Trudeau Okay, very simply: normal computers work by...[crowd laughs, applauds]
KRIS: Welcome to Moonshot - the show exploring the worlds biggest ideas and the people making them happen. I’m Kristofor Lawson.
ANDREW: And I’m Andrew Moon.
KRIS: And you’re listening to a press conference from 2016 where Canadian Prime Minister Justin Trudeau was jokingly questioned about the quantum leap that is taking place in the computing world… and it turned out he wasn’t quite ready to led that joke slide.
Justin Trudeau: No no no, don't interrupt me. when you walk out of here, you will know more--well no, some of you will know far less about quantum computing, but most of you. Normal computers work,either there's power going through a wire or not: it's one or a zero. They're binary systems. What quantum states allow for is much more complex information to be encoded into a single bit. Regular computer bit is either a one or a zero: on or off; a quantum state can be much more complex than that because, as we know, things can be both particle and wave at the same times, and the uncertainty around quantum states allows us to encode more information into a much smaller computer. So that's what's exciting about quantum computing and that's where we're going. [ applause] Don't get me going on this, or we'll be here all day. Trust me. [crowd laughs]
ANDREW: As the Canadian PM so eloquently explained there, quantum computing takes advantage of the ability of subatomic particles to exist in more than one state at any time. And if these systems really do become a day to day reality, they would be able to solve problems that our normal computers just can’t… as Eric Ladizinsky - co-founder of quantum computing startup D-Wave explained back at Wired conference 2014.
Eric Ladizinsky: “Imagine you had a math problem. Well I’ll make it easier… let’s say you go to the Library of Congress, there’s 50 million books, in one of them I secretly put an X on one of the pages, I tell you to go into that library. I want you to find the x, you have 5 minutes, you say, well okay that’s going to take me many lifetimes to go through those books sequentially. I’m not even going to take that problem. But if somehow I could put you in this magical quantum superposition in which you were in 50 million parallel realities and in each one you try a different book. Now the other important feature is that you have to be able to collaborate with all your other selves. And these aren’t clones of yourself the same physical matter is somehow playing out all these different roles. Talking to your other selves we call coherent evolution, you’re kind of coherent to your other selves and you can share information… Well now you find the book very quickly and that’s quantum computing in a nutshell.”
KRIS: Stop and think about that just for one second. From weather forecasting to artificial intelligence, chemical engineering, and drug design, quantum computers have the potential to completely change the way that we think about solving complex problems.
ANDREW: But how far away are we from really seeing quantum computers used to tackle the world’s biggest, meatiest, problems? We’ll have more on this great big topic right after this break.
Anna Phan: The basis of today's computing is we think that any problem is sort of solvable with a supercomputer, as long as we have enough memory or we have enough CPU or we have enough time to solve that problem. There are actually certain problems that these classical computers will never be able to solve.
KRIS: This is Anna Phan, a researcher for IBM based in Melbourne. IBM has been doing an awful lot of research in the quantum computing space. This topic is very much a buzz word in technology circles at the moment, but what exactly is it? To understand, we actually need to know how our existing computers - including the phone in your pocket - work right now.
KRIS: You heard this earlier from Justin Trudeau’s impromptu explanation, but most computers store information in binary units called bits - which can take a value of zero or one. It’s kind of like having many on and off switches that control the device. Now this system works really great for traditional computing however, a quantum computer uses qubits, which are created using atoms or photons and can then store the information in many different states. As you scale up the number of qubits - the power of your quantum computer scales also allowing you to solve more complex problems and algorithms.
KRIS: And that’s a really important point to understand. Quantum computing isn’t about making the next iPhone 100 times faster. Quantum computers are really about solving for big, complicated problems.
Alberto Peruzzo: Multiple qubits is where the power comes.
ANDREW: This is Alberto Peruzzo, head of the Quantum Photonics Lab at RMIT University in Melbourne.
Alberto Peruzzo: When you have to identify qubit, you need two parameters. To identify for two qubits you need four parameters. And every time you add a qubit to your string of information, you basically encode, you double the encoded information in that. And that's where the power comes from. So if you have three bits you need three parameters, but if you have three qubits, you need eight parameters.
KRIS: And it’s this ability to rapidly increase the number of parameters available in the system that allows quantum computing to solve some of nature's greatest and most complex problems.
Alberto Peruzzo: There are problems in chemistry and physics that even if you had a computer of the size of the whole planet, you will never be able to solve. So it's just at some point you can make a problem that is big enough. The rules are any capacity that we could ever think about performing with this classical methods.
Anna Phan: An example of such is simulating molecules down to their most fundamental interactions. Consider caffeine in your cup of coffee. This is, surprisingly, is complex enough that it takes 10 to 48 classical bits to try to simulate and understand that. So that's, 10 to the 48 is almost all of the atoms on earth at the moment, but, on a quantum computer, this would only take about 100, 160 to 200 qubits to understand.
