Quantum Computing is Tomorrow's Computing, Today
26:11 Share

Hello, hello. Malcolm Gladwell here. I want to tell you about a new series we're launching at Pushkin Industries on the 1936 Olympic Games. Adolf Hitler's Games. Fascism, anti-Semitism, racism, high Olympic ideals, craven self-interest, naked ambition, illusion, delusion, all collide in the long, contentious lead-up to the most controversial Olympics in history. The Germans put on a propaganda show, and America went along with all of it. Why?

This season on Revisionist History, the story of the games behind the games. Listen to this season of Revisionist History wherever you get your podcasts. If you want to hear episodes before they're released to the public, subscribe to Pushkin Plus on Apple Podcasts or at pushkin.fm slash plus.

Hello, hello. This is Smart Talks with IBM, a podcast from Pushkin Industries, iHeartMedia, and IBM about what it means to look at today's most challenging problems in a new way. I'm Malcolm Gladwell. In this episode, I'll be discussing the capabilities of quantum computing with Dr. Dario Gill. Dr. Gill is the Senior Vice President and Director of IBM Research.

He's also recognized globally as one of the brightest minds in the quantum computing industry. What we know is that it's theoretically sound and possible and that we are making very significant progress towards achieving that goal. And that's why this is a quest of doing something that has never been done before. But it is definitely possible.

Earlier this year, at the Wall Street Journal's virtual CIO Network Summit, Dr. Gill proclaimed that the next 10 years will be the decade in which quantum really comes of age. So what is quantum computing and how will it transform over the next decade? And in what ways will quantum computers change the way we interact with technology? Let's dive in.

Thanks for joining us, Dr. Gill. It's a real pleasure. Thank you for having me, Malcolm. It's a pleasure to be with you. I wanted to start by telling us a little bit about yourself. You did your graduate work at MIT. What did you study there?

I studied nanotechnology at MIT. I joined the nanostructures laboratory. It was in the department of electrical engineering and computer science under Professor Hank Smith, who was one of the pioneers of the field of nanofabrication. How do you manipulate and build the incredibly small part of our world? Yeah. And so that sets you up.

for the world that you're in now, right? This, the quantum computing flows naturally out of the idea that there's plenty of room at the bottom.

Yes, it is because of the different theories of physics. One that is, of course, extraordinarily relevant for the world of the small is quantum physics. So if we are to understand how matter behaves at the atomic scale and electronic structure and the interactions and what occurs at the level of materials, you have to understand what is happening in the world of quantum physics. For those who are

not from a technical background, can you give me the simplest definition of what quantum computing is? You know, we're all accustomed to using computers. And the foundation of the computers we use every day and our smartphones is the idea of bits or binary digits.

Interestingly, this is an idea that the fascination that all the complexity in the world, we could reuse it in a mode of communication with just zeros and ones dates back as far back as life means. But it was really in the 2040s.

in the 1940s and 50s that Cloud Shannon, which is one of the great leaders in the world of computing, told us we could create these incredibly sophisticated modes of communication and computation just by being able to map all the complexity and information in the world to strings of zeros and ones. And computers are machines that manipulate zeros and ones very, very efficiently. So in quantum computing, it actually revisits that idea

It turns out that the most fundamental building block of computation is not the zero and one, not the bit, but something known as the qubit, short for quantum bit. And at its heart, it melts that idea of information with the idea of physics. So what quantum computers do is they manipulate information, exploiting the laws of quantum physics to be able to do calculations that are simply impossible to do.

If you just use the binary digit, the zeros and ones, it's a richer way to represent information and manipulate information by exploiting the properties of quantum mechanics to do things that are impossible, basically. Yeah. Would you say...

you can tackle problems now that would be impossible mechanically. What is that? Can you represent the difference in capacity of these two ways of computing? How much of a gap is there between quantum and conventional computing? In its full potential, the gap is an exponential. So let me explain what I mean by that. If you want to simulate

So let's say, to be very practical, that you want to build a better battery technology for electric cars. So those are based on lithium chemistry. And if you want to, say, build a battery that is longer lasting or safer or can charge faster. So now you have in front of you a material science problem.

And what you can do is go through the periodic table and see all the different elements and figure out how you are going to combine them to create the material that has the properties you like. Okay, so how can you go about doing that? Well, one approach is to do it empirically. Just try. And humans have been doing that since time immemorial, right? Combine elements and see how it works.

Another methodological approach is if you have a theory of how things work, you could try to solve the problem long form and see if you can have a closed form solution to the problem. And a third angle that really came about with the advent of computers is you could simulate it, right? You could use computers to mimic how atoms behave and use those equations and try to do the calculations to see what the properties would be. What's the problem?

The problem is that no matter how big computers we use today, the number of variables that we have to compute over is roughly correlated to the number of electrons and electron orbitals present in those elements. So the more sophisticated a material we got to make, the more interactions between these electrons we got to be able to calculate.

And that number grows exponentially. Pretty soon, we need to have a computer with more components than there are atoms in the universe. So it's impractical. So what do we do? We approximate things. And when we approximate, we don't get the right answer. So we are in this stuck loop of rate of progress. What is interesting on quantum is that for modeling those kinds of problems,

Instead of having an exponential, meaning the more electrons we add, you know, the number of calculations blowing up. Now it's a relation that looks more like linear, meaning I only need one more qubit, roughly speaking, to model another electron.

So even if I have a complex molecule where I need, you know, dozens or hundreds of orbital calculations that I need to do, I will need a machine with dozens of hundreds of qubits rather than a classical machine with 10 trillion transistors, right, that we don't know how to make.

Oh, I see. So it's not an extension or a derivative from the kind of computers that we've been using. It's an entirely new class of computers. That's exactly right. And that is what's so interesting. So there will be classical computing and quantum computing. That's how important this is, right? That you phrased it very nicely, which is not just another evolution, is we've actually left no element of the assumption of the current information computational model as sacred, right? Not even the bit, right?

has survived the quantum information view of the world, right? The very foundation had to be revisited. So where are we? How close are we to having quantum computers? Actually, you described that task of trying to figure out how to make a better battery. When do you think we'll be able to use quantum computers for a task like that?

We already have built quantum computers. Actually, IBM was the first company in the world in 2016 to build a small quantum computer and make it universally available. So the first part of the answer is like we already have quantum computers. So you can learn how to program them. You can start mapping problems around how you do it. The challenge we have is that very difficult to build these machines. So we have not yet crossed the path.

where they can do things that are of practical value that our classical machines cannot. So we got to keep an eye now of when that crossover is going to happen. And that is something that is this frontier of information computation that will likely happen in the next few years. And then that begins a whole new space of opportunity. Yeah.

You had said earlier that we're at the stage now where the machines make errors. What's the source of the difficulty at the moment? Yeah, Tim, you can build these machines that have special properties to represent information in unique ways that gives them exponential power compared to classical machines.

The machines are subject to errors, but there's both a theory and a way to implement an error correction technique that would allow us to compute indefinitely and with like minimum levels of errors such that it would be a practical value.

But realizing that large scale machine is still a significant journey with a lot of scientific and engineering breakthroughs need to occur. But what we know is that it's theoretically sound and possible and that we are making very significant progress towards achieving that goal. And that's why this is a quest of doing something that has never been done before. But it is definitely possible.

Yeah, yeah. You began with the example of the electric battery. Give me another example of an industry or a problem which would be well served by using a quantum computer.

There's three categories that quantum is going to make a difference. Simulating nature, the world of mathematics, linear algebra that matters to machine learning and other problems, and the third is world of search and graphs and what you can do with them that matter a great deal. But I want to give an example that's very famous of the consequences of quantum computing, which is some of the implications that it will have for cybersecurity and for security in general.

And this came from a very, very famous algorithm that gave re-energizing of the field of quantum in the 1990s called Shor's algorithm. And it came from Peter Shor, who was then at Bell Labs and now is a professor at MIT. And he published an algorithm

That took a problem that has to do with factoring. So basically the problem is if you take two prime numbers that are, say, large, and you multiply those two prime numbers together, and you get the product, the final number of the product of those two. If doing the multiplication is very easy to do, anybody can do it, right? It's just multiplying two numbers. But if I gave you the product...

And I ask you, can you tell me, given this number, what two prime numbers compose that product? Turns out to be very, very computationally expensive. And in his algorithm, he showed that if you had a sufficiently large quantum computer, you could do this efficiently. And you say, well, why does that matter? Well, it turns out it matters because that's the basis of how we do encryption today.

and how we secure all forms of communications and financial systems and everything, where basically your private key is your prime number. I would have another private key, which is my prime number. Those two numbers are secret. And when we multiply them, that's the public key that we share over the Internet and so on, the protocol. Everybody sees our public key, but they cannot calculate our private keys. But if you had a large enough quantum computer, now you could.

