WEBVTT - Smart Talks with IBM and Malcolm Gladwell: Quantum Computing is Tomorrow’s Computing, Today 0:00:04.760 --> 0:00:08.640 Hello, Hello. This is Smart Talks with IBM, a podcast 0:00:08.920 --> 0:00:13.319 from Pushkin Industries, High Heart Media and IBM about what 0:00:13.320 --> 0:00:16.240 it means to look at today's most challenging problems in 0:00:16.280 --> 0:00:22.960 a new way. I'm Malcolm Gladwell. In this episode, I'll 0:00:22.960 --> 0:00:27.360 be discussing the capabilities of quantum computing with Dr Dario 0:00:27.480 --> 0:00:31.520 gil Dr Gill is the senior Vice president and director 0:00:31.840 --> 0:00:36.440 of IBM Research. He's also recognized globally as one of 0:00:36.440 --> 0:00:41.400 the brightest minds in the quantum computing industry. But what 0:00:41.560 --> 0:00:44.720 we know is that is theoretically sound and possible, and 0:00:44.800 --> 0:00:48.560 that we are making very significant progress towards achieving good goal. 0:00:48.600 --> 0:00:51.280 And that's why this is a quest of doing something 0:00:51.320 --> 0:00:54.240 that has never been done before, but it is definitely possible. 0:00:54.720 --> 0:00:57.880 Earlier this year, at the Wall Street Journals Virtual CIO 0:00:58.040 --> 0:01:01.480 Network Summit, Dr Gil pro claimed that the next ten 0:01:01.560 --> 0:01:05.080 years will be the decade in which quantum really comes 0:01:05.080 --> 0:01:08.920 of age. So what is quantum computing and how will 0:01:08.920 --> 0:01:12.520 it transform over the next decade and in what ways 0:01:12.560 --> 0:01:17.400 will quantum computers change the way we interact with technology. 0:01:18.800 --> 0:01:29.240 Let's dive in. Thanks for joining us. Start to kill. 0:01:29.440 --> 0:01:32.600 It's a real pleasure thank you for having me, Malcolm, 0:01:32.640 --> 0:01:34.720 it's a pleasure to be with you. I wanted to 0:01:35.720 --> 0:01:37.880 start my talent us a little bit about yourself. You 0:01:37.920 --> 0:01:41.160 did your graduate work at m I T. What did 0:01:41.160 --> 0:01:45.800 you study there? I studied nanotechnology at m i T. 0:01:46.440 --> 0:01:49.440 I joined the Nanostructure's laboratory. It was in the Department 0:01:49.440 --> 0:01:53.520 of Electrical Engineering and Computer Science and their professor Hank Smith, 0:01:54.000 --> 0:01:56.840 who was one of the pioneers of the field of 0:01:56.920 --> 0:02:00.000 nano fabrication. You know, how do you manipulate and build 0:02:00.640 --> 0:02:04.640 incredibly small part of our world? And so that sets 0:02:04.680 --> 0:02:08.680 you up for the world that you're in now. Right 0:02:08.760 --> 0:02:14.520 this the quantum computing flows naturally out of the idea 0:02:14.560 --> 0:02:17.400 that this plenty of room at the bottom. Yes, it 0:02:17.560 --> 0:02:22.160 is because of the different theories of physics. One that 0:02:22.280 --> 0:02:25.040 is of course extraordinarily irrelevant for the world of the 0:02:25.160 --> 0:02:29.960 small is quantum physics. So if we are to understand 0:02:31.280 --> 0:02:35.720 how matter behaves at the atomic scale and electronic structure 0:02:35.720 --> 0:02:38.760 and the interactions and uh what occurs at the level 0:02:38.760 --> 0:02:42.440 of materials, you have to understand what is happening in 0:02:42.480 --> 0:02:46.280 the world of quantum physics. For those who are not 0:02:46.440 --> 0:02:49.720 from a technical background, can you give us the give 0:02:49.800 --> 0:02:53.880 me this the simplest definition of what quantum computing is. 0:02:54.840 --> 0:02:57.799 You know, we're all accustomed to using computers today, and 0:02:57.840 --> 0:03:01.200 at the foundation of the computers we use every day 0:03:01.240 --> 0:03:05.720 and our smartphones is the idea of bits or binary digits. 