Scott Aaronson: What is a Quantum Computer?

00:00:00.000 | - As you've said, quantum computing,

00:00:03.800 | at least in the 1990s, was a profound story

00:00:06.900 | at the intersection of computer science,

00:00:08.560 | physics, engineering, math, and philosophy.

00:00:10.920 | So there's this broad and deep aspect to quantum computing

00:00:14.880 | that represents more than just the quantum computer.

00:00:17.440 | But can we start at the very basics?

00:00:19.760 | What is quantum computing?

00:00:22.280 | - Yeah, so it's a proposal for a new type of computation,

00:00:27.760 | or let's say a new way to harness nature to do computation

00:00:31.720 | that is based on the principles of quantum mechanics.

00:00:34.800 | Okay, now the principles of quantum mechanics

00:00:37.080 | have been in place since 1926.

00:00:40.600 | You know, they haven't changed.

00:00:42.640 | You know, what's new is how we wanna use them.

00:00:45.680 | Okay, so what does quantum mechanics say about the world?

00:00:50.240 | You know, the physicists, I think, over the generations,

00:00:53.120 | you know, convinced people that that is an unbelievably

00:00:55.760 | complicated question and, you know,

00:00:57.760 | just give up on trying to understand it.

00:01:00.240 | I can let you in, not being a physicist,

00:01:03.080 | I can let you in on a secret,

00:01:04.480 | which is that it becomes a lot simpler

00:01:07.080 | if you do what we do in quantum information theory

00:01:10.320 | and sort of take the physics out of it.

00:01:12.440 | So the way that we think about quