WEBVTT - What's so exciting about quantum computing? (featuring Scott Aaronson) 0:00:06.440 --> 0:00:09.720 Hey, everyone, Daniel here. You know you hear the word 0:00:10.000 --> 0:00:13.960 quantum a lot, like a lot a lot, mostly in 0:00:14.080 --> 0:00:16.560 places where it makes no sense. I see it on 0:00:16.800 --> 0:00:21.160 advertisements for pest control companies or financial firms or laser 0:00:21.239 --> 0:00:25.360 tattoo joints. Okay, lasers actually are kind of quantum, so 0:00:25.440 --> 0:00:28.600 maybe that one works. But the point is that the 0:00:28.640 --> 0:00:32.360 word is so ubiquitous that it's come to mean very little. 0:00:32.800 --> 0:00:37.200 But it does have meaning. It evokes something modern and mysterious, 0:00:37.280 --> 0:00:40.800 because in the last one hundred years, physics has discovered 0:00:40.960 --> 0:00:44.920 that the world is very mysterious and operates under very 0:00:44.960 --> 0:00:47.720 different rules than the ones we are used to the 0:00:47.760 --> 0:00:51.159 world that we experience, the one that Aristotle and Copernicus 0:00:51.200 --> 0:00:54.240 and Galleo and Newton and even Einstein we're trying to 0:00:54.320 --> 0:00:58.320 understand is something of an illusion. When we peel back 0:00:58.440 --> 0:01:01.880 that layer of reality and see what's happening underneath, we 0:01:01.920 --> 0:01:06.200 discover that the rules of the microscopic universe are very different, 0:01:06.319 --> 0:01:10.119 almost alien to our intuition. And so of course people wondered, 0:01:10.160 --> 0:01:13.160 almost immediately, hey, what can I do with this? Can 0:01:13.200 --> 0:01:15.520 I take advantage of that to make my video games 0:01:15.560 --> 0:01:18.720 faster or hack into my school computer and change my 0:01:18.800 --> 0:01:22.120 grades because hey, what else is fundamental physics? 0:01:22.200 --> 0:01:22.959 Good form? All right? 0:01:23.440 --> 0:01:26.720 But jokes aside, quantum computing is something you hear a 0:01:26.760 --> 0:01:29.319 lot about these days. There's a lot of information and 0:01:29.400 --> 0:01:33.000 plenty of misinformation out there. Is it just another marketing 0:01:33.040 --> 0:01:36.600 ploy like quantum wasps appers or is it like the 0:01:36.640 --> 0:01:41.720 first computing revolution which completely transformed our society, our economy, 0:01:41.840 --> 0:01:44.960 and our daily lives. So today on the podcast, we're 0:01:45.000 --> 0:01:48.200 excited to be talking to Professor Scott Arenson, one of 0:01:48.200 --> 0:01:51.120 the world's top experts in quantum computing and a renowned 0:01:51.120 --> 0:01:54.480 communicator who can help us answer the question should we 0:01:54.520 --> 0:01:58.880 be excited about quantum computing? Welcome to Daniel and Kelly's 0:01:58.920 --> 0:02:00.000 Extraordinary Universe. 0:02:13.200 --> 0:02:17.040 Hello, I'm Kelly Waiter Smith, and I sort of think 0:02:17.080 --> 0:02:20.600 I understand how quantum computing works, but I'm never really sure. 0:02:21.040 --> 0:02:21.239 Hi. 0:02:21.320 --> 0:02:24.840 I'm Daniel. I'm a particle physicist, and I simultaneously understand 0:02:24.880 --> 0:02:26.800 and don't understand quantum computing. 0:02:27.919 --> 0:02:30.959 That's very appropriate. I think I think I understand it 0:02:31.080 --> 0:02:33.240 enough to get that joke, which makes me feel pretty good. 0:02:33.360 --> 0:02:34.320 Oh, you're pretty clever. 0:02:34.680 --> 0:02:38.360 So what is your favorite product that has been pitched 0:02:38.360 --> 0:02:40.800 as being quantum that makes you laugh. 0:02:42.919 --> 0:02:45.240 I once sat next to somebody on airplane who told 0:02:45.240 --> 0:02:49.040 me she was a quantum messuse. I was really wondering 0:02:49.120 --> 0:02:50.839 about how that worked, and I asked her a bunch 0:02:50.840 --> 0:02:54.000 of questions and I didn't reveal my expertise, but I 0:02:54.040 --> 0:02:56.960 didn't learn anything about quantum mechanics, and that conversation just 0:02:57.040 --> 0:02:57.920 something about. 0:02:57.639 --> 0:02:58.959 People, you know, yeah, yep. 0:02:59.320 --> 0:03:01.600 It was it like maybe she's giving you a massage, 0:03:01.600 --> 0:03:03.760 maybe she's not, and you don't know until you pay. 0:03:04.040 --> 0:03:06.640 That's the thing that reveals if it happened or not. 0:03:07.880 --> 0:03:10.160 No, it was more like spooky action at a distance, 0:03:10.280 --> 0:03:13.320 because she massages you without actually touching you. She like 0:03:13.560 --> 0:03:17.360 waves her hands nearby, and that somehow, through quantum mechanics 0:03:17.400 --> 0:03:21.200 influences your muscles and joints and whatever. So she was 0:03:21.280 --> 0:03:23.399 charging a lot of money. Also, people were paying her 0:03:23.560 --> 0:03:25.080 big bucks for this effect. 0:03:25.240 --> 0:03:27.520 I would be so grumpy if I paid for a 0:03:27.560 --> 0:03:30.120 massage like that, like you really got to get into 0:03:30.160 --> 0:03:33.880 my muscles if I'm paying you. But anyway, okay, well 0:03:33.880 --> 0:03:37.120 that's amazing. I once heard it was a quantum self 0:03:37.120 --> 0:03:40.440 help talk, and the whole time I just couldn't, could 0:03:40.480 --> 0:03:44.800 not understand how quantum, like how thinking about it from 0:03:44.800 --> 0:03:47.080 a quantum way was helpful at all, And it just 0:03:47.160 --> 0:03:48.560 left me completely baffled. 0:03:48.680 --> 0:03:51.040 I think people just latching onto the idea that the 0:03:51.080 --> 0:03:54.280 world is fundamentally different from the world we thought it was, 0:03:54.320 --> 0:03:57.240 that it works in a different way, it follows different rules, 0:03:57.600 --> 0:04:00.360 and that feels a little bit like magic. You can 0:04:00.400 --> 0:04:03.480 sort of grab onto that and smear your business with 0:04:03.560 --> 0:04:06.640 a little bit of that sparkly quantum physics stuff, And 0:04:06.680 --> 0:04:10.600 this feels like maybe you can do something which seemed impossible. 0:04:11.160 --> 0:04:14.720 When I think quantum is particularly likely to get used 0:04:14.720 --> 0:04:17.280 in that way because it's confusing. I mean, maybe it's 0:04:17.320 --> 0:04:18.760 not confusing to the people who study it, but for 0:04:18.880 --> 0:04:22.680 like lay people, it just sounds so unclear. And I 0:04:22.680 --> 0:04:26.359 think that you can feel like you understand it and yeah, 0:04:26.400 --> 0:04:28.720 but definitely feel like using the word quantum makes you 0:04:28.800 --> 0:04:31.040 sound smart and then you just run with it. And 0:04:31.120 --> 0:04:35.440 so I'm excited to see what our audience thinks quantum 0:04:35.520 --> 0:04:38.640 computing is all about, because I until I married Zach, 0:04:39.279 --> 0:04:42.160 who likes giving me literally three hour lectures on car 0:04:42.240 --> 0:04:45.520 rides about what quantum computing is. Well, I'm driving and 0:04:45.839 --> 0:04:49.600 can't escape. I didn't feel like I knew. So let's 0:04:49.600 --> 0:04:52.560 see what does the you know, what does the general public, 0:04:53.040 --> 0:04:56.080 in particular our audience think quantum computing is. 0:04:56.360 --> 0:04:58.320 So I reached out to our listeners and I asked 0:04:58.320 --> 0:05:03.240 them what's different about it quantum computer. Here's what listeners 0:05:03.440 --> 0:05:04.080 had to say. 0:05:04.880 --> 0:05:10.119 Quantum computers do multiple calculations at once, and each time 0:05:10.200 --> 0:05:17.360 you add a bit it goes up by exponential numbers faster. 0:05:17.440 --> 0:05:20.600 But especially they do not overheat. 0:05:22.279 --> 0:05:24.520 I don't know anything about quantum computers. 0:05:24.760 --> 0:05:27.400 You can actually get an answer faster because you can 0:05:27.600 --> 0:05:30.600 determine the probabilities rather than brute forcing it. 0:05:31.000 --> 0:05:35.840 Quantum computer works with superpositions instead of ones and zeros. 0:05:36.440 --> 0:05:41.400 That gives it advantages for some calculations, but makes it 0:05:41.839 --> 0:05:43.720 worse for other calculations. 0:05:43.960 --> 0:05:47.120 But quantum computer is like having five hundred and eighty 0:05:47.200 --> 0:05:52.040 thousand people all typing one word of the book, all 0:05:52.080 --> 0:05:54.000 at the same time, but in order. 0:05:54.440 --> 0:05:57.360 A quantum computer is different because you'll never show with 0:05:57.440 --> 0:05:59.599 it to turn it on and off again or off 0:05:59.680 --> 0:06:00.479 and on again. 0:06:00.839 --> 0:06:04.599 A quantum particle can be in. 0:06:04.680 --> 0:06:09.039 A superposition on off or on and off at the 0:06:09.040 --> 0:06:09.600 same time. 0:06:10.600 --> 0:06:14.479 A quantum computer can use three bits I think spin up, 0:06:14.560 --> 0:06:17.719 spin down, and I think something in between. 0:06:18.279 --> 0:06:24.120 A quantum computer runs on quantum mechanics, which makes it 0:06:24.680 --> 0:06:30.360 capable of calculating the probabilities in a situation where randomness 0:06:30.760 --> 0:06:32.560 is involved. 0:06:33.160 --> 0:06:38.640 Does a quantum computer not use binary code zeros and 0:06:38.760 --> 0:06:41.960 ones so there can be more options, which. 0:06:41.720 --> 0:06:42.960 Makes it more powerful. 0:06:43.080 --> 0:06:47.480 Possibly quantum computers use q bits that are super cooled 0:06:47.640 --> 0:06:52.400 and have many many states rather than the binary transistor 0:06:52.800 --> 0:06:54.960 logic that only has off and on. 0:06:56.000 --> 0:06:59.600 Quantum computing is all about probabilities. 0:07:00.080 --> 0:07:02.359 With a quantum computer, you can't tell if it's on 0:07:02.560 --> 0:07:04.599 or off until you hope the box it came in. 0:07:05.360 --> 0:07:10.840 A quantum computer uses quantum fluctuation for its processor and 0:07:11.000 --> 0:07:15.440 explores every possible answer instantaneously. 0:07:17.560 --> 0:07:20.480 How this is useful for our computation is still beyond 0:07:20.520 --> 0:07:21.320 my understanding. 