0:03:05.919 --> 0:03:09.440 And interestingly, this is an idea that the fascination that 0:03:09.600 --> 0:03:12.640 all the complexity in the world we could reuse it 0:03:12.919 --> 0:03:15.640 in a mode of communication, which is zeros and ones 0:03:15.919 --> 0:03:18.920 dates back as far back as Lightnings, but it was 0:03:19.000 --> 0:03:22.320 really in the twenty century. In the nineteen fourties and fifties, 0:03:22.400 --> 0:03:26.360 a Cloud Shannon, which is one of the great leaders 0:03:26.360 --> 0:03:29.640 in the world of computing, told us we could create 0:03:29.760 --> 0:03:34.560 these incredibly sophisticated modes of communication and computation just by 0:03:34.600 --> 0:03:37.560 being able to map all the complexity and information in 0:03:37.560 --> 0:03:39.920 the world to strings of zeros and ones, and computers 0:03:39.920 --> 0:03:43.960 are machines that manipulate zeros and ones very very efficient. 0:03:44.000 --> 0:03:48.120 So in quantum computing it actually revisits that idea. It 0:03:48.160 --> 0:03:51.480 turns out that the most fundamental building block of computation 0:03:51.600 --> 0:03:53.800 is not the zero on one, not the bit. That's 0:03:53.840 --> 0:03:57.560 something known as the cubit sure for quantum bit, and 0:03:57.720 --> 0:04:00.920 at its heart it melts that idea ye of information 0:04:01.760 --> 0:04:06.240 with the idea of physics. So what quantum computers do 0:04:06.320 --> 0:04:12.440 is they manipulate information, exploiting the loss of quantum physics 0:04:12.680 --> 0:04:17.119 to be able to do calculations that are simply impossible 0:04:17.160 --> 0:04:20.640 to do if you just use the binary digit the 0:04:20.720 --> 0:04:24.240 zeros and ones. It's a richer way to represent information 0:04:24.760 --> 0:04:29.200 and manipulate information by exploiting the properties of quantum mechanics 0:04:29.200 --> 0:04:33.480 to do things are impossible classically. Yeah, would you say 0:04:34.200 --> 0:04:37.120 you can tackle problems now that would be impossible mechanically. 0:04:37.480 --> 0:04:41.160 What is that? Could you represent the difference in capacity 0:04:41.200 --> 0:04:44.080 of these two ways of computing, like how how much 0:04:44.120 --> 0:04:47.640 of a gap is here between quantum and conventional computer? 0:04:48.120 --> 0:04:50.480 You know, and it's full potential the gaps it's an 0:04:50.480 --> 0:04:54.280 exponential So let me let me explain what I mean 0:04:54.320 --> 0:04:59.400 by that. If you want to simulate nature, so let's 0:04:59.440 --> 0:05:02.240 say very practical that you want to build a better 0:05:02.279 --> 0:05:05.880 battery technology for electric cons right, so those are based 0:05:05.880 --> 0:05:09.360 on lithium chemistry. And if you want to say, build 0:05:09.960 --> 0:05:13.960 a battery that is longer lasting or safer, contract faster. 0:05:14.400 --> 0:05:16.560 So now you have in front of your material science 0:05:16.600 --> 0:05:19.520 problems and what you can do is go through the 0:05:19.560 --> 0:05:22.440 periodic table and see all the different elements and figure 0:05:22.480 --> 0:05:24.919 out how you are going to combine them to create 0:05:24.920 --> 0:05:27.800 the material that has the properties you like. Okay, so 0:05:27.839 --> 0:05:30.760 how can you go about doing that? Well? One approach 0:05:31.240 --> 0:05:34.480 is to, uh, do it empirically, just try and humans 0:05:34.480 --> 0:05:37.000 have been doing that since time me memorial right, combine 0:05:37.080 --> 0:05:41.120 elements and see hy works. Another methodological approach is if 0:05:41.120 --> 0:05:43.520 you have a theory of how things work, you could 0:05:43.560 --> 0:05:46.120 try to solve the problem long form and see if 0:05:46.160 --> 0:05:48.839 you can have a close form solution to the problem. 0:05:48.880 --> 0:05:52.200 And a third angle that really came about with the 0:05:52.200 --> 0:05:54.960 advent of computers is you could simulate it. Right. You 0:05:55.000 --> 0:05:59.440 could use computers to mimic how atoms behave and use 0:05:59.480 --> 0:06:01.880 those equation issues and try to do the calculations to 0:06:02.000 --> 0:06:05.599 see what the properties would be. What's the problem. The 0:06:05.680 --> 0:06:09.640 problem is that no matter how big computers we use today, 0:06:09.760 --> 0:06:13.520 the number of variables that we have to compute over 0:06:14.200 --> 0:06:18.600 is roughly correlated to the number of electrons and electron 0:06:18.720 --> 0:06:24.240 orbitals present in those elements. So