KRIS: And if after all that you’re still confused about the idea of a qubit - you wouldn’t be alone. We’ve been looking at this topic for a while now and it’s still pretty confusing. So here’s Will Zeng - Director of Evangelism and Special Projects at quantum computing startup Rigetti to explain.
Will Zeng: You should think about it as the amount of memory you have in your quantum computer. And by that I mean it's actually not the only thing you'd use to measure the performance of a quantum computer, like when you look at a laptop you're not just looking at your hard disc, you're looking at your RAM, you're looking at your clock speed, you're looking at a whole bunch of different parameters. And when you look at a quantum computer it's the same sort of thing. You want to look at the number of qubits, the quality of those qubits, the way they're connected matters as well, and then other features like their clock speed, and the latency with which you can access that quantum computing system. And when we think about a quantum computer we try and take into account all of those different features, because as we've discovered more, and more of those different parameters by working with users who work with our systems, we've realised how much they matter as well as the number of qubits
ANDREW: So from Will’s explanation, the parallels become a little clearer between qubits and regular computer memory. More qubits, faster computer - but it’s not the only factor.
ANDREW: Now these qubits can come in many different varieties and forms based on the technology and the types of molecules that a company might use in development. As Will just mentioned - the number of qubits might be important but what’s also valuable is the quality of those qubits, how those qubits are connected, and what they’re made from.
Alberto Peruzzo: As an example you could use a trapped ions, you could use atoms, you could use superconductors, you can use a particles of lights which are called photons. And depending on which technology, which particle you're going to use, you might face different technological issues. So this basically branches out into different research fields. They all want to build a quantum computer, but they have different approaches. Is a bit like using different materials to build a chair or a table. And then there is a little bit of competition between these different particles. My group is specialised on using photonics. So we encode the quantum information in states of single photons which are single particles of light.
KRIS: Now Alberto says while there are significant challenges with using photons in quantum computing - the big advantage will be when it comes to networking these systems. Because using photons you’ll be able to connect multiple quantum computers and send information easily between them.
Alberto Peruzzo: Photons are flying qubits. So a photon light moves very fast... and if you put in the right material will maintain its properties. So you can send the quantum information across an optical fibre or through air very, very far with minimal noise. If you try and do the same with other particles, it would be impossible.
KRIS: Unlike Alberto’s team at RMIT, over at IBM they’re using superconducting qubits which is what many companies, including Rigetti, are also using. And they’ve been able to make significant advancements in this technology.
Anna Phan: In that space, we've been able to create more and more qubits over the years, and connect them, and try to run more algorithms, and more complicated algorithms on them.
KRIS: Because quantum computers are great at solving for complexity - they’re perfect to a research environment. And everyone we spoke with for this episode says the future involves having both quantum and regular computers sitting side by side because they both have their own strengths.
Alberto Peruzzo: We cannot really tell that if in the next few years we will find more applications for quantum computers, but as of now e don't expect it to be able to run Microsoft Word on a quantum computer. But if you want to detect a disease or predict the behaviour of a drug that will definitely have an advantage.
ANDREW: You might not be surprised to know that there aren’t actually many quantum computing systems around - and that’s for a number of reasons. Not only is the process of creating qubits incredibly complex, to actually operate one of these systems at an optimal level, you need to keep it in an environment as close as possible to absolute zero.
Anna Phan: We build these systems and we put them into something called a dilution refrigerator. These actually work at 15 to 20 milli kelvin. So that's 15 to 20 milli kelvin above absolute zero. That's actually colder than outer space. And so that is needed so we can actually control these systems to the extent that we need to do these computations.
ANDREW: And for those of you playing at home who may not remember the temperature conversions from science class, my hand is going up. Absolute Zero is equivalent to –273.15 degrees Celsius or -469.67 degrees Fahrenheit, which is pretty darn cold. And to keep the system running at that temperature requires an awful lot of gear.
Anna Phan: There is a lot of infrastructure around the refrigerator where the chips are laid, as well as the room temperature electronics needed to control the systems, and then connect to the internet so that other people can use them.
Alberto Peruzzo: And this is where we keep the quantum computer. This is a cryostat, it can reach one kelvin of temperature.
KRIS: That's pretty cold.
Alberto Peruzzo: It's pretty cold. It's not as cold as superconducting cubit technology requires but for photonics it's enough. So, what we need it for is to cool down the single photon sources and the single photon detectors that need to operate at very low temperature for noise issues, and because our detectors are based on superconductors the superconductors need to be cooled below the critical temperature and when a photon gets absorbed by this superconductor it brings the super [inaudible] to conductor. You can use this to monitor the presence of, the detection of, single photons.
KRIS: I mean it looks like something you might see in almost like a sci-fi sort of film or something like that.