0:07:21.960 --> 0:07:26.320 Computer uses quantum technology to search all the random variables 0:07:26.320 --> 0:07:28.920 that possibly happen and come up a handurd solution. 0:07:29.000 --> 0:07:34.160 I think quantum computers process really small numbers. Do they 0:07:34.200 --> 0:07:39.360 do computations on extremely tiny numbers? I'm not sure what 0:07:40.160 --> 0:07:42.400 is different about a quantum computer that. 0:07:42.520 --> 0:07:48.760 Makes use of superposition of states to give nearly infinite possibilities. 0:07:48.120 --> 0:07:51.160 The size of the processor or the flip flops. So 0:07:51.280 --> 0:07:55.800 now we're down on a quantum level versus the smallest 0:07:55.840 --> 0:07:57.440 transistors we have at this moment. 0:07:57.880 --> 0:08:00.640 Other them its fast are really about nothing. 0:08:01.440 --> 0:08:04.480 What's different about a quantum computer. I don't really know 0:08:04.640 --> 0:08:08.800 much about how quantum computers work, but there's something fundamentally 0:08:08.880 --> 0:08:12.080 different about how the bits, essentially we're being. 0:08:12.000 --> 0:08:12.520 On or off. 0:08:13.160 --> 0:08:15.920 And that's not something I know the details of. 0:08:16.520 --> 0:08:20.920 I'm not surprised that our audience had detailed and insightful 0:08:21.040 --> 0:08:22.280 responses to this question. 0:08:22.920 --> 0:08:24.440 That is awesome, But you know. 0:08:24.520 --> 0:08:27.880 You and I decided this is a complicated question, and well, 0:08:27.920 --> 0:08:29.880 I'm sure Daniel could have explained it on his own. 0:08:30.280 --> 0:08:32.240 We thought, you know, it might be a good idea 0:08:32.440 --> 0:08:35.880 to bring in Scott Aronson, who's an expert on quantum computing, 0:08:35.920 --> 0:08:37.440 has been working on it for a really long time, 0:08:37.720 --> 0:08:40.920 and so we're bringing you a conversation that's a collaboration 0:08:41.160 --> 0:08:45.600 between Daniel and Scott Aronson. So, Daniel, where do we start? 0:08:45.920 --> 0:08:48.800 So I think the place to start with understanding quantum 0:08:48.840 --> 0:08:52.760 computers is first with the word computer, Like what do 0:08:52.880 --> 0:08:56.040 we mean when we say computer? And this isn't just 0:08:56.200 --> 0:08:59.360 like philosophical rabbit hole. I think it's important to think 0:08:59.360 --> 0:09:02.040 about what a can is, what a computer does, not 0:09:02.240 --> 0:09:05.480 just the kinds of computers that we have, but other 0:09:05.679 --> 0:09:09.600 kinds of computers, what possible computers there are, because then 0:09:09.640 --> 0:09:11.960 we can compare the different kinds of computers, we can 0:09:12.080 --> 0:09:14.880 understand what they have in common and what they don't 0:09:14.920 --> 0:09:17.920 have in common. Because computers are not just something that 0:09:18.000 --> 0:09:20.520 sits on your desk or the thing inside your phone 0:09:20.559 --> 0:09:24.439 that makes it go. A computer basically is something that computes. 0:09:24.440 --> 0:09:27.000 It's something that gives you an answer to a question, 0:09:27.760 --> 0:09:30.000 you know, and that can be something very simple, like 0:09:30.240 --> 0:09:32.800 you can have a baseball and you throw the baseball. 0:09:33.080 --> 0:09:35.319 That gives you the answer to what happens when you 0:09:35.400 --> 0:09:37.480 throw a baseball, and that can be kind of a 0:09:37.559 --> 0:09:40.000 complicated calculation. You know, you have to take into account 0:09:40.040 --> 0:09:42.400 gravity and air resistance and the spin of the Earth 0:09:42.440 --> 0:09:44.320 and all this kind of stuff, and the baseball does 0:09:44.400 --> 0:09:47.120 that for you. It computes the answer to that question. 0:09:47.559 --> 0:09:50.439 And that's not a very exciting computer because that's basically 0:09:50.640 --> 0:09:53.440 all it can do other than let you play baseball. 0:09:53.960 --> 0:09:57.000 That one just answers one question. In general, we're interested 0:09:57.000 --> 0:09:59.880 in computers that can do lots of different kinds of calculations, 0:10:00.480 --> 0:10:04.040 and one kind of computer is you. You know, back 0:10:04.160 --> 0:10:06.679 in the middle of last century, what they called a 0:10:06.760 --> 0:10:10.000 computer was a person like at NASA who sat at 0:10:10.040 --> 0:10:14.480 a table doing calculations pencil and paper to figure out 0:10:14.920 --> 0:10:16.520 how are we going to shoot this rocket and what 0:10:16.640 --> 0:10:18.880 angle do we need to go at, because that was 0:10:18.920 --> 0:10:21.760 the best way to do those calculations. And humans pretty 0:10:21.800 --> 0:10:24.800 good at doing lots of different calculations. So a computer 0:10:24.960 --> 0:10:27.280 can include the thing on your desk, the thing in 0:10:27.360 --> 0:10:31.040 your phone, a baseball, even a person, right, and all 0:10:31.120 --> 0:10:33.880 these computers, some of them are good at different kinds 0:10:33.880 --> 0:10:37.320 of calculations, Like the ball is excellent at calculating where 0:10:37.400 --> 0:10:40.400 the ball is going to go, but terrible at everything else. 0:10:40.600 --> 0:10:43.160 A human is pretty good at some kinds of things 0:10:43.320 --> 0:10:46.079 and not that great at other kinds of things. Like 0:10:46.200 --> 0:10:49.079 a human is really good at doing things like spotting 0:10:49.160 --> 0:10:52.360 a tiger in the grass. Your brain is really efficient 0:10:52.800 --> 0:10:55.960 at noticing things that look like predators, not so great 0:10:55.960 --> 0:10:58.760 at other things like adding up all the numbers between 0:10:58.920 --> 0:11:00.960 one and a trillion. I mean, maybe you can find 0:11:01.000 --> 0:11:03.440 some mathematical shortcut, but if you have to actually add 0:11:03.559 --> 0:11:05.200 them up, it would take you a long time. It's 0:11:05.320 --> 0:11:08.240 not the kind of thing your brain is efficient at now. 0:11:08.280 --> 0:11:10.439 The kind of computer where, of course are interested in. 0:11:10.840 --> 0:11:14.480 It's not baseballs or tigers. It's the kind that are programmable, 0:11:14.559 --> 0:11:17.959 the kind that can basically do any kind of calculation. 0:11:18.360 --> 0:11:20.000 And this is something that's kind of incredible that we 0:11:20.080 --> 0:11:22.600 can even do. We have a question we want answer, 0:11:22.640 --> 0:11:25.839 it's some calculation we want done, and we build a 0:11:25.920 --> 0:11:29.360 computer which is like a machine, build a physical objects 0:11:29.720 --> 0:11:32.479 that we can manipulate in a way to do our calculation. 0:11:32.960 --> 0:11:36.719 That's the amazing thing about programmable computers that when you 0:11:36.800 --> 0:11:38.199 build it, you don't have to know what kind of 0:11:38.240 --> 0:11:40.160 calculations you want it to do. You can build it 0:11:40.240 --> 0:11:42.320 in such a way that it can do almost any 0:11:42.559 --> 0:11:46.959 kind of calculation. That's sort of amazing. We're taking a 0:11:47.000 --> 0:11:49.719 real world problem and then we're building a device that 0:11:49.800 --> 0:11:52.680 can calculate the answer as long as we can represent 0:11:52.840 --> 0:11:56.600 that real world problem somehow in the language of that computer. 0:11:57.280 --> 0:11:59.679 So these are deep issues and fascinating questions. To help 0:11:59.760 --> 0:12:03.640 us understand what is computing and how quantum computing is 0:12:03.760 --> 0:12:08.400 different from regular normal computing, we talk to Professor Scott Arenson. 0:12:08.760 --> 0:12:11.800 He he's a professor of theoretical computer science at the 0:12:11.880 --> 0:12:14.720 University of Texas at Austin. He's worked with open AI 0:12:15.000 --> 0:12:18.120 on the theoretical foundations of AI safety. He writes a 0:12:18.360 --> 0:12:21.880 wonderful blog about quantum computing called stettl Optimized, and he's 0:12:21.920 --> 0:12:25.719 maybe most famously the author of quantum computing since democratis. 0:12:25.880 --> 0:12:28.079 And he's a lot of fun to talk to. So 0:12:28.160 --> 0:12:30.319 we're very glad to have Scott on the show. And 0:12:30.360 --> 0:12:33.400 our first question for Scott, who's a professor of theoretical 0:12:33.480 --> 0:12:37.640 computer science, is what is theoretical computer science anyway? 0:12:38.040 --> 0:12:40.880 So it's a field that's usually considered to have started 0:12:41.080 --> 0:12:45.080 with Alan Touring in the nineteen thirties, who came up 0:12:45.120 --> 0:12:47.920 with a theoretical model of what a computer could be, 0:12:48.120 --> 0:12:49.960 which was called the touring machine. 0:12:50.240 --> 0:12:52.520 So when Scott is talking about a touring machine, he's 0:12:52.559 --> 0:12:55.559 not talking about any kind of computer anybody's ever actually built. 0:12:56.000 --> 0:12:59.600 This is like the computer science equivalent of a thought experiment, right, 0:13:00.080 --> 0:13:03.079 It's a thought computer. It's something Alan Turing came up with. 0:13:03.559 --> 0:13:06.959 It's the simplest kind of computer he could imagine, and 0:13:07.080 --> 0:13:09.640 it's a computer that, again you would never build. But 0:13:09.720 --> 0:13:13.200 it's very simple. Imagine a computer that has a tape 0:13:13.559 --> 0:13:16.720 on which there are written numbers zeros and ones, for example, 0:13:16.760 --> 0:13:18.679 and you can read the tape. It can move the 0:13:18.760 --> 0:13:21.520 tape forwards and backwards. They can also write to the tape, 0:13:21.840 --> 0:13:23.520 so you can do all the basic stuff that we 0:13:23.600 --> 0:13:26.960 think computers can do. Right. It can read in information, 0:13:27.160 --> 0:13:29.280 it can do a calculation and decide what it's going 0:13:29.360 --> 0:13:31.280 to do next. It can write onto the tape, so 0:13:31.360 --> 0:13:34.920 it can modify that. It can store information. And Turing 0:13:35.000 --> 0:13:37.959 did something really cool, which is that he showed that 0:13:38.120 --> 0:13:40.880 a Turing machine is very simple, basic kind of thought. 