the more sophisticated and 0:06:24.279 --> 0:06:28.160 material we've gotta make, the more interactions between these electrons 0:06:28.240 --> 0:06:31.920 we gotta be able to calculate, and that number grows exponentially, 0:06:32.120 --> 0:06:36.200 you know, pretty soon we need to have a computer, 0:06:36.560 --> 0:06:39.320 you know, with more components than there are atoms in 0:06:39.360 --> 0:06:42.200 the universe. Right, so it's impractical. So what do we do. 0:06:42.240 --> 0:06:44.920 We approximate place, and when we approximate, we don't get 0:06:44.920 --> 0:06:47.880 the right answer. So we are in this stock loop 0:06:48.279 --> 0:06:51.880 of rate of progress. What is interesting on quantum is 0:06:51.920 --> 0:06:55.760 that for modeling those kinds of problems, instead of having 0:06:56.360 --> 0:06:59.680 an exponential meaning the more electrons we add, you know, 0:07:00.040 --> 0:07:04.240 the number of calculations blowing up. Now it's a relation 0:07:04.320 --> 0:07:08.080 that looks more like linear, meaning I only need one 0:07:08.120 --> 0:07:12.800 more cubit roughly speaking, to model another electron. So even 0:07:12.840 --> 0:07:16.000 even have a complex molecule where I need you know, 0:07:16.280 --> 0:07:21.000 dozens or hundreds of of orbital calculations and I need 0:07:21.000 --> 0:07:23.520 to do I would need a machine with dozens of 0:07:23.600 --> 0:07:29.080 hundreds of cubits rather than a classical machine with ten 0:07:29.200 --> 0:07:31.600 trillion transistors. Right that we don't know how to make. 0:07:33.280 --> 0:07:36.320 It's not an extension or a derivative from the kind 0:07:36.320 --> 0:07:39.400 of computers that we've been using. It's an entirely new 0:07:39.480 --> 0:07:42.920 class of computers. That's exactly right, And that is what's 0:07:42.920 --> 0:07:46.560 so interesting. So there will be classical computing and quantum computing. 0:07:47.080 --> 0:07:50.200 That's how important this is, right, that you phrased it 0:07:50.320 --> 0:07:54.040 very nicely, which is not just another evolution. Is we've 0:07:54.080 --> 0:07:56.880 actually left no element of the assumption of the current 0:07:56.880 --> 0:08:00.840 information on computational model as sacred. Right, not even the 0:08:00.920 --> 0:08:05.720 bit has survived the quantum information view of the world. Right, 0:08:05.880 --> 0:08:09.240 the very foundation had to be revisited. So where are 0:08:09.360 --> 0:08:13.960 we How close are we having quantum computers? Actually that 0:08:14.040 --> 0:08:17.280 you you describe that that task kept trying to figure 0:08:17.320 --> 0:08:19.280 out how to make a better battery. When do you 0:08:19.320 --> 0:08:21.320 think we'll be able to use quantum computers for a 0:08:21.320 --> 0:08:26.600 task like that? We already have built quantum computers. Actually, 0:08:26.680 --> 0:08:28.600 Iban was the first company in the world in two 0:08:28.600 --> 0:08:30.840 thousand and sixteen to build a small quantum computer and 0:08:30.880 --> 0:08:34.200 making universally available. So the first part of the answer 0:08:34.280 --> 0:08:36.960 is like, we already have quantum computers, so you can 0:08:37.040 --> 0:08:39.960 learn how to program them. You can start mapping problems 0:08:40.080 --> 0:08:42.560 around how you do it. The challenge we have is 0:08:42.600 --> 0:08:45.760 that very difficult to build these machines, so we have 0:08:45.920 --> 0:08:50.040 not yet crossed the path where they can do things 0:08:50.120 --> 0:08:54.000 that are of practical value that or classical machines cannot. 