Alberto Peruzzo: Yeah. This cryostat contains a number of shields so that you can basically separate stages of cooling and also protect it from external radiation, the devices, and it also contains a large number of electrical and optical connections that we can then use to access the chipper at the bottom of the cryostat.
KRIS: Now Alberto says the size of the machinery needed to house the quantum computing chip can be tens of square metres in size… all for a chip that measures a just a few centimetres. So in the end you’re probably not going to have a quantum computer sitting on your desk at any point in the near future… but that won’t really matter - because IBM and Rigetti are both working on building ecosystems for researchers and developers to use these quantum computers remotely. And we’ll take a look at these systems right after the break.
KRIS: Welcome back to Moonshot - I’m Kristofor Lawson. And as I mentioned before the break, quantum computers are physically pretty large due to all the equipment needed to keep them in their optimal operating state… which is as cold as possible. Which means most people won’t have direct access to one of these systems. But if you’re a researcher out there wanting to solve a complex algorithm, IBM and Rigetti are both working on cloud-like systems to allow you to leverage their quantum resources.
Anna Phan: So two years ago, we put a couple of quantum computers on the cloud for people to use. We put a five qubit computer up there. We allow people to use them using a drag and drop interface. Then from there, we realised that drag and drop was obviously to basic for some of the more sophisticated developers out there, and then released a API and SDK, where people could programme these quantum computers using Python.
Anna Phan: People can run chemistry and optimization and artificial intelligence applications without really understanding the quantum computing, just like you use other computing packages today.
Will Zeng: We call ourselves a full stack quantum computing company. We make hardware and software for quantum computing that we provide through our quantum cloud services platform.
KRIS: That’s Will Zeng again from Rigetti. Rigetti’s goal is to build a 128 qubit system, which is much higher than the record set by Google with 72 qubits in early 2018 and is very similar to IBM, and because they’re looking to find a sustainable business they’re allowing researchers to use their software on a cloud based platform.
Will Zeng: So one of the things we realised quite early on is that in order to make the best use of this technology we need to grow a field, an ecosystem, of developers and users, who can apply it. And the best way to do that was with an open platform where we actually put our prototypes out quite early in their development for people to start to use them, to start to figure out the programming models.
KRIS: Rigetti have taken a very platform driven approach towards quantum computing - they released a Python API for quantum programming, with open source Python libraries.
Will Zeng: We already had users working on an 8 qubit, and 19 qubit processors that we've had available as part of our Forest Cloud platform, and this next series of chips is a generation that's going to get us to enough quantum memory to start to pursue something we call quantum advantage… And by our estimates you need something around ... certainly more than 60, more like 100, 128 qubits to start to build quantum advantage applications. So these are applications where you use the quantum computer to solve business problems faster, or cheaper, than any other computing solution.
KRIS: Right, OK, you're seeing that as sort of like the point where the quantum computing system becomes more powerful than using any other form of computer system to solve a problem?
Will Zeng: That's right, and it's going to happen in different verticals at different times, and it's going to continue to evolve. I mean for classical computing we're still discovering new ways to apply them, but the platform we're building on top of our 128 qubit chip, including the hardware, that chip, and then also the quantum cloud services software platform on top.. really go together to produce a platform to start to find quantum advantage for the first time.
ANDREW: After 18 months of feedback from researchers from 30 countries and 90 million different experiments, Rigetti has recently revamped their cloud platform.
Will Zeng: The way it used to work, is you downloaded some Python libraries, and you've got an API key and you could write some code in our language, and they would compile to an instruction set that you'd send over an API to the quantum computer, and you'd get back an answer.
ANDREW: The system is now called Rigetti Quantum Cloud Services, and it’s driven by an integrated quantum classical model. Now before you ask that means where regular computers work alongside quantum ones.
Will Zeng: So we've actually taken the step of now, instead of just having a quantum computer, giving you an API, we're actually building an integrative quantum classical data centre here in California, that houses our quantum computers and classical compute resources in house.
KRIS: Ok so you really see at least the immediate future of quantum computing being in this space where it sits alongside a traditional system and the quantum computing system takes over when there's benefit in the quantum system, and the traditional system does what it's best placed at doing?
Will Zeng: That's right yeah, and I don't just think that's in the near term, I think that for a very, very long time that's really how it's going to work. And I want to really emphasise that the coupling between the quantum and classical has to be extremely tight to get the best out of both resources. We're shaving off milliseconds here, and also the programming model has to be tight to make the best out of the technology on the quantum side that's going to be available in the next 5 to 10 years.
KRIS: Now looking at traditional computing systems, and traditional processing chips, there was always this concept of Moore's Law which says that the computing power would double every couple of years, but does that theory actually hold true when it comes to quantum computing?