0:13:40.960 --> 0:13:44.319 Computer can do any kind of calculation that any computer 0:13:44.480 --> 0:13:46.800 could do. So if it was doable on a computer, 0:13:47.040 --> 0:13:49.200 you could do it on a Turing machine. That doesn't 0:13:49.240 --> 0:13:50.880 mean a Turing machine is the best way to do 0:13:50.920 --> 0:13:52.600 it or the right way to do it, but because 0:13:52.640 --> 0:13:55.200 the Turning machine is so simple, it let you think 0:13:55.280 --> 0:13:57.079 about the kind of things computers could do. 0:13:57.400 --> 0:13:57.559 Well. 0:13:57.600 --> 0:13:58.760 You could say, if you can do it on a 0:13:58.840 --> 0:14:00.840 Turing computer, then you could do it on any computer. 0:14:01.200 --> 0:14:03.840 And it's easier to prove theorems and think about stuff 0:14:04.040 --> 0:14:06.800 on a Touring computer. And again, a Turing machine doesn't 0:14:06.880 --> 0:14:08.920 lend me you to thinking about the kinds of computer 0:14:09.000 --> 0:14:11.439 that we've been building. It can explore the kind of 0:14:11.520 --> 0:14:12.959 things any computer can do. 0:14:13.600 --> 0:14:16.400 When Touring used the word computer in the nineteen thirties, 0:14:16.520 --> 0:14:18.839 you know, that was what it meant people, usually women, 0:14:19.160 --> 0:14:22.840 who were hired to do computation. And he came up 0:14:22.920 --> 0:14:25.720 with his model of the Touring machine by trying to 0:14:25.920 --> 0:14:29.040 idealize what such a person would be doing. You know 0:14:29.200 --> 0:14:31.800 that they'd be reading symbols on a piece of paper, 0:14:32.360 --> 0:14:35.720 maybe you know, crossing them off, replacing them with new symbols, 0:14:36.080 --> 0:14:39.000 you know, moving back and forth on the paper. They 0:14:39.040 --> 0:14:41.840 would always be doing so according to some rule that 0:14:41.920 --> 0:14:44.520 they knew or had been taught. There is nothing in 0:14:44.600 --> 0:14:47.280 the concept that is tied to you know, it has 0:14:47.360 --> 0:14:50.200 to be built out of transistors that are etched onto 0:14:50.280 --> 0:14:51.400 wafers of silicon. 0:14:51.920 --> 0:14:52.040 Right. 0:14:52.160 --> 0:14:54.480 That happens to be the way that we do it 0:14:54.640 --> 0:14:58.760 now because it worked so spectacularly. Well, okay, but you 0:14:58.840 --> 0:15:01.520 know a computer could in principle be made out of 0:15:01.600 --> 0:15:04.400 billiard balls, right, It could be made out of jets 0:15:04.440 --> 0:15:04.840 of water. 0:15:05.360 --> 0:15:05.520 Right. 0:15:05.600 --> 0:15:08.480 You know, these would maybe not be very reliable. 0:15:08.000 --> 0:15:12.320 Methods, and also doesn't have to be programmable and infinitely powerful. Right, 0:15:12.320 --> 0:15:15.560 anytime I build an experiment, I'm building a computer. 0:15:16.280 --> 0:15:20.800 Until very recently, you know, there was use in analog computers, right, 0:15:20.920 --> 0:15:25.600 and people just building analog systems to simulate some physical process. 0:15:25.960 --> 0:15:29.080 You know, at some point digital computing became so good 0:15:29.480 --> 0:15:31.840 that it eliminated the use cases for that. 0:15:32.880 --> 0:15:35.320 So it's great that the kind of digital computers we've 0:15:35.360 --> 0:15:38.120 been building can basically solve any problem. I mean, there 0:15:38.160 --> 0:15:40.680 are cobbyists to that, there are things turn computers can't do. 0:15:41.400 --> 0:15:44.800 But the crucial thing for understanding quantum computing and computing 0:15:44.880 --> 0:15:48.080 more generally is that some problems are hard, and some 0:15:48.280 --> 0:15:51.120 problems are easy. Some problems you can do quickly, and 0:15:51.200 --> 0:15:54.400 some problems you can do very very slowly. And different 0:15:54.520 --> 0:15:57.000 kind of computers are going to be good at different 0:15:57.120 --> 0:15:59.560 kinds of things, so the same with it. Like you 0:16:00.000 --> 0:16:02.560 as a person, you are a computer. You are fast 0:16:02.680 --> 0:16:06.280 or slow at particular problems. Our style of computers, what 0:16:06.360 --> 0:16:09.400 we call classical digital computers. These are good at some 0:16:09.600 --> 0:16:12.200 kinds of things and slow at other kinds of things. 0:16:12.720 --> 0:16:16.560 It's a feature of the kind of computers we've long developed, which, 0:16:16.680 --> 0:16:19.000 of course you know isn't the only way to do it. 0:16:19.160 --> 0:16:22.280 That's where we're going. But the origins of the digital 0:16:22.360 --> 0:16:24.760 computer that we've been using go all the way back 0:16:24.760 --> 0:16:27.400 to the middle of last century with John von Neuman. 0:16:27.720 --> 0:16:30.920 John von Neuman, who built some of the first digital computers, 0:16:31.120 --> 0:16:35.400 was you know, directly inspired by tourings insights, in particular 0:16:35.560 --> 0:16:38.200 the idea that you didn't want to build a different 0:16:38.320 --> 0:16:41.920 machine for each possible function. You wanted to build one 0:16:42.120 --> 0:16:46.400 universal machine and then have it simulate any other machine by. 0:16:46.360 --> 0:16:48.000 Giving it an appropriate program. 0:16:48.200 --> 0:16:50.880 This is, you know, an idea that today is so 0:16:51.080 --> 0:16:53.960 obvious that, you know, it's hard to convince my students 0:16:54.040 --> 0:16:56.920 that it was ever non obvious. So today what we 0:16:57.040 --> 0:17:00.520 do in theoretical computer science is often about trying to 0:17:00.680 --> 0:17:03.960 understand what problems can be solved efficiently. 0:17:04.119 --> 0:17:07.680 All right, And so when Scott talks about computers being efficient, 0:17:08.240 --> 0:17:11.439 he's thinking about whether they're good at this kind of problem? 0:17:11.600 --> 0:17:13.920 Can they do it quickly? As a problem gets bigger 0:17:13.960 --> 0:17:17.040 and bigger, do they become unbearably slow at solving it? 0:17:17.240 --> 0:17:19.480 Like you can add up the numbers between one and 0:17:19.600 --> 0:17:22.040 five pretty quickly, you can add up the number between 0:17:22.080 --> 0:17:25.040 one and one hundred, little less quickly. If it gets 0:17:25.080 --> 0:17:27.840 to one in a trillion, it's hopeless, right, And so 0:17:27.960 --> 0:17:30.399 for digital computers there are some problems that can be 0:17:30.520 --> 0:17:33.800 solved in principle but would take a very very long time, 0:17:33.880 --> 0:17:36.640 and the kind of computer we've been building. An example 0:17:36.720 --> 0:17:39.399 of that problem is checking to see if a number 0:17:39.680 --> 0:17:43.000 is prime. Like you can tell that eleven is prime 0:17:43.280 --> 0:17:45.840 because you can't think of any two numbers that multiply 0:17:46.000 --> 0:17:49.040 themselves together to give you eleven because it is prime. 0:17:49.440 --> 0:17:52.800 If I give you an arbitrary number seven, four hundred 0:17:52.840 --> 0:17:55.760 and seventeen, how do you know whether it's prime. Well, 0:17:55.800 --> 0:17:58.399 a computer can do this by, for example, checking all 0:17:58.440 --> 0:18:00.879 the numbers that go into it. That's just brute forcing it. 0:18:01.119 --> 0:18:03.440 There are some more clever algorithms we'll hear about later, 0:18:03.760 --> 0:18:06.480 but essentially it's very slow at checking prime numbers. So 0:18:06.720 --> 0:18:09.400 computers can do a lot of things, and the way 0:18:09.600 --> 0:18:12.440 we've usually built computers are good at some things and 0:18:12.640 --> 0:18:15.440 slow at other things. So we ask Scott, who thinks 0:18:15.440 --> 0:18:18.200 about this a lot, what kind of things our normal 0:18:18.280 --> 0:18:21.680 computers are good at and our normal computers are bad at? 0:18:22.440 --> 0:18:22.720 Sure? 0:18:23.200 --> 0:18:25.960 So simulating physics, you know, which you mentioned, is a 0:18:26.040 --> 0:18:29.520 great example of you know, something that computers have been 0:18:29.640 --> 0:18:32.760 used for since the very beginning, right and in some sense, 0:18:32.800 --> 0:18:36.520 you know, the entire program of physics since Galileo and 0:18:36.640 --> 0:18:40.200 Newton has been to you know, understand nature. Well. It 0:18:40.320 --> 0:18:43.520 off that we can put the initial conditions into a 0:18:43.600 --> 0:18:46.879 computer and just have a computer tell us what is 0:18:46.960 --> 0:18:47.919 going to happen. 0:18:48.080 --> 0:18:50.720 Simulating a differential equation, you know. 0:18:50.840 --> 0:18:56.600 In order to model fluids or gravitational dynamics, celestial mechanics. 0:18:57.040 --> 0:18:59.639 These are our famous examples of things that you know, 0:18:59.720 --> 0:19:00.600 comput uters. 0:19:00.320 --> 0:19:02.879 Are very good at. There are issues here, you know, 0:19:02.960 --> 0:19:04.560 that have to do with discritization. 0:19:05.320 --> 0:19:09.960 Nature is traditionally modeled in physics as using a continuum 0:19:10.040 --> 0:19:13.840 of numbers, and computers in the touring sense can only 0:19:13.920 --> 0:19:17.200 deal with discrete quantities with bits with ones and zeros. 0:19:17.280 --> 0:19:19.560 So we need some way to deal with that, right, 0:19:19.680 --> 0:19:23.680 Typically we take continuous quantities and we truncate them, we 0:19:23.840 --> 0:19:27.440 represent them only to a finite precision, and then you know, 0:19:27.560 --> 0:19:29.040 we have to understand how. 0:19:29.000 --> 0:19:31.480 Much error that's going to cause and so forth. 0:19:32.040 --> 0:19:35.119 But you know, in terms of what computers can do efficiently, 0:19:35.320 --> 0:19:37.600 you know, I think the classic examples that you know, 0:19:37.680 --> 0:19:40.760 we would teach in computer science are going to be 0:19:40.880 --> 0:19:44.360 things like, Okay, you know all of the arithmetic operations 0:19:44.440 --> 0:19:48.960 that we learn in school, right, adding to integers, you know, multiplying, 0:19:49.200 --> 0:19:52.879 dividing given Also, you know the thing that Google Maps 0:19:52.960 --> 0:19:55.960 does for us, right, find the shortest route between two 0:19:56.040 --> 0:20:00.280 given cities, you know, or between two addresses in the 0:20:00.359 --> 0:20:03.840 distance between each address and the immediately neighboring ones. 0:20:04.000 --> 0:20:06.240 Right, it's finding the shortest path in a graph. 