0:08:54.320 --> 0:08:56.640 So we gotta keep an eye now of when that 0:08:56.720 --> 0:08:59.439 crossover is going to happen, and that is something that 0:08:59.559 --> 0:09:03.000 is this tier of information and computation that would likely 0:09:03.040 --> 0:09:07.080 happen in the next few years, and then that begins, 0:09:07.200 --> 0:09:10.680 you know, a whole a whole new space right of opportunity. Yeah, 0:09:11.480 --> 0:09:14.520 you had said earlier with the stage now where the 0:09:14.600 --> 0:09:17.640 machines make errors, what's the source of the difficulty at 0:09:17.640 --> 0:09:22.400 the moment. Yeah, then you can build these machines that 0:09:22.520 --> 0:09:26.200 have special properties to represent information in unique ways that 0:09:26.240 --> 0:09:30.199 gives them exponential power compared to classical machines. The massines 0:09:30.360 --> 0:09:34.280 are subject to errors, but there's both a theory and 0:09:34.320 --> 0:09:37.880 a way to implement an error correction technique that would 0:09:37.880 --> 0:09:42.080 allow us to compute in definitely and with like minimum 0:09:42.160 --> 0:09:44.560 levels of errors. That's that it would be a practical value. 0:09:45.040 --> 0:09:49.600 But realizing that large scale machine is still a significant 0:09:49.640 --> 0:09:52.320 journey with a lot of scientific and engineering breakthroughs need 0:09:52.360 --> 0:09:54.679 to occur. But what we know is that it's theoretically 0:09:54.720 --> 0:09:58.520 sound and possible, and that we are making very significant 0:09:58.559 --> 0:10:01.040 progress towards achieving the goal. And that's why this is 0:10:01.040 --> 0:10:04.360 a quest of doing something that has never been done before, 0:10:04.400 --> 0:10:08.480 but it is definitely possible. Yeah, you began with the 0:10:08.720 --> 0:10:12.240 example of the electric battery. Give me another example of 0:10:12.320 --> 0:10:16.480 a of an industry or a problem which would be 0:10:16.520 --> 0:10:22.080 well served by using a quantum computer. These three categories 0:10:22.360 --> 0:10:24.920 that quantum is going to make a difference similar in nature. 0:10:25.400 --> 0:10:28.600 The world of mathematics, linear algebra that matters to maschine 0:10:28.640 --> 0:10:31.200 learning and other problems. And the third is world of 0:10:31.240 --> 0:10:33.880 search and graphs and what you can do with them 0:10:33.960 --> 0:10:35.920 that matter a great deal. But I want to give 0:10:36.120 --> 0:10:40.360 an example that's very famous of the consequences of quantum computing, 0:10:40.720 --> 0:10:42.680 which is some of the implications that it will have 0:10:42.760 --> 0:10:46.720 for cybersecurity and for security in general. And this came 0:10:46.760 --> 0:10:51.000 from a very very famous algorithm that gave re energizing 0:10:51.160 --> 0:10:54.640 of the field of quantum in the ninety nineties called 0:10:54.679 --> 0:10:58.240 Short's algorithm, and it came from Peter Short, who was 0:10:58.400 --> 0:11:01.000 then at bad LAPS and now is a professor at 0:11:01.120 --> 0:11:06.000 m I T. And he published an algorithm that took 0:11:06.040 --> 0:11:10.360 a problem that has to do with factoring. So basically, 0:11:10.400 --> 0:11:13.440 the problem is if you take two prime numbers that 0:11:13.520 --> 0:11:17.439 are say large, and you multiply those two prime numbers together, 0:11:18.120 --> 0:11:20.800 and you get the product, the final number of the 0:11:20.840 --> 0:11:25.520 product of those of those two If doing the multiplication, 0:11:25.760 --> 0:11:27.720 it's very easy to do. Anybody can do it, right. 0:11:27.760 --> 0:11:30.040 It's just multiplying two numbers. But if I give you 0:11:30.080 --> 0:11:33.280 the product and I ask you, can you tell me, 0:11:33.800 --> 0:11:38.200 given this number, what two prime numbers composed? That product 0:11:38.200 --> 0:11:42.520 turns out to be very, very computationally expensive. And he 0:11:42.840 --> 0:11:45.559 in his algorithm, he showed that if you had a 0:11:45.600 --> 0:11:49.160 sufficient in large quantum computer you could do this efficiently. 0:11:49.520 --> 0:11:52.120 And you say, well, what does that matter, Well, it 0:11:52.160 --> 0:11:54.840 turns out it matters is because as the basis of 0:11:54.880 --> 0:11:58.439 how we do encryption today and how we secure all 0:11:58.520 --> 0:12:03.360 forms of communications and financial systems and everything, where basically 0:12:03.960 --> 0:12:07.680 your private key is your prime number. Malcome, I would 0:12:07.720 --> 0:12:10.240 have another private key which is my prime number. Those 0:12:10.240 --> 0:12:13.920 two numbers are secret, and when we multiply them, that's 0:12:13.960 --> 0:12:16.240 the public key that we share over the Internet and 0:12:16.240 --> 0:12:19.440 so on the protocol. Everybody sees our public key, but 0:12:19.520 --> 0:12:22.560 they cannot calculate or private keys. But if you had 0:12:22.600 --> 0:12:26.160 a large enough quantum computer, now you could. So there's 0:12:26.160 --> 0:12:30.360 a big implication that the encryption protocols of the world 0:12:30.720 --> 0:12:36.520 need to be changed to prevent future quantum computers from decryption. 