So there's a big implication that the encryption protocols of the world need to be changed to prevent future quantum computers from decryption. So that's not the fault of quantum computers, but it's an example of the consequences. We build all sorts of assumptions in the world about what problems are easy and hard to do mathematically. And this technology will alter that equation. Yeah, yeah. There was something I was thinking about when you were talking about

I was imagining the world of clinical trials of a promising new drug, which are now conducted in exactly the same way, basically, as they were conducted 100 years ago. You put it in people and observe differences between, you know, the experimental arm and the control arm. And it's because the task of modeling a drug's interaction with very, very different

human beings is too complicated. Is this the kind of thing that down the road we might be able to simulate a drug trial? Yes, the opportunities of these advancements is to accelerate discovery. The rate of time from invention to realizing the capability to compress it very significantly, perhaps by a factor of 10x in time or 10x in cost.

So, yeah, you're absolutely right that the only path that we have to improve our experimental capacity to be able to determine and compress how efficient we can do it is to be more sophisticated as to how much we need to test.

And the trade-off, the thing that you can balance is if I can compute essentially to do virtual experiments, but to do it with the level of accuracy that is required and the level of fidelity that we would see in the real world, then it's a net gain. And indeed, that is going to be one of the main vectors of use cases and applications for quantum computing. I was thinking about the mRNA COVID vaccines, which were

Conceived and developed using the most cutting edge science imaginable and then tested using the least cutting edge science imaginable. You went from this dazzling feat of 21st century genetic biomedicine and then you painstakingly rounded up people.

brought them in, gave them shots, asked them questions, had them fill out forms. I mean, it's like, and that's also an example of this. You just brought up this thing about what happens when we combine this new technology with existing technologies. In that hypothetical case, that's a combination. You're taking this

brand new field of biomedicine and marrying it to a way of revolutionizing the clinical aspect of medicine. That's two systems in combination create a kind of exponential change in your outcome.

Yeah. And that's the part that we always struggle as humans, right? Because since, you know, time progresses linearly for us, the fact that there's these exponentials in the form of technology is an example. But we've also seen exponentials. I think people are understanding a little bit better, you know, tragically in this case, the power of exponentials in the context of a pandemic.

But the fact that these exponentials are present in our world and our universe and that through technology you get these combinations of technology that allows you to create them

It's something that is both the source of massive opportunity and aspects that have to do with governance of how we need to be smart enough to be able to guide them properly. But you're right. I mean, that's a good example what you brought up in terms of an experimental capability of mRNA. And interestingly enough,

the sort of more theoretical unification of some of these ideas is that mRNA technology is again rooted on the idea that biology is information.

And that if we're able to, in this case, decode, like in this case involves the genetic sequencing of the virus, and from there figure out what parts of the code I need to bring back into your immune system to be able to fight it efficiently, in a ways about being able to read information, process information, send it back to you, and you yourself are the computer, right, with this new program to deal with the biology of it. So,

bringing information on that, on how efficiently we compute it, you know, how you conduct clinical trials, all of that aspect of it is the opportunity to have more mastery over our environments. Yeah. Can I ask you a personal question? If you look over the history of technology, every now and again, there are people who are in these magical moments where they are aware that the thing they're working on is going to dramatically transform the world they live in.

You can imagine someone working in Edison's lab or someone working in the Manhattan Project in the desert in, you know, in 1943. Or, you know, we can all identify you're in that position.

You know, I believe that to my core. And I indeed, like, I feel that way. And the team feels this way, that we have assembled a team that is the finest team in the world, that it is designing and imagining and creating these quantum computers. And there's not a doubt in our mind that as difficult as this quest is,

It has that potential. It's one of those things that it answers the equation of what is possible to do with technology. It is one of these things that will be definitely in the history books in terms of information and computation and what it means.

And I think that that brings us an enormous amount of energy into us, right? Because when we come to work every day and when we see the progress we're making is this feeling of being absolutely at the cutting edge where every day that the team makes progress is the actual boundary of knowledge that

and possibilities in the field. And it just feels magical, right? And both our successes and the challenges as we push forward, you know, are colored by this passion of saying, boy, but this is a frontier of humanity. And, you know, and we're all working together to, you know, do it as well as we know how to.

Let's explore this idea, the potential of combining these different computing forms. Give me some more practical examples of what combinations look like. If we're going to put them in the proper context, what's happening with technologies like quantum and AI, I like to say that they need to fit in the context of a method. And the method that we're most passionate about is not a new one. It's a scientific method.

Our thesis is that we should expand the reach of the scientific method. And for the most important problems that we're confronting, let's take global warming or fighting pandemics, accelerating the rate of discovery is incredibly important, right? This aspect of time. So here's the question. How can these advances in computing accelerate the scientific method?