0:20:06.600 --> 0:20:08.880 Okay. Now, there are a bunch of things that turn 0:20:09.000 --> 0:20:13.480 out to have clever efficient algorithms, even though it is 0:20:13.520 --> 0:20:17.000 sort of totally not obvious a priori that they would. 0:20:17.520 --> 0:20:21.560 For example, I give you a bunch of students, you know, 0:20:21.720 --> 0:20:25.520 I tell you who is willing to be roommates with whom? Okay, 0:20:25.680 --> 0:20:29.240 and now I ask you pair off as many willing 0:20:29.359 --> 0:20:32.680 roommates as you can. Right. This is called the maximum 0:20:32.800 --> 0:20:36.879 matching problem. Find you know, maximum number of pairs of 0:20:36.960 --> 0:20:40.600 people who are willing to room with each other A priori. 0:20:40.800 --> 0:20:44.479 That seems like that might require an exponential search, right, 0:20:44.560 --> 0:20:48.040 It might require considering an astronomical number of possibilities. 0:20:48.520 --> 0:20:51.359 But it was discovered in the nineteen sixties that it doesn't. 0:20:51.680 --> 0:20:54.320 There is an efficient way to solve that, That is, 0:20:54.640 --> 0:20:57.080 there's a way to solve it that scales only like 0:20:57.880 --> 0:21:02.399 the cube of the number of potential roommates something like that, 0:21:02.640 --> 0:21:03.840 rather than exponentially. 0:21:04.640 --> 0:21:07.000 Another great example would be linear. 0:21:06.760 --> 0:21:10.720 Programming, right, one of the most important problems in industrial 0:21:11.080 --> 0:21:14.679 you know, operations research, things like that, where you're given 0:21:14.760 --> 0:21:18.119 a bunch of linear constraints on some variables like this 0:21:18.359 --> 0:21:20.920 one plus this one can be at most ten, this 0:21:21.119 --> 0:21:23.800 one minus this one has to be at least eight, 0:21:23.960 --> 0:21:26.800 and so forth, and you're looking for a solution that 0:21:26.960 --> 0:21:30.840 satisfies all of those linear constraints. Okay, that also has 0:21:30.880 --> 0:21:35.760 an efficient solution primality. Right, I give you five thousand 0:21:35.800 --> 0:21:38.879 digit number, and I ask you is it prime or composite? 0:21:39.520 --> 0:21:39.639 Right? 0:21:39.760 --> 0:21:42.040 Well, you know you can try some simple things. 0:21:42.119 --> 0:21:44.639 You know, if it ends in an even number, or 0:21:44.720 --> 0:21:47.800 a zero or a five, right, then it's composite. But 0:21:48.320 --> 0:21:51.879 you know you can check if three, if seven, if eleven, 0:21:52.040 --> 0:21:55.280 go into it, right, But more generally, right, this is 0:21:55.480 --> 0:21:59.359 a famous problem in math. It's even extremely important in 0:21:59.480 --> 0:22:05.040 cryptoc Modern cryptography sort of uses gigantic prime numbers as 0:22:05.119 --> 0:22:07.080 one of its central ingredients. 0:22:07.200 --> 0:22:09.919 Okay, Now it turns out that there is. 0:22:10.160 --> 0:22:14.440 An efficient algorithm that tells you whether a huge number 0:22:14.640 --> 0:22:15.840 is prime or composite. 0:22:16.119 --> 0:22:18.440 Okay, it was discovered in the nineteen seventies. 0:22:18.920 --> 0:22:22.520 There are probabilistic methods, and in two thousand and two 0:22:22.760 --> 0:22:25.200 even a deterministic method was discovered. 0:22:25.480 --> 0:22:27.440 Okay, you're very very non obvious. 0:22:27.720 --> 0:22:31.560 Now, Crucially, these methods only tell you if the number 0:22:31.640 --> 0:22:34.920 is prime or composite. If it's composite, they don't tell 0:22:35.000 --> 0:22:36.680 you what the prime factors. 0:22:36.280 --> 0:22:39.080 Are, all right. So those are some great examples of 0:22:39.119 --> 0:22:41.720 what classical computers are pretty good at. What kind of 0:22:41.800 --> 0:22:44.560 things are they slow at? Again, a great example is 0:22:44.720 --> 0:22:48.679 prime factors. And this is really important because it turns 0:22:48.720 --> 0:22:51.560 out that it's useful to a lot of people that 0:22:51.640 --> 0:22:54.639 computers are slow at this. Like, if you know that 0:22:54.760 --> 0:22:58.240 computers can't crack this puzzle quickly, you can use it 0:22:58.400 --> 0:23:01.679 as a way to protect your information. The whole field 0:23:01.760 --> 0:23:06.440 of cryptography, of building codes and protecting information relies on 0:23:06.600 --> 0:23:10.200 some things being easy and some things being hard for 0:23:10.320 --> 0:23:11.080 computers to do. 0:23:11.680 --> 0:23:15.440 The belief that finding the prime factors is hard is 0:23:15.560 --> 0:23:21.119 actually also central to modern cryptography. So modern cryptography, you 0:23:21.240 --> 0:23:24.760 have to be able to generate huge prime numbers quickly 0:23:25.240 --> 0:23:28.440 multiply them together quickly, which we know how to do 0:23:28.560 --> 0:23:32.119 all of that. That's how we generate the cryptographic keys 0:23:32.240 --> 0:23:35.080 called the public keys, that people can use to send 0:23:35.200 --> 0:23:38.080 us encrypted messages. But then the way that it works 0:23:38.320 --> 0:23:41.399 is that in order to decrypt the message, we. 0:23:41.560 --> 0:23:43.560 Think, you need to know the prime factors. 0:23:44.000 --> 0:23:46.720 If anyone had a fast way to find the prime 0:23:46.840 --> 0:23:51.200 factors of a gigantic composite number, most of the cryptography 0:23:51.280 --> 0:23:53.000 that protects the Internet would be broken. 0:23:53.200 --> 0:23:54.800 Okay, so it is crucial that. 0:23:54.880 --> 0:23:58.120 You know, after half a century of effort, at least 0:23:58.160 --> 0:24:02.040 the best publicly known methods for factoring numbers. 0:24:02.160 --> 0:24:02.320 You know. 0:24:02.840 --> 0:24:05.440 Of course, if the NSA you know knew something better, 0:24:05.600 --> 0:24:07.960 then you know, I would have no reason to know that. 0:24:08.320 --> 0:24:11.520 But the best publicly known methods you know use an 0:24:11.520 --> 0:24:15.560 amount of time that scales exponentially with the number of digits, 0:24:15.720 --> 0:24:19.040 or more precisely, exponentially with the cube root of the 0:24:19.160 --> 0:24:20.040 number of digits. 0:24:20.920 --> 0:24:24.400 So when Scott talks about scaling with the number of digits, 0:24:24.600 --> 0:24:26.920 what he's talking about is how long it will take 0:24:27.000 --> 0:24:30.520 the computer to solve the puzzle, depending on the length 0:24:30.960 --> 0:24:34.560 of the password. Problems that get much harder very quickly 0:24:34.640 --> 0:24:37.280 when you add digits to your password. Those are good 0:24:37.400 --> 0:24:40.040 for cryptography because it makes it easy to make the 0:24:40.119 --> 0:24:44.440 problem impossible even if computers get faster. Like if computers 0:24:44.600 --> 0:24:47.520 suddenly tomorrow get twice as faster next year they're five 0:24:47.600 --> 0:24:49.880 times as fast. You don't want people to be able 0:24:49.920 --> 0:24:52.280 to crack your passwords. The good thing about the prime 0:24:52.359 --> 0:24:54.560 number puzzle is you can just add a couple more 0:24:54.640 --> 0:24:57.119 digits to our keys to our passwords, and now the 0:24:57.200 --> 0:25:00.840 puzzle is impossible. Again, this problem is easy to make 0:25:01.160 --> 0:25:05.439 much harder because it gets exponentially harder as it gets bigger. 0:25:05.600 --> 0:25:08.639 So it's easy to keep making the problem harder faster 0:25:08.800 --> 0:25:11.920 than computers are getting better at the problem. That's the 0:25:12.040 --> 0:25:14.680 key to these exponential problems. And there are a bunch 0:25:14.720 --> 0:25:16.960 of problems like this that are very hard to solve 0:25:17.359 --> 0:25:19.560 for our classical digital computers. 0:25:19.960 --> 0:25:23.359 So factoring is a famous example of a problem that 0:25:23.520 --> 0:25:27.480 might be exponentially hard for all we know. Okay, but 0:25:27.800 --> 0:25:31.879 there's maybe an even more famous class of problems that 0:25:31.960 --> 0:25:36.000 are believed to be exponentially hard. And this includes, like 0:25:36.280 --> 0:25:41.359 most of the problems in combinatorial search and optimization that 0:25:41.480 --> 0:25:44.600 people care about in practice, and so examples would be, 0:25:45.200 --> 0:25:49.000 I give you the distances between every city and every other. 0:25:49.440 --> 0:25:52.159 I ask you to find the shortest path that visits 0:25:52.240 --> 0:25:57.800 every city. That's the famous traveling salesman problem traveling salesperson. 0:25:57.480 --> 0:25:58.080 Call it today. 0:25:58.600 --> 0:26:01.680 Or I give you the mess of a bunch of suitcases. 0:26:02.160 --> 0:26:04.200 I asked, can they all fit in the trunk of 0:26:04.240 --> 0:26:06.560 your car? That's a problem with which many of us 0:26:06.640 --> 0:26:07.359 have experience. 0:26:07.720 --> 0:26:10.120 The answer is always yes, you have to try. 0:26:10.080 --> 0:26:12.920 Arranging them in a different way. Or you know, I 0:26:13.000 --> 0:26:15.480 give you a jigsaw puzzle, you know, can you solve it? 0:26:15.640 --> 0:26:17.960 To make it hard, let's imagine a jigsaw puzzle with 0:26:18.080 --> 0:26:21.199 no picture on it, okay, or a sudoku. 0:26:21.600 --> 0:26:23.639 The answer is always you can solve. It's just a 0:26:23.720 --> 0:26:27.199 question of how many curse words are going to be the. 0:26:27.200 --> 0:26:29.919 Answer yes, yes, And if it's thousands of pieces, then 0:26:29.960 --> 0:26:32.679 are we talking about more curse words than there are 0:26:32.800 --> 0:26:34.640 atoms in the observable universe. 0:26:52.200 --> 0:26:54.399 So we've reminded ourselves what it digital computer can do. 0:26:54.680 --> 0:26:57.040 Some things it can do really well and very quickly, 0:26:57.440 --> 0:26:59.520 and other things it does more slowly, and as a 0:26:59.560 --> 0:27:02.760 problem to bigger, becomes unbearably slow. But the topic of 0:27:02.840 --> 0:27:06.080 today's episode, of course, is other kinds of computers. What 0:27:06.240 --> 0:27:09.120 if you decided to build a computer using something other 0:27:09.280 --> 0:27:13.160 than zeros and ones. Remember, computer is just some arrangement 0:27:13.320 --> 0:27:16.000 of a physical system that lets you do a calculation. 