0:12:37.320 --> 0:12:40.160 So that's not the fault of quantum computers, but it's 0:12:40.160 --> 0:12:43.920 an example of the consequences we build all sorts of 0:12:43.960 --> 0:12:46.400 assumptions in the world about what problems are easy and 0:12:46.480 --> 0:12:50.360 hard to do mathematically, and uh, and this technology will 0:12:50.400 --> 0:12:54.440 alter that equation. Yeah. Yeah, But there was something I 0:12:54.480 --> 0:12:57.439 was thinking about when you were talking. I was imagining 0:12:57.960 --> 0:13:01.920 the world of clinical trials of a promising new drug, 0:13:02.240 --> 0:13:06.200 which are now conducted in exactly the same way basically 0:13:06.240 --> 0:13:09.120 as they were conducted hundred years ago. You put it 0:13:09.120 --> 0:13:13.160 in people and observe differences between you know, the experimental 0:13:13.160 --> 0:13:17.280 i'm and a control are. And it's because the task 0:13:17.360 --> 0:13:23.960 of modeling drugs interaction with very very different human beings 0:13:24.080 --> 0:13:26.680 is too complicated. This is the kind of thing that 0:13:26.760 --> 0:13:29.800 down the road we might be able to simulate a 0:13:29.880 --> 0:13:36.560 drug trial. Yes, some the opportunities of of these advancements 0:13:36.600 --> 0:13:41.200 is to accelerate discovery, the rate of time from invention 0:13:41.360 --> 0:13:46.320 to realizing the capability to compress it very significantly, perhaps 0:13:46.400 --> 0:13:49.400 by a factor of ten x in time or ten 0:13:49.600 --> 0:13:54.040 x in cost. So um, yeah, you're You're absolutely right, 0:13:54.440 --> 0:13:58.719 Uh that the only path that we have to improve 0:13:59.280 --> 0:14:02.640 or expert mental capacity to be able to determine and 0:14:02.720 --> 0:14:04.800 compress how efficient we can do it is to be 0:14:04.840 --> 0:14:07.280 more sophisticated as to how much we need to test, 0:14:08.080 --> 0:14:11.000 and they trade off. The thing that you can balance 0:14:11.080 --> 0:14:14.479 is if I can compute essentially to do virtual experiments, 0:14:15.080 --> 0:14:17.200 but to do it with the level of accuracy that 0:14:17.360 --> 0:14:20.600 is required and the revel of fidelity that we would 0:14:20.600 --> 0:14:23.000 see in the real world, then it's a net game. 0:14:23.480 --> 0:14:26.760 And indeed that is going to be you know, one 0:14:26.760 --> 0:14:29.240 of the main vectors of use cases and applications for 0:14:29.360 --> 0:14:32.240 quantum computer. I was thinking about the the m R 0:14:32.280 --> 0:14:36.480 and A COVID vaccines, which were conceived, they developed using 0:14:37.160 --> 0:14:42.120 the most cutting edge science imaginable, and then tested using 0:14:42.200 --> 0:14:45.440 the least cut edge science. Right. You went from this 0:14:45.640 --> 0:14:51.640 dazzling feed of century, you know, genetic biomedicine, and then 0:14:51.720 --> 0:14:57.280 you painstakingly rounded up people, brought them in, gave them shots, 0:14:57.480 --> 0:14:59.520 you know, ask them questions, had them fill out fill 0:14:59.560 --> 0:15:02.560 out four. I mean, it's like and that's also an 0:15:02.600 --> 0:15:05.160 example of this. You just you just brought this this 0:15:05.360 --> 0:15:08.960 thing about what happens when we combine this new technology 0:15:09.000 --> 0:15:12.840 with existing technologies. In that hypothetical case, that's a combination. 