So let's peel the layer. What is behind the scientific method if we look at it very, very simply? We would say it's the act of learning from the past. So you've got to know and exploit the knowledge that has been accumulated. That is typically in the form of documents and books, etc. You need to then be able to generate hypotheses that can be verified or nullified. You've got to conduct experiments and then you've got to share it

with a community for feedback and go through the loop again. You say, well, how can these technologies help you? Take the first one.

to search and learn from the past. So AI, in the form of natural language processing, in the form of being able to process documents and build huge graphs with which to search knowledge that already existed, is greatly helping us. I mean, we live it in a day-to-day life by, you know, like the power of search, right, of information on the web. But as scientists, you

You can do this if you can greatly enhance the ability to read scientific literature and see its connections and help you as a scientist acquire information fast. So that's a use of AI for the search. Then you go to the next step, generate a hypothesis. Well, to generate hypotheses, there's a beautiful new area in AI called generative models.

We may be a little bit more familiar with the use of neural networks inside AI to do the task of classification, right? If I give you images, you give me labels, right? Say, well, this is a yellow car, a red car, and so on, and it gets done with a neural network.

Perhaps people are less familiar with using now some of these neural networks to do generation in terms of classification. So I give you, you know, I say, hey, design to me a chair that looks like an avocado, right? And the system can automatically give you hundreds of thousands of different designs and so on, right?

So now you can use this generative capability to imagine new molecules back to connected to our idea about chemistry and lithium chemistry. So I have these properties I want. Give me molecules that may fit that criteria.

And if I have AI that creates these generative models, I want to verify whether they may work as they want. So now I can use a quantum computer, right, in the future to say, do they work like they say, because I'm simulating a model of chemistry. So now I'm combining AI and quantum and simulation to be able to do this better. Then the next step is, well, let's realize it in practice. Let's do experimentation now.

So I can have robots that synthesize with chemistry that are AI guided to optimally create the synthetic route and the programming steps with which to create the molecules and so on. So I like to think about is take a method that we know works, the scientific method.

Think about it as a method and now ask yourself how the loop of technologies that we're creating can enhance it and improve it in concert with scientists and humans. And that is what I think is going to have revolutionary potential. Because, and I'll close with the idea of what a difference it made. You brought up mRNA. What a difference it made to have the tools available to us to compress the time to discovery from the average time of 14 years for a vaccine to under one.

And think of the implications that have. Well, in future pandemics, in climate change, how are we going to compress that time to discovery? And that's going to be the power of the scientific methods supercharged with computing, including quantum and AI. One last quick question. It sounds like it is impossible to be a pessimist and work on quantum computing.

I like that. So probably true, you know, because when you have so many challenges and so many difficulties, it takes a particular type of people to have the courage to be able to overcome them.

But it gets combined when you know that the theory is very sound and correct. The fact that we haven't been able to realize the technology that allows a theory to be expressed is...

in itself, a source of energy, right? And indeed, like you cannot be, you know, a pessimist if you want to be at the vanguard of the creation of this technology. And also the implications of it are so profound for some of the most fundamental problems that that is another source of optimism required for the technology. Yeah. This has been so fascinating. Thank you so much. I've really enjoyed this, Dr. Gill. Thank you so much.

Thank you again to Dr. Dario Gil for his insights about the future of quantum computing. It will be fascinating to see how the conversions of old and new can revolutionize the way we live and communicate. Smart Talks with IBM is produced by Emily Rostak with Carly Migliori and Catherine Giraudoux. Edited by Karen Shakerji.

Engineering by Martin Gonzalez. Mixed and mastered by Jason Gambrell. Music by Gramascope. Special thanks to Molly Socha, Andy Kelly, Mia LaBelle, Jacob Weisberg, Hedda Fane, Eric Sandler, and Maggie Taylor, and the teams at 8 Bar and IBM. Smart Talks with IBM is a production of Pushkin Industries and iHeart Media. You can find more episodes at ibm.com slash smarttalks.

You'll find more Pushkin podcasts on the iHeartRadio app, Apple Podcasts, or wherever you like to listen. I'm Malcolm Gladwell. See you next time.

Hello, hello. Malcolm Gladwell here. I want to tell you about a new series we're launching at Pushkin Industries on the 1936 Olympic Games. Adolf Hitler's Games. Fascism, anti-Semitism, racism, high Olympic ideals, craven self-interest, naked ambition, illusion, delusion, all collide in the long, contentious lead-up to the most controversial Olympics in history. The Germans put on a propaganda show, and America went along with all of it. Why?

This season on Revisionist History, the story of the games behind the games. Listen to this season of Revisionist History wherever you get your podcasts. If you want to hear episodes before they're released to the public, subscribe to Pushkin Plus on Apple Podcasts or at pushkin.fm slash plus.