0:27:16.560 --> 0:27:19.280 Doesn't have to be based on like bits that flip 0:27:19.359 --> 0:27:22.600 between zero and one and half precise values. What if 0:27:22.640 --> 0:27:25.720 you found something in the universe that operated differently, that 0:27:25.880 --> 0:27:29.360 wasn't so crisp, right, that operated under a different set 0:27:29.400 --> 0:27:32.760 of rules that you could then exploit to do different 0:27:32.840 --> 0:27:35.840 kinds of calculations, or to have some calculations be faster 0:27:36.040 --> 0:27:38.560 or slower. That would be awesome because it would complement 0:27:38.680 --> 0:27:39.879 our current computer system. 0:27:40.000 --> 0:27:40.119 Right. 0:27:40.280 --> 0:27:43.080 We already know that this is possible in principle. You 0:27:43.160 --> 0:27:45.600 know my example of a baseball that can do a 0:27:45.800 --> 0:27:49.160 very precise calculation that includes all sorts of effect wind 0:27:49.240 --> 0:27:51.480 resistance and the tug of Jupiter and all sorts of 0:27:51.520 --> 0:27:54.800 stuff that would be very laborious for a traditional computer 0:27:54.920 --> 0:27:57.159 to do. It can do it very very quickly, in 0:27:57.280 --> 0:27:59.920 like the time you throw a baseball. But the question, 0:28:00.080 --> 0:28:02.800 and of course, is can you make a programmable computer 0:28:03.320 --> 0:28:05.960 not a specialized one off thing like a baseball, A 0:28:06.080 --> 0:28:09.720 programmable computer that is good at the kinds of problems 0:28:10.080 --> 0:28:14.000 that classical computers are slow at. And that's where quantum 0:28:14.000 --> 0:28:18.200 mechanics comes in, because quantum mechanics does operate under different rules, 0:28:18.240 --> 0:28:21.040 and we can exploit those rules to build a different 0:28:21.160 --> 0:28:24.480 kind of computer. The crucial things to understand about quantum 0:28:24.480 --> 0:28:27.240 mechanics in just a few minutes are that rather than 0:28:27.320 --> 0:28:30.880 having to have definitive states like this bit is zero 0:28:31.080 --> 0:28:33.760 or this bit is one, quantum bits from what we 0:28:33.840 --> 0:28:36.720 call cubits, can be in a superposition of states. A 0:28:36.800 --> 0:28:39.440 superposition just means it has a chance to be in 0:28:39.560 --> 0:28:42.160 more than one state, So rather than being a zero 0:28:42.520 --> 0:28:44.400 or a one, it can be like, well, this one 0:28:44.440 --> 0:28:46.320 has a thirty percent chance of being a zero and 0:28:46.400 --> 0:28:48.719 a seventy percent chance of being a one. That one 0:28:48.800 --> 0:28:51.200 over there has a ninety percent chance of being a 0:28:51.320 --> 0:28:53.240 zero and a ten percent chance of being a one. 0:28:53.800 --> 0:28:56.440 So a superposition just means two things laid on top 0:28:56.480 --> 0:28:59.280 of each other doesn't have to be zero or one. 0:28:59.640 --> 0:29:03.080 It can be some probability of zero or one. And 0:29:03.200 --> 0:29:06.040 the fact that it can maintain these superpositions lets it 0:29:06.160 --> 0:29:10.040 do something that classical bits can't do, which is interfere 0:29:10.480 --> 0:29:13.040 If you've heard of the double slit experiment, this is 0:29:13.160 --> 0:29:16.000 like a beam of photons that go through two slits, 0:29:16.320 --> 0:29:18.680 and maybe the photons come from one slit, and maybe 0:29:18.680 --> 0:29:21.800 the photons go through another slit, and the possibility that 0:29:21.880 --> 0:29:25.320 they've gone through both slits interferes with each other, creating 0:29:25.360 --> 0:29:28.320 this interference pattern on the final screen. And it's the 0:29:28.440 --> 0:29:31.080 fact that the photons can be in a superposition of 0:29:31.120 --> 0:29:33.880 those states, that they can be maybe through slit one 0:29:33.960 --> 0:29:36.960 and maybe through slit two. Those possibilities are the things 0:29:37.080 --> 0:29:40.960 doing the interfering. So that interference is very important part 0:29:41.280 --> 0:29:44.520 of what quantum systems can do and what classical systems 0:29:44.600 --> 0:29:46.760 cannot do because the classical system like if you throw 0:29:46.760 --> 0:29:49.560 a base ball through a double slit experiment, it either 0:29:49.600 --> 0:29:51.480 went through the left one or through the right one. 0:29:51.640 --> 0:29:55.000 There's no superposition and there's no interference. Now these are 0:29:55.080 --> 0:29:57.560 quantum systems that can do this weird thing of having 0:29:57.720 --> 0:30:01.440 multiple possibilities at the same time. But then when quantum 0:30:01.480 --> 0:30:05.360 systems interact with classical systems, when you ask the quantum system, hey, 0:30:05.880 --> 0:30:08.240 what is the value of this bit? Because eventually you 0:30:08.360 --> 0:30:10.280 want to know the answer from your computer, right you 0:30:10.360 --> 0:30:12.720 have to make a measurement, you have to do something, 0:30:12.760 --> 0:30:15.760 you interact with it. That's when the universe picks one 0:30:15.840 --> 0:30:18.400 of the options. So there's a spread of possible outcomes 0:30:18.440 --> 0:30:21.920 for any quantum interaction. There's a superposition that describes the 0:30:22.000 --> 0:30:25.400 various possibilities, and measurement is the thing that collapses it 0:30:25.680 --> 0:30:29.000 that makes the universe pick one of those outcomes. So 0:30:29.120 --> 0:30:31.560 this is the way the universe works in the microscopic scale, 0:30:31.560 --> 0:30:34.160 which is weird and amazing and very different from our 0:30:34.280 --> 0:30:37.880 experience and the macroscopic scale, and it opens the door 0:30:37.960 --> 0:30:41.920 to doing different kinds of computation. Remember a computer, it's 0:30:42.000 --> 0:30:45.200 just a physical system we've arranged to calculate something we're 0:30:45.240 --> 0:30:47.840 interested in. We take advantage of the way that physical 0:30:47.880 --> 0:30:50.800 system works to represent some kind of calculation and hope 0:30:50.800 --> 0:30:52.800 that that computer that we've built is good at that 0:30:52.960 --> 0:30:56.520 kind of calculation. And because quantum computers work with these 0:30:56.640 --> 0:30:59.280 very different rules, it means they can do different kinds 0:30:59.320 --> 0:31:02.200 of calculation, and they can do different calculations quickly, and 0:31:02.240 --> 0:31:05.840 they have different strengths and weaknesses than classical computers. So 0:31:05.960 --> 0:31:07.440 here's Scott telling us more about that. 0:31:08.080 --> 0:31:12.240 Let's now imagine a computer that operates by these principles 0:31:12.240 --> 0:31:15.560 of quantum mechanics, we'll call it a quantum computer, so. 0:31:15.840 --> 0:31:16.920 A classical computer. 0:31:17.360 --> 0:31:21.120 Typically, for simplicity, we like to imagine it as having 0:31:21.200 --> 0:31:23.840 a state that's built up out of bits out of 0:31:23.880 --> 0:31:26.760 you know, zeros are ones, right, and if there's one 0:31:26.800 --> 0:31:30.360 thousand bits, then there's two to one thousand power possible 0:31:30.440 --> 0:31:34.880 configurations of that computer's memory, okay. But now with a 0:31:34.960 --> 0:31:39.320 quantum computer, there's going to be vastly more configurations than that, okay, 0:31:39.360 --> 0:31:43.040 because if I have even one quantum bit or you know, 0:31:43.120 --> 0:31:46.360 what we call a cube bit, right, then it can 0:31:46.440 --> 0:31:49.920 have some amplitude for being zero and some amplitude for 0:31:50.040 --> 0:31:53.160 being one at the same time. So it can be 0:31:53.280 --> 0:31:56.280 in what we call a superposition of the zero state 0:31:56.400 --> 0:31:57.920 and the one state, you know, at. 0:31:57.880 --> 0:31:59.560 Least before we look at it. 0:32:00.160 --> 0:32:02.560 Once we make a measurement, then we'll force it to 0:32:02.720 --> 0:32:05.040 snap to either zero or one, and then it will 0:32:05.560 --> 0:32:07.800 probabilistically collapse to one or the other. 0:32:08.120 --> 0:32:10.960 But before we look it can be in this superposition state. 0:32:11.560 --> 0:32:13.960 And now if I have let's. 0:32:13.680 --> 0:32:17.400 Say three cubits, okay, then it's not enough to give 0:32:17.480 --> 0:32:20.920 amplitudes for each cubit separately from the others. Right, The 0:32:21.040 --> 0:32:24.600 rules of quantum mechanics are unequivocal. I have to give 0:32:24.680 --> 0:32:27.720 an amplitude that the three bits are zero zero zero. 0:32:28.240 --> 0:32:30.160 I have to give an amplitude that there are zero 0:32:30.320 --> 0:32:32.440 zero one. I have to give an amplitude that they're 0:32:32.520 --> 0:32:35.040 zero one zero, you know, and so on, so I 0:32:35.120 --> 0:32:36.640 have to give eight amplitudes. 0:32:37.200 --> 0:32:40.400 So it's sort of more information dense because the amount 0:32:40.400 --> 0:32:43.560 of information grows exponentially. But why does that allow us 0:32:43.640 --> 0:32:46.360 to do different kinds of computation? Why does that allow 0:32:46.400 --> 0:32:50.160 us to solve different kinds of problems efficiently than classical computers. 0:32:50.360 --> 0:32:52.240 Yes, I mean, of course that's the point, and that's 0:32:52.280 --> 0:32:54.680 where this is ultimately headed. But we have to be 0:32:54.840 --> 0:32:58.000 very careful about it, because if you take these three 0:32:58.120 --> 0:32:59.600 cubits and you measure. 0:32:59.320 --> 0:33:01.560 Them, don't see those eight numbers. 