0:15:12.880 --> 0:15:18.800 You're taking this brand new field of biomedicine and marrying 0:15:18.840 --> 0:15:22.840 it to a a way of of revolutionizing the clinical 0:15:22.880 --> 0:15:28.000 aspect of medicine. That's two systems in combination create a 0:15:28.080 --> 0:15:33.240 kind of exponential change in your outcome. Yeah, and and 0:15:33.240 --> 0:15:36.080 and that's the part that we always struggle as humans, 0:15:36.160 --> 0:15:39.560 right because since you know, time progresses linearly for us, 0:15:40.080 --> 0:15:42.560 the fact that there's these exponentials in the form of 0:15:42.600 --> 0:15:45.560 technology is an example. But we've also seen exponentials. I 0:15:45.560 --> 0:15:48.680 think people are understanding a little bit better, uh, you know, 0:15:48.760 --> 0:15:51.480 tragically in this case the power of exponentials in the 0:15:51.520 --> 0:15:55.320 context of a pandemic. But the fact that these exponentials 0:15:55.360 --> 0:15:58.000 are present in our in our world and our universe, 0:15:58.400 --> 0:16:02.359 and that through technology you get these combinations of technology 0:16:02.360 --> 0:16:06.320 that allows you to create them. It's something that is 0:16:06.360 --> 0:16:09.920 both the source of massive opportunity and aspects that have 0:16:10.000 --> 0:16:12.720 to do with with with governance of how we need 0:16:12.760 --> 0:16:15.320 to be smart enough to be able to guide them properly. 0:16:15.640 --> 0:16:17.640 But you're right, I mean, that's a good example where 0:16:17.640 --> 0:16:20.880 you brought up in terms of an experimental capability of 0:16:21.160 --> 0:16:23.920 m R n A and UH and and interestingly enough, 0:16:24.560 --> 0:16:27.880 the sort of more theoretical unification or some of these 0:16:27.880 --> 0:16:31.720 ideas is that MR and A technology is again rooting 0:16:31.760 --> 0:16:36.440 on the idea that biology is information and and that 0:16:36.840 --> 0:16:40.520 if we're able to in this case UH the code 0:16:40.640 --> 0:16:43.640 like in this case involves in the genetic sequencing of 0:16:43.760 --> 0:16:46.840 the virus and from there figure out what parts of 0:16:46.880 --> 0:16:49.040 the code I need to bring back into your immune 0:16:49.040 --> 0:16:51.360 system to be able to find it efficiently in a 0:16:51.440 --> 0:16:55.200 ways about being able to read information, process information, send 0:16:55.240 --> 0:16:58.360 it back to you and you yourself are are the 0:16:58.440 --> 0:17:01.400 computer right with this new program to deal with the 0:17:01.440 --> 0:17:05.200 biology of it, So bringing information on that and how 0:17:05.200 --> 0:17:08.440 efficiently we computed, you know, how you conduct clinical trials. 0:17:08.640 --> 0:17:11.160 All of that aspect of it is is the opportunity 0:17:11.200 --> 0:17:16.639 to have more mastery overall environments. Can you ask a 0:17:16.800 --> 0:17:21.120 personal question if you look over the history of technology, 0:17:21.400 --> 0:17:24.000 every now and again, there are people who are in 0:17:24.040 --> 0:17:27.840 these magical moments where they are aware that the thing 0:17:27.880 --> 0:17:31.760 they're working on is going to dramatically transform the world 0:17:31.760 --> 0:17:35.400 I live in. You can imagine someone working in Edison's 0:17:35.480 --> 0:17:39.479 lab or someone working in the Manhattan project in the 0:17:39.520 --> 0:17:44.480 desert in you know, in or you know that we 0:17:44.480 --> 0:17:49.160 can all identify you're in that position. You know, I 0:17:49.200 --> 0:17:52.800 believe that to my core, and I indeed like I 0:17:53.160 --> 0:17:56.800 feel that way, and the team feels this way. That 0:17:57.040 --> 0:18:00.399 we have assembled a team that is a fine esteem 0:18:00.440 --> 0:18:04.480 in the world that it is designing and imagining and 0:18:04.520 --> 0:18:08.520 creating these quantum computers. And there's not a doubt in 0:18:08.520 --> 0:18:11.560 our mind that as difficult as this quest is, it 0:18:11.720 --> 0:18:14.560 has that potential. It's one of those things that it 0:18:14.680 --> 0:18:18.240 answers the equation of what is possible to do with technology. 