Will Zeng: We're working on it. There's not a tonne of data yet but so far we've made our first qubit in early 2016, and we've about doubled the number of qubits on our chips every six months since then. Keeping with the roadmap that we have that trend will broadly stay. But again it still remains to be seen exactly what the cadence will be, you know, is it going to be 18 months, is it going to be 2 years, are we going to do tick tocks between increasing qubit quality and number.
ANDREW: As we mentioned earlier in the show scaling quantum computing allows you to solve big, complex problems much faster. And one of the big fears is that quantum computers won’t just be used for good. They could also be used to decrypt information that had previously been impervious to hackers.
Anna Phan: There are a couple of quantum computing algorithms that will possibly break today's encryption methods, in terms of how we secure our computers today. However, the level of quantum computing power assumes to run these algorithms, assumes the device that needs thousands or even millions of thought tolerant qubits. These are something which the industry is still working towards and is many, many years away.
Alberto Peruzzo: There are two points in the debate that this is not potentially a problem. One is that there are many encryption algorithms that do not rely on prime number factorization. So the first solution is just take something that doesn't encrypt using prime numbers. And the second thing is that before we're going to solve large enough problems. Before we have a quantum computer large enough to solve that problem to actually threats the cryptography community, we need a very, very large number of qubits and I feel it's very far.
KRIS: Despite the very remote possibility of breaking encryption, the community is already talking about ways to protect data in a quantum computing world.
Will Zeng: Within the last year, NIST, the National Institute for Standards and Technology in the United States, which maintains the cryptography and encryption standards put out a call asking for quantum secure encryption methods. So to look to see how they can update encryption standards to be resilient to quantum computers, even though relevant sized quantum computers for that application are still decades out. So I think yeah, over the long term it's going to change how we think about encryption but that's going to, I think, that change is going to pale in the way quantum computing, quantum cloud services, are going to change how we think about optimization, machine learning, design of drugs and molecules and stuff like that. Like I think that's going to be really where the impact is first.
KRIS: As we’ve heard - quantum computing can solve these really big burning problems that scientists and researchers have held for decades. And for Anna and the team at IBM, the goal is to really focus on building the infrastructure needed to help solve those questions, and provide the world with much-needed answers.
Anna Phan: As a technology company, we want to build technology that helps the world. Currently, we can see the limitations of our current technology. So we're trying to build technology that in the future will be able to solve problems we can't do today.
Anna Phan: I would really like to see us try to lead to revolutionary breakthroughs in materials in drug discovery. I know that this is ... it might not happen, and this could be a long way off, but I think in terms of personalised medicine and helping people's lifestyles in the future it'd be really good to see if we can in turning it into trying drug discovery, really trying to go into drug design.
KRIS: What excites you most about the possibilities of quantum computing?
Anna Phan: The possibility of quantum computing is being able to solve problems that today we think are unsolvable. We make so many assumptions when we simulate systems that we just assume that we can't do it from scratch. The hope is that quantum computers in the future will be able to do this, so we can understand these quantum mechanical systems from the bottom up, rather than making assumptions top down.
Anna Phan: And then, hopefully, the computers that we use to solve these types of problems will be able to be used to solve other problems that we think can't be solved today. Sort of like how classical computers were first built to calculate missile trajectories and break interesting codes for war time, but now use them for everything.
KRIS: If quantum computers do reach this point of quantum advantage where they become better at solving problems then traditional computing systems - it will certainly change the way that researchers look at complex problems. But there is one big issue when it comes to this technology… and it’s the fact that it’s just super confusing. We’ve been looking at this for weeks and to be honest we’re still fairly confused about some of the details because each expert says a slightly different thing due to the different approaches.
KRIS: A lot of the videos you see online with people trying to explain the technology dive into topics around superposition and quantum entanglement… but your average person won’t find it easier to understand quantum computing if you start using too much of this language that they really don’t understand. So we’ve chosen not to dive too far into those ideas and I don’t think we’d do you a service by trying to explain them.
KRIS: Now this lack of understanding could easily flow into the ranks of government but to be honest, how often do government actually understand technical issues. But to it’s credit - it’s worth pointing out the the US house of representatives recently passed a bill called the National Quantum Initiative Act, which is about trying to make sure the US remain a leader in quantum technology. It’s a great first move and it would be exciting to see other countries really embrace developments in this field at a political level.
KRIS: In Australia - where I’m from - our Australian of the Year which is among the highest honours one can achieve is Professor Michelle Simmons - who has been working to make quantum chips from silicon. It’s exciting technology and we tried to get Michelle on the show but being the Australian of the Year is a pretty time consuming job.
KRIS: But even through all this confusion about the technology the idea of being able to one day solve significant problems in chemistry, and being able to design drugs specifically to each individual person’s needs, is just really exciting. And I hope that we continue to see more advancements in this technology, for everybody’s sake.