0:33:01.840 --> 0:33:06.560 Now, the cubits collapse and you just see three bits each, 0:33:06.720 --> 0:33:09.880 you know, with some probability. Right now, if I have 0:33:10.280 --> 0:33:13.880 one thousand cubits, right, then that's two to the thousand 0:33:14.000 --> 0:33:17.600 power amplitudes to keep track of their state, right, which is, 0:33:17.680 --> 0:33:19.240 you know, more than you could write down in the 0:33:19.320 --> 0:33:22.280 whole observable universe. Okay, so there seems to be that 0:33:22.440 --> 0:33:26.600 this exponentiality, you know, beneath the surface. And this is 0:33:26.720 --> 0:33:30.760 certainly a problem if you wanted to simulate quantum mechanics 0:33:30.920 --> 0:33:35.000 on a conventional computer. Okay. And actually chemists and physicists 0:33:35.040 --> 0:33:38.160 have known that for generations, right. They're like, once they 0:33:38.240 --> 0:33:41.880 started trying to apply the Schrodinger equation to you know, 0:33:42.040 --> 0:33:46.120 calculate the behavior of even quite simple molecules, right, they 0:33:46.200 --> 0:33:49.200 have to write down a wave function that has like 0:33:49.680 --> 0:33:53.640 more and more dimensions, you know, the more electrons you add, right, 0:33:53.800 --> 0:33:57.000 and this you very quickly get the problems that would 0:33:57.080 --> 0:34:00.480 tax you know, the fastest supercomputers of today. You know, 0:34:00.600 --> 0:34:04.200 let alone of the nineteen fifties when they started doing this, right, 0:34:04.760 --> 0:34:07.880 And so you know, chemists and physicists invented all sorts 0:34:07.960 --> 0:34:13.080 of hacks and approximation methods for dealing with these exponentially 0:34:13.280 --> 0:34:16.920 large wave functions. Okay, But it was not until the 0:34:17.080 --> 0:34:21.840 early eighties that a few physicists like Feineman and Deutsch, 0:34:22.239 --> 0:34:25.200 you know, started saying, well, if nature is giving us 0:34:25.320 --> 0:34:28.320 this computational lemon, you know, it is making it so 0:34:28.480 --> 0:34:31.920 hard for us to simulate atomic physics on computers. 0:34:32.160 --> 0:34:34.520 Then why don't we make lemonade? That is, you know, 0:34:34.600 --> 0:34:35.680 why don't we build. 0:34:35.440 --> 0:34:40.520 A computer that itself would exploit quantum mechanical principles, you 0:34:40.600 --> 0:34:42.240 know what they called a quantum computer. 0:34:42.719 --> 0:34:44.520 And then what would that be good for? 0:34:44.960 --> 0:34:47.560 Well, if nothing else, it would be good for simulating 0:34:47.640 --> 0:34:50.680 quantum mechanics itself, right, And that was sort of the 0:34:50.800 --> 0:34:55.080 original application that they had in mind, And forty years later, 0:34:55.239 --> 0:34:58.080 I think that that's still, honestly, you know, the most 0:34:58.120 --> 0:35:01.680 important application economic that we know. A bet that's the 0:35:01.760 --> 0:35:02.960 truth of the matter, right. 0:35:03.360 --> 0:35:04.239 Anyone who is. 0:35:04.320 --> 0:35:08.680 Trying to design better batteries or better solar cells, or 0:35:09.040 --> 0:35:14.520 high temperature superconductors, or better chemical reactions for making fertilizer 0:35:15.040 --> 0:35:18.719 or better drugs that you know buind a receptor in 0:35:18.800 --> 0:35:21.600 a certain way, right, they're basically dealing with a many 0:35:21.680 --> 0:35:25.480 body quantum mechanics problem. And these problems can be incredibly 0:35:25.600 --> 0:35:31.279 hard for classical computers for reasons that you know, ultimately 0:35:31.719 --> 0:35:35.400 come from the exponentiality of the wave function, right, and 0:35:35.680 --> 0:35:38.960 a quantum computer could potentially help with any of that. 0:35:39.920 --> 0:35:43.560 Scott is pointing out a really crucial feature of quantum systems. 0:35:44.000 --> 0:35:46.080 Not only do they have these cubits that can be 0:35:46.200 --> 0:35:49.279 in superposition, but the cubits can be entangled with each other, 0:35:49.320 --> 0:35:52.040 which means the value in one bit can be linked 0:35:52.120 --> 0:35:54.680 to the value in another bit, and that's what makes 0:35:54.719 --> 0:35:58.080 them much more information dense than classical computers. It's like, 0:35:58.160 --> 0:36:01.160 instead of having three independent axes where you can just 0:36:01.239 --> 0:36:03.400 pick a number along the access, you have three pieces 0:36:03.440 --> 0:36:06.800 of information. You semble those into a three dimensional space, 0:36:07.239 --> 0:36:09.600 so now you have a three D volume of information. 0:36:09.960 --> 0:36:13.400 So they're capable of storing information much more densely because 0:36:13.400 --> 0:36:17.280 of these connections between the bits, which creates this information space, 0:36:18.040 --> 0:36:21.279 and this makes it very hard for classical computers to 0:36:21.480 --> 0:36:25.120 simulate a quantum system. It takes a lot of normal bits, 0:36:25.200 --> 0:36:29.120 classical zero one bits to calculate what a quantum system 0:36:29.160 --> 0:36:31.840 will do or what a cbe bit will do. So 0:36:32.000 --> 0:36:34.880 the first thing that quantum computers could be good for 0:36:35.200 --> 0:36:37.239 is to just describe quantum systems. It's kind of a 0:36:37.320 --> 0:36:40.959 natural application. You know, the quantum computer follows similar rules 0:36:41.040 --> 0:36:43.440 to the quantum system, and so it's natural to describe 0:36:43.480 --> 0:36:45.560 it in that way, the same way like the baseball 0:36:45.640 --> 0:36:47.560 follows the rules of the baseball. So it's a good 0:36:47.600 --> 0:36:50.480 way to calculate what a baseball will do. But of 0:36:50.560 --> 0:36:53.319 course we wonder, like, is that all quantum computers can 0:36:53.400 --> 0:36:58.320 do simulate some nerdy quantum experiment or are quantum computers 0:36:58.360 --> 0:37:02.440 also good at doing other things? Here's Scott telling us 0:37:02.440 --> 0:37:02.920 all about it. 0:37:03.400 --> 0:37:04.840 That was the original promise. 0:37:05.320 --> 0:37:07.360 But you know, as long as that was sort of 0:37:07.880 --> 0:37:10.560 the only promise, I think, you know, this remained very 0:37:10.680 --> 0:37:16.440 much a niche interest of some weird physicists pursuing this idea. 0:37:16.560 --> 0:37:18.840 And you know, the eighties, the early nineties. 0:37:19.280 --> 0:37:22.759 Now, the big discovery that put quantum computing on the 0:37:22.880 --> 0:37:25.080 map for you know, most of the rest of the 0:37:25.160 --> 0:37:29.239 world was that a quantum computer can sometimes also help 0:37:30.400 --> 0:37:33.880 to get exponential speed ups, even for problems that have 0:37:34.040 --> 0:37:37.160 nothing to do with quantum mechanics, at least for a 0:37:37.280 --> 0:37:39.560 few very specific such problems. 0:37:40.120 --> 0:37:42.640 Can we predict these kinds of problems in advance? 0:37:43.000 --> 0:37:45.239 Welcome to what my colleagues and I have been trying 0:37:45.280 --> 0:37:46.680 to do for the last thirty years. 0:37:46.880 --> 0:37:48.960 Yeah, I mean, for the whole history of this field. 0:37:49.080 --> 0:37:51.520 We are trying to figure out what is the border 0:37:51.680 --> 0:37:55.200 between what is efficiently solvable by a quantum computer and 0:37:55.280 --> 0:37:57.560 what isn't and we know a lot about it, but 0:37:57.719 --> 0:37:58.319 you know, there is. 0:37:58.320 --> 0:37:59.920 A great deal that we still don't know. 0:38:00.560 --> 0:38:04.759 The big discovery that sort of started quantum computing as 0:38:04.800 --> 0:38:07.279 a field, I would say, you know, as opposed to 0:38:07.400 --> 0:38:10.600 just an idea, came in nineteen ninety four, Okay, and 0:38:10.680 --> 0:38:14.040 that was when Peter Shore, who was a mathematician then 0:38:14.120 --> 0:38:18.400 at Belle Ebs, discovered that there is a fast quantum 0:38:18.520 --> 0:38:20.560 algorithm for factoring numbers. 0:38:21.120 --> 0:38:23.600 Okay, So he discovered that the factoring. 0:38:23.200 --> 0:38:28.120 Problem, the problem of factoring a huge composite number into primes, 0:38:28.719 --> 0:38:33.640 and some various closely related problems of central importance in 0:38:33.800 --> 0:38:39.680 modern cryptography are all solvable on a quantum computer using 0:38:39.760 --> 0:38:43.279 a number of steps that grows like the size of 0:38:43.360 --> 0:38:47.040 the number. You know that you're trying to factor squared maybe, 0:38:47.400 --> 0:38:49.480 but not exponentially with the size. 0:38:49.280 --> 0:38:49.800 Of the number. 0:38:50.200 --> 0:38:52.680 So this is a big deal because, as you're saying, 0:38:53.040 --> 0:38:55.080 you know, we can use billiard balls to calculate how 0:38:55.120 --> 0:38:57.440 billiard balls move, and we can use quantum systems to 0:38:57.480 --> 0:39:00.600 simulate quantum systems. But now we're using a quotum system 0:39:00.840 --> 0:39:03.279 to describe something that's fundamentally not quantum. So it gives 0:39:03.320 --> 0:39:05.319 us a clue that like, maybe we can open up 0:39:05.320 --> 0:39:06.800 a whole new category of problems. 0:39:07.000 --> 0:39:10.400 It is totally not obvious a priori that a quantum 0:39:10.440 --> 0:39:12.800 computer should help you for factoring numbers. 0:39:13.160 --> 0:39:15.480 You know, what does that have to do with quantum mechanics? 0:39:15.960 --> 0:39:20.120 Right? And of course this problem is hugely important because, 0:39:20.280 --> 0:39:22.920 for better or worse, we base the whole security of 0:39:23.040 --> 0:39:27.439 the modern Internet on the belief that factoring is hard. Okay, 0:39:27.560 --> 0:39:30.000 What Sure was saying is that if and when someone 0:39:30.120 --> 0:39:34.759 builds a large quantum computer, a scalable quantum computer with 0:39:34.920 --> 0:39:38.320 you know, thousands or millions of cubits, then that is 0:39:38.400 --> 0:39:41.440 no longer true. Okay, then you can break all of 0:39:41.520 --> 0:39:44.160 the encryption that we use to protect the Internet. 0:39:44.640 --> 0:39:47.160 So a bunch of things happened, you know after that. 0:39:47.360 --> 0:39:49.719 You know, one was people you know kind of like 0:39:49.880 --> 0:39:52.960 with the story of Rumpelstiltskin, Right, It's like, if you 0:39:53.080 --> 0:39:56.080 can spin this much straw into gold, and then why 0:39:56.160 --> 0:39:59.960 not more? And people said, well, maybe all of the exponential, 0:40:00.000 --> 0:40:03.120 really hard problems that we're you know, dealing with, maybe 0:40:03.200 --> 0:40:05.120 quantum computers can solve all of them. 0:40:05.320 --> 0:40:08.600 Is there any way to intuitively understand the idea here? 0:40:08.760 --> 0:40:11.719 Like what it is about quantum computing that makes this 0:40:12.120 --> 0:40:13.320 problem easier to do. 0:40:13.800 --> 0:40:16.239 I teach a whole undergrad course where you know, at 0:40:16.320 --> 0:40:18.040 the end of it, I hope that people will have 0:40:18.200 --> 0:40:19.480 the intuition for these things. 