0:18:18.320 --> 0:18:20.640 It is one of these things that will be definitely 0:18:21.040 --> 0:18:24.560 in the history books in terms of information and computation 0:18:24.640 --> 0:18:28.119 and what it means. And I think that that brings 0:18:28.240 --> 0:18:32.040 us an enormous amount of energy into us, right because 0:18:32.040 --> 0:18:34.080 when we come to work every day and when we 0:18:34.080 --> 0:18:37.480 see the progress we're making, is this feeling of being 0:18:37.520 --> 0:18:40.159 absolutely at the cutting edge where every day that the 0:18:40.200 --> 0:18:44.520 team makes progress is the actual boundary of knowledge and 0:18:44.960 --> 0:18:48.800 possibilities in the field. And it just feels magical, right, 0:18:48.920 --> 0:18:51.919 And both our successes and the challenges as as we 0:18:52.040 --> 0:18:57.000 push forward, you know, are colored by this this passion 0:18:57.040 --> 0:18:59.480 of saying boy, but this is this is a frontier 0:18:59.520 --> 0:19:03.000 of human and uh, you know, and we're all working 0:19:03.000 --> 0:19:05.120 together to uh you know, do it as well as 0:19:05.119 --> 0:19:09.160 we know how to. Let's explore this idea, the potential 0:19:09.200 --> 0:19:13.760 of combining these different computing forms. Give me some more 0:19:13.760 --> 0:19:18.399 practical examples of what combinations look like if we're gonna 0:19:18.640 --> 0:19:22.639 put them in the proper context. What's happening with technologies 0:19:22.640 --> 0:19:27.320 like quantum and AI. I'd like to say that they 0:19:27.359 --> 0:19:30.040 need to fit in the context of a method. And 0:19:30.240 --> 0:19:33.439 the method that we're most passionate about, it's not a 0:19:33.480 --> 0:19:37.160 new one, is the scientific method. Or thesis is that 0:19:37.320 --> 0:19:42.120 we should expand the reach of the scientific method, and 0:19:42.200 --> 0:19:45.520 for the most important problems that we're confronting. Let's take 0:19:45.640 --> 0:19:51.560 global warming or fighting pandemics. Accelerating the right of discovery 0:19:52.760 --> 0:19:58.119 is incredibly important right this aspect of time. So here's 0:19:58.119 --> 0:20:01.960 the question, how can this advances in computing accelerate the 0:20:01.960 --> 0:20:05.679 scientific method? So let's peel the layer. What is behind 0:20:05.720 --> 0:20:08.520 the scientific method? If we look at it very very simply, 0:20:09.200 --> 0:20:12.199 we would say is the act of learning from the past. 0:20:12.800 --> 0:20:16.040 So you gotta you know, know and exploit the knowledge 0:20:16.040 --> 0:20:18.920 that has been accumulated that is typically in the form 0:20:19.000 --> 0:20:23.560 of documents, uh, books, etcetera. You need to then be 0:20:23.640 --> 0:20:28.200 able to generate hypothesis that can be verified or nullified. 0:20:28.880 --> 0:20:31.919 You've gotta conduct experiments and then you gotta share it 0:20:32.640 --> 0:20:35.679 with a community for feedback and go through the loop again. 0:20:36.160 --> 0:20:39.520 You say, well, how can these technologies help you? Take 0:20:39.520 --> 0:20:42.000 the first one to search and learn from the past. 0:20:42.200 --> 0:20:45.720 So AI in the form of natural language processing, in 0:20:45.760 --> 0:20:48.960 the form of being able to process documents and build 0:20:49.000 --> 0:20:52.520 huge graphs with which to search knowledge that already existed. 0:20:53.520 --> 0:20:56.159 It's greatly helping us. I mean we we live it 0:20:56.359 --> 0:20:58.520 in a day to day life by you know, like 0:20:58.600 --> 0:21:01.480 the power of search right of information the web. But 0:21:01.560 --> 0:21:05.240 as as scientists, you can do this if you can 0:21:05.560 --> 0:21:09.200 greatly enhance the ability to read scientific literature and see 0:21:09.200 --> 0:21:12.800 its connections and help you as a scientist acquire information fast. 0:21:13.080 --> 0:21:15.920 So that's a use of AI for the search. Then 0:21:15.920 --> 0:21:19.600 you go the next step generate a hypothesis. Well, to 0:21:19.680 --> 0:21:23.639 generate hypothesis, there's a beautiful new area in and I 0:21:23.800 --> 0:21:27.040 called generative models. We may be a little bit more 0:21:27.040 --> 0:21:30.040 familiar with the use of neural networks in SATURAI to 0:21:30.080 --> 0:21:34.159 do the task of classification. Right. If I give you images, 0:21:34.280 --> 0:21:36.280 you give me labels. Right, I said, well, this is 