0:40:19.560 --> 0:40:19.640 Right. 0:40:19.680 --> 0:40:21.680 But like, if there were a one sentence way to 0:40:21.760 --> 0:40:24.400 say the intuition, then you wouldn't have needed Shore and 0:40:24.480 --> 0:40:27.600 Grover to discover these things, right, you know, it could 0:40:27.600 --> 0:40:29.080 have been obvious from the beginning. 0:40:29.280 --> 0:40:32.319 Right. But let me say this, right, so, you know. 0:40:32.400 --> 0:40:36.320 What almost every popular article about quantum computing wants to 0:40:36.440 --> 0:40:41.520 say is something that sounds really appealing and is totally wrong. Okay. 0:40:42.080 --> 0:40:44.680 In fact, Zach Kelly's husband and I made a whole 0:40:44.800 --> 0:40:47.279 cartoon about exactly this eight years ago. 0:40:47.520 --> 0:40:49.839 So throw cold water on some clickbait for us. What's 0:40:49.920 --> 0:40:52.440 wrong about quantum computing descriptions? 0:40:52.840 --> 0:40:56.160 What almost every popular writer has wanted to say is 0:40:56.239 --> 0:41:01.000 that a quantum computer just tries every possible solution in parallel. 0:41:01.200 --> 0:41:04.600 You know, it tries each one in a different parallel universe, 0:41:04.800 --> 0:41:07.200 or you know, all of them in superposition or whatever, 0:41:07.280 --> 0:41:11.120 and then somehow magically the best one gets picked. Right. 0:41:11.560 --> 0:41:14.720 If that were how it worked, then quantum computers would solve, 0:41:14.840 --> 0:41:19.319 not only factoring, but also np complete problems. Right, they 0:41:19.360 --> 0:41:23.720 would break not only the cryptosystems that currently protect the Internet, 0:41:23.840 --> 0:41:27.360 but they would break all other possible cryptosystems you know 0:41:28.000 --> 0:41:30.360 that are based on hard problems. Okay, but that is 0:41:30.480 --> 0:41:33.000 not what we believe that quantum computers can do, right. 0:41:33.120 --> 0:41:35.279 We believe that they're more limited than that. So the 0:41:35.400 --> 0:41:37.400 question is why are they more limited? Okay? 0:41:37.719 --> 0:41:41.280 And it all has to do with the restrictions of measurement. 0:41:41.760 --> 0:41:41.880 Right. 0:41:42.280 --> 0:41:46.400 It's true that with a quantum computer you can create 0:41:46.760 --> 0:41:50.719 an equal superposition over all the possible solutions to your 0:41:50.800 --> 0:41:54.080 hard problem. That's even an easy thing to do if 0:41:54.120 --> 0:41:57.760 you have a quantum computer, right, like create a superposition 0:41:58.239 --> 0:42:00.480 or each of these two to the thousand and power 0:42:00.640 --> 0:42:04.200 possible solutions has some amplitude. You know, that's just like 0:42:04.719 --> 0:42:07.960 very simple, it's done, Okay. The trouble is for a 0:42:08.040 --> 0:42:10.800 computer to be useful, at some point you have to 0:42:10.840 --> 0:42:13.560 look at it. You have to measure, you have to 0:42:13.680 --> 0:42:17.000 get an output. You know, you have to read something out. Okay. 0:42:17.080 --> 0:42:19.600 And if you took an equal superposition over all the 0:42:19.680 --> 0:42:23.000 answers and you just measured it, not having done anything else, 0:42:23.440 --> 0:42:26.040 then the rules of quantum mechanics are very clear on 0:42:26.160 --> 0:42:28.960 what you're going to see. It's a completely random answer. 0:42:29.440 --> 0:42:32.120 And if you had just wanted a completely random answer, 0:42:32.400 --> 0:42:34.239 then you could have just flipped a coin a bunch 0:42:34.280 --> 0:42:36.879 of times, or just you know, use a random number 0:42:37.040 --> 0:42:40.160 generator that's inside your classical computer. Right, You didn't need 0:42:40.280 --> 0:42:42.520 to spend all these billions of dollars to build. 0:42:42.360 --> 0:42:43.160 A quantum computer. 0:42:43.680 --> 0:42:47.879 Right, So the only hope of getting an advantage from 0:42:47.960 --> 0:42:52.400 a quantum computer is to exploit the way that these amplitudes, 0:42:52.920 --> 0:42:59.400 you know, being complex numbers, work differently from conventional probabilities. Okay, 0:42:59.520 --> 0:43:04.320 And the central thing that amplitudes can do that probabilities 0:43:04.400 --> 0:43:07.399 cannot do is that they can interfere with each other. 0:43:07.920 --> 0:43:10.880 They can cancel each other out. And so with a 0:43:10.960 --> 0:43:15.520 quantum computer in particular, the idea with every algorithm for 0:43:15.640 --> 0:43:19.080 a quantum computer is that you are trying to choreograph 0:43:19.280 --> 0:43:22.320 things in such a way that for each wrong answer, 0:43:22.560 --> 0:43:25.239 because each answer that you don't want to see, some 0:43:25.560 --> 0:43:29.480 contributions to its amplitude are positive and others are negative, 0:43:29.800 --> 0:43:31.440 so they're canceling each other out. 0:43:32.040 --> 0:43:32.799 Whereas for the. 0:43:32.880 --> 0:43:35.880 Right answer, the answer you do one you want all 0:43:35.960 --> 0:43:40.320 the contributions to its amplitude to reinforce each other, okay, 0:43:40.440 --> 0:43:44.600 to add up constructively. If you can arrange that, then 0:43:44.640 --> 0:43:46.800 when you look, you're going to see the right answer 0:43:47.160 --> 0:43:48.600 with a large probability. 0:43:49.160 --> 0:43:50.319 That's the name of the game. 0:43:50.800 --> 0:43:53.640 Now. The hard part is you have to do all 0:43:53.719 --> 0:43:56.480 of that even though you yourself don't know in advance 0:43:56.560 --> 0:43:57.920 which answer is the right one. 0:43:58.280 --> 0:44:00.320 You know, if you already knew, what would be the point? 0:44:00.840 --> 0:44:03.680 Right, And you have to do all of this faster 0:44:04.120 --> 0:44:07.920 than even the cleverest classical algorithm could do the same thing. 0:44:08.080 --> 0:44:09.520 Otherwise what would be the point? 0:44:09.880 --> 0:44:13.760 Okay, so nature is giving us this really really bizarre 0:44:13.920 --> 0:44:15.920 hammer and a priori. 0:44:16.040 --> 0:44:17.520 It's not obvious whether. 0:44:17.320 --> 0:44:20.160 There's any nails that that hammer can hit, you know, 0:44:20.320 --> 0:44:24.640 other than just the obvious one of simulating quantum physics itself, right, 0:44:24.680 --> 0:44:26.640 And it really it took more than a decade for 0:44:26.719 --> 0:44:30.640 people to discover those nails and factoring the problem that 0:44:30.760 --> 0:44:33.960 Peter Shore designed his algorithm for. That was the first 0:44:34.280 --> 0:44:37.719 big example, and some people hope that that would be 0:44:37.880 --> 0:44:42.120 followed by a flood of other examples, and you know, unfortunately, 0:44:42.160 --> 0:44:45.319 you know, thirty years later, factoring remains one of our 0:44:45.400 --> 0:44:46.640 pre eminent examples. 0:44:47.320 --> 0:44:49.919 So I have a crazy question for you. You're telling 0:44:50.000 --> 0:44:53.879 me that quantum computers work by maintaining all of these 0:44:54.000 --> 0:44:57.480 different amplitudes and the superpositions, but that we don't have 0:44:57.680 --> 0:44:59.839 access to all the superpositions because we have to take 0:44:59.880 --> 0:45:02.200 the measurement. So we have to play clever games with 0:45:02.360 --> 0:45:05.560 interference so that we can use this system to do 0:45:05.680 --> 0:45:08.440 something useful. But we don't have access to the superpositions 0:45:08.480 --> 0:45:12.000 because of the measurement, only because we are classical objects 0:45:12.080 --> 0:45:15.800 and classical interactions with quantum systems collapse the measurement. What 0:45:16.000 --> 0:45:19.960 if we were tiny, we were microscopic, we were quantum, 0:45:20.560 --> 0:45:23.440 could then we use quantum systems in a way that 0:45:23.560 --> 0:45:26.880 didn't collapse all their wave functions and access all of 0:45:27.040 --> 0:45:27.960 those amplitudes. 0:45:28.480 --> 0:45:30.200 So I hate to break it to you, us being 0:45:30.280 --> 0:45:33.560 microscopic wouldn't help. Right, You can be as tiny as 0:45:33.600 --> 0:45:36.600 you like, But if you interact with a quantum system 0:45:36.920 --> 0:45:40.040 in a way that carries away the information about you know, 0:45:40.160 --> 0:45:43.000 which branch of the superposition we're we in, then that 0:45:43.200 --> 0:45:46.560 has exactly the same effect of collapsing the state. Right, 0:45:46.640 --> 0:45:50.400 So it's not our physical bigness that's the issue. It's that, 0:45:50.560 --> 0:45:55.600 you know, sort of we want an answer in our universe, right, 0:45:55.840 --> 0:45:58.279 what we can say, You know, there are people who 0:45:58.400 --> 0:46:02.320 are very gung ho about the interpretation of quantum mechanics, 0:46:02.360 --> 0:46:05.960 where they would say, collapse is not real, right, collapse 0:46:06.080 --> 0:46:09.800 is just a figment of our limited perspective. Right. Really, 0:46:09.880 --> 0:46:12.879 what's going on is it's just the Schrodinger equation all 0:46:12.960 --> 0:46:15.920 the way. And so they would say that whenever you know, 0:46:16.080 --> 0:46:19.600 you measure a quantum computer that's in a superposition, Actually 0:46:19.640 --> 0:46:23.000 the whole universe then splits into all these different branches, 0:46:23.480 --> 0:46:25.319 and each branch is equally real. 0:46:25.840 --> 0:46:28.520 We have an experience of one of them where. 0:46:28.320 --> 0:46:31.400 We perceive some answer. So a many worlder, that's what 0:46:31.600 --> 0:46:34.560 these people are called, right, a Mani worlder would say 0:46:34.960 --> 0:46:38.760 that there is some branch of the universal wave function 0:46:39.000 --> 0:46:42.000 in which you do get the right answer right, in 0:46:42.120 --> 0:46:45.719 which you try all the possible answers in superposition. You know, 0:46:45.800 --> 0:46:48.919 they would say, well, there's some branch where you get 0:46:49.000 --> 0:46:52.480 lucky and you measure the right answer, right. And they love, 0:46:52.920 --> 0:46:56.440 you know, all sorts of crazy thought experiments, like you know, 0:46:56.600 --> 0:46:58.480 quantum suicide, right, Like why not. 0:46:58.960 --> 0:47:00.640 Use a quantum random number? 0:47:00.760 --> 0:47:04.760 Generator to pick a lottery ticket and then just decide 0:47:04.840 --> 0:47:07.320 to kill yourself if it doesn't end up being winning, 0:47:07.680 --> 0:47:10.040 and then in all the branches of the wave function 0:47:10.200 --> 0:47:11.040 where you still. 0:47:10.880 --> 0:47:13.320 Exist, you know you'll have won the lottery. 0:47:13.719 --> 0:47:17.560 Now, I do not recommend that any listeners try that, Okay. 0:47:18.440 --> 0:47:22.040 I don't think there's any principle of reason that says, 0:47:22.120 --> 0:47:24.920 you know, you get to condition on being in a 0:47:25.000 --> 0:47:26.960 branch of the wave function where you're alive. 0:47:27.880 --> 0:47:30.479 Yeah, it sounds like a cool premise for a streaming show, 0:47:30.560 --> 0:47:31.680 but not a way to live your life. 0:47:31.920 --> 0:47:33.800 That is the one thing that I can say in 0:47:34.080 --> 0:47:36.560 the direction of what you were hoping for, and it's 0:47:36.640 --> 0:47:38.000 not very useful, I'm afraid. 0:47:54.440 --> 0:47:54.719 All right. 0:47:54.760 --> 0:47:56.759 So you're sort of famous for, you know, throwing cold 0:47:56.840 --> 0:48:01.360 water on mis explanations of quantum computing and maybe even overhype. 0:48:01.800 --> 0:48:04.000 Let me flip that around and ask you what is 0:48:04.160 --> 0:48:07.160 the most under hyped aspect of quantum computing. What is 0:48:07.200 --> 0:48:09.719 the thing that you're most excited about that would make 0:48:09.760 --> 0:48:12.160 you like, take your family money and invest it in 0:48:12.239 --> 0:48:14.799 somebody's quantum startup if you heard them doing it, if 0:48:14.840 --> 0:48:15.200 I had that. 0:48:15.320 --> 0:48:16.480 Kind of risk tolerance. 0:48:16.520 --> 0:48:18.600 There are so many things that I knew about when 0:48:18.640 --> 0:48:20.800 they were small, that I should have been investing in. 0:48:20.960 --> 0:48:23.759 And you know, clearly it's probably for the best that 0:48:23.880 --> 0:48:27.520 I became a professor and not an investor. I mean, 0:48:27.600 --> 0:48:29.520 one thing I think that a lot of people don't 0:48:29.560 --> 0:48:32.239 realize is that, you know, we are just within the 0:48:32.320 --> 0:48:36.719 last year the experimentalists have gotten very very close to 0:48:37.480 --> 0:48:41.160 the degree of control, you know, over cubits that would 0:48:41.200 --> 0:48:44.960 be needed to build a scalable quantum computer with with 0:48:45.080 --> 0:48:49.160 what we call quantum error correction, which is the technology 0:48:49.239 --> 0:48:51.680 that would ultimately allow you to sort of keep your 0:48:51.800 --> 0:48:57.200 cubits isolated, keep them from being prematurely measured by their environment, 0:48:57.640 --> 0:49:00.480 and you know, do an arbitrarily long quantum come computation 0:49:00.640 --> 0:49:02.920 with them. Right, So people have been talking about these 0:49:03.000 --> 0:49:05.920 ideas for thirty years now, and you know, so some 0:49:06.080 --> 0:49:09.600 people you may have gotten fatigue with quantum computing. What 0:49:09.719 --> 0:49:13.600 we've known since the mid nineteen nineties is that if 0:49:13.680 --> 0:49:18.080 you can control two pairs of cubits well enough, like 0:49:18.200 --> 0:49:21.239 if you can get the error the noise in a 0:49:21.400 --> 0:49:25.080 two cubit interaction to be sufficiently small, then there are 0:49:25.160 --> 0:49:28.680 these very very clever quantum error correcting codes that can 0:49:28.760 --> 0:49:31.280 get you the rest of the way that can encode 0:49:31.640 --> 0:49:35.920 like a single logical cubit across an entangled state of 0:49:36.160 --> 0:49:39.239 tens or hundreds of physical cubits, you know, in such 0:49:39.280 --> 0:49:42.160 a way that you can survive and recover from, you know, 0:49:42.280 --> 0:49:43.640 an error on any one of. 0:49:43.640 --> 0:49:44.719 The physical cubits. 0:49:45.040 --> 0:49:48.800 And you know, these codes have the effect of pushing 0:49:48.920 --> 0:49:53.319 your effective error rate down closer and closer to zero. Right, 0:49:53.600 --> 0:49:57.000 but only after you've passed this critical point at which 0:49:57.160 --> 0:50:00.840 the error correction becomes a net win, which it starts 0:50:00.920 --> 0:50:04.600 making things better rather than making them worse. It's almost 0:50:04.600 --> 0:50:06.960 as if like you have to pass the critical mass 0:50:07.239 --> 0:50:10.200 for a nuclear reaction. Right, if you're halfway there, you 0:50:10.239 --> 0:50:13.000 don't get half the reaction, right, you know, you need 0:50:13.120 --> 0:50:16.080 to pass criticality, okay, And so that's sort of been 0:50:16.160 --> 0:50:19.480 the engineering goal of the people who, unlike me, have 0:50:19.719 --> 0:50:23.239 labs and not just blackboards, right, who are actually trying 0:50:23.320 --> 0:50:25.759 to build these devices. You know, that's been their goal 0:50:25.880 --> 0:50:28.439 for thirty years, and I think you know, many people 0:50:28.600 --> 0:50:31.960 might not realize just how close they are. So basically, 0:50:32.080 --> 0:50:34.000 you know, the estimate is that you want to be 0:50:34.040 --> 0:50:37.920 able to, you know, apply a two cubit gait, that is, 0:50:38.000 --> 0:50:41.400 you know, do a desired operation on two cubits with 0:50:41.560 --> 0:50:45.839 about ninety nine point nine nine percent accuracy, about four 0:50:45.960 --> 0:50:48.960 nines of accuracy, and that ought to be enough to 0:50:49.120 --> 0:50:53.680 get this quantum error correction, you know, self sustaining reaction started. 0:50:54.120 --> 0:50:57.000 And when I joined the field, which was in the 0:50:57.120 --> 0:51:01.120 late nineteen nineties, such as a few years after shores 0:51:01.160 --> 0:51:04.279 and grovers algorithms had been discovered, right, it would have 0:51:04.360 --> 0:51:07.080 been amazing if you could do like a two cubit 0:51:07.200 --> 0:51:10.920 gate with fifty percent accuracy, right, Like that would have 0:51:10.960 --> 0:51:12.959 been a nature paper, okay. But then at some point 0:51:13.040 --> 0:51:17.040 the fifty percent became ninety percent, and then that became 0:51:17.200 --> 0:51:20.839 ninety five ninety nine percent, and within the last year 0:51:21.000 --> 0:51:24.840 we've seen like ninety nine point nine percent accuracy, okay, 0:51:25.040 --> 0:51:29.040 in several different groups. You know, the neutral atoms grew 0:51:29.200 --> 0:51:34.279 Querra and Boston, the trapped ions which like Quantinuum in 0:51:34.400 --> 0:51:39.279 Colorado does superconducting cubits, which at Google and IBM are doing. 0:51:39.840 --> 0:51:42.120 So you know, so a bunch of different approaches are 0:51:42.280 --> 0:51:45.560 you know, being pursued simultaneously, but you know, several of 0:51:45.680 --> 0:51:48.160 them are getting up to this, like ninety nine point 0:51:48.239 --> 0:51:51.440 nine percent accuracy, and it looks like just one more 0:51:51.560 --> 0:51:53.920 nine and you should be at the point where you 0:51:53.960 --> 0:51:57.120 know this self sustaining reaction works. So I think that's 0:51:57.200 --> 0:52:00.960 the central case for optimism right now. For just looking 0:52:01.040 --> 0:52:03.040 at the hour raid as a function of year, it 0:52:03.120 --> 0:52:05.440 looks like either you get there in the next decade 0:52:05.760 --> 0:52:07.600 or else something surprising. 0:52:07.160 --> 0:52:10.080 Happens, you know that explains why you didn't get there. 0:52:10.640 --> 0:52:13.320 That's one aspect of the story that maybe is not 0:52:13.520 --> 0:52:14.600 so well appreciated. 0:52:15.600 --> 0:52:17.719 All right, So that was our interview with Scott in 0:52:17.760 --> 0:52:21.640 conversation about quantum computing. I think the answer to the 0:52:21.760 --> 0:52:25.239 question should you be excited about quantum computing is yes. 0:52:25.760 --> 0:52:30.560 This potentially represents a whole new era of computing computers 0:52:30.760 --> 0:52:33.520 that are good at solving different kinds of problems than 0:52:33.560 --> 0:52:37.160 our traditional computers and opens our minds to the way 0:52:37.320 --> 0:52:40.560 computation can be done. Maybe there are other kinds of computers, 0:52:40.640 --> 0:52:44.080 not just classical or quantum computers, but other kind of things. 0:52:44.400 --> 0:52:46.840 Ways we can take advantage of the universe to do 0:52:46.960 --> 0:52:50.200 the calculations that we want to do. Thanks Scott very 0:52:50.280 --> 0:52:52.640 much for joining us, and thank you to everybody for listening. 0:52:52.800 --> 0:52:53.640 Tune in next time. 0:53:00.719 --> 0:53:04.520 Daniel and Kelly's Extraordinary Universe is produced by iHeartRadio. We 0:53:04.600 --> 0:53:06.960 would love to hear from you, We really would. 0:53:07.200 --> 0:53:09.840 We want to know what questions you have about this 0:53:10.160 --> 0:53:11.760 Extraordinary Universe. 0:53:11.920 --> 0:53:14.839 We want to know your thoughts on recent shows, suggestions 0:53:14.880 --> 0:53:17.879 for future shows. If you contact us, we will get 0:53:17.920 --> 0:53:18.279 back to you. 0:53:18.560 --> 0:53:22.000 We really mean it. We answer every message. Email us 0:53:22.120 --> 0:53:24.720 at Questions at Danielankelly dot. 0:53:24.719 --> 0:53:26.560 Org, or you can find us on social media. We 0:53:26.640 --> 0:53:30.520 have accounts on x, Instagram, Blue Sky and on all 0:53:30.600 --> 0:53:34.280 of those platforms. You can find us at d and Kuniverse. 0:53:34.440 --> 0:53:35.920 Don't be shy write to us