0:21:36.280 --> 0:21:38.560 a yellow car, a red car, and so on, and 0:21:38.600 --> 0:21:40.919 it gets done with an all network. Perhaps people are 0:21:40.960 --> 0:21:44.040 less familiar with using now some of these new networks 0:21:44.080 --> 0:21:47.680 to do generation in terms of classification. So I give 0:21:47.720 --> 0:21:50.040 you you know, I say, hey, designed to me a 0:21:50.160 --> 0:21:54.160 chair that looks like an avocado, right, and the system 0:21:54.280 --> 0:21:57.760 can automatically give you hundreds of thousands of different designs 0:21:57.760 --> 0:22:00.760 and so on. Right. So, so now you can use 0:22:00.920 --> 0:22:05.399 this generative capability to imagine new molecules back to connect 0:22:05.400 --> 0:22:08.119 it to our idea about chemistry and lithium chemistries that 0:22:08.320 --> 0:22:11.640 have these properties. I want give me molecules that may 0:22:11.680 --> 0:22:15.600 fit that criteria. And if I have an eye that 0:22:15.640 --> 0:22:18.600 creates these generative models, I want to verify whether they 0:22:18.640 --> 0:22:20.320 may work as they want. So now I can use 0:22:20.359 --> 0:22:22.600 a quantum computer right in the future to say do 0:22:22.640 --> 0:22:25.680 they work like they like? They say, because I'm simbulating 0:22:25.720 --> 0:22:28.240 a model of of chemistry is now and combining AI 0:22:28.440 --> 0:22:31.879 and quantum and simulation to be able to do this better. 0:22:32.240 --> 0:22:34.560 Then the next step says, well, let's realize it in practice. 0:22:34.640 --> 0:22:38.600 Let's do experimentation now, so I can have robots that 0:22:38.760 --> 0:22:42.720 synthesized with chemistry that are AI guided to optimally create 0:22:42.760 --> 0:22:46.240 the synthetic round, and the programming steps with which to 0:22:46.280 --> 0:22:49.879 create the molecules and so on. So I like to 0:22:49.920 --> 0:22:52.560 think about it is take a method that we know works, 0:22:52.800 --> 0:22:56.159 the scientific method, think about it as a method, and 0:22:56.200 --> 0:22:59.200 now ask yourself how the loop of technologies that we're 0:22:59.280 --> 0:23:03.000 creating can enhance it and improve it in concert with 0:23:03.040 --> 0:23:06.280 scientists and humans. And that is what I think is 0:23:06.280 --> 0:23:09.760 going to have revolutionary potential because I'm I'm closed with 0:23:09.800 --> 0:23:12.400 the idea of what a difference it made you brought 0:23:12.480 --> 0:23:14.840 up m r n A, What a difference it made 0:23:15.080 --> 0:23:18.000 to have the tools available to us to compress the 0:23:18.080 --> 0:23:20.760 time to discovery from the average time of fourteen years 0:23:20.760 --> 0:23:23.600 for a vaccine to under one. And think of the 0:23:23.600 --> 0:23:29.159 implications that happen well in future pandemics, in in climate change. 0:23:29.680 --> 0:23:32.359 How are we going to compress a time to discovery? 0:23:33.000 --> 0:23:34.680 And and that's going to be the power of the 0:23:34.720 --> 0:23:39.919 scientific methods supercharged with computing, including quantum and act mm HM. 0:23:40.720 --> 0:23:43.760 When I ask a question, it sounds like it is 0:23:43.760 --> 0:23:47.080 impossible to be a pessimist and work on quantum computing. 0:23:50.359 --> 0:23:54.199 I like that so so uh probably true, you know, 0:23:54.680 --> 0:23:58.840 because when you have so many challenges and uh and 0:23:58.960 --> 0:24:02.119 so many difficulty it takes a particular type of people 0:24:02.280 --> 0:24:05.760 to have the courage to be able to overcome them. 0:24:05.760 --> 0:24:09.480 But it gets combined when when you know that the 0:24:09.480 --> 0:24:14.679 theory is very sound and correct, the fact that we 0:24:14.760 --> 0:24:18.119 haven't been able to realize the technology that allows that 0:24:18.280 --> 0:24:22.840 theory to be expressed is in itself a source of energy, 0:24:23.040 --> 0:24:26.919 right And indeed, like you, you cannot be, you know, 0:24:27.160 --> 0:24:29.440 a pessimist if you want to be at the banguard 0:24:29.640 --> 0:24:33.120 of the creation of this technology. And also the implications 0:24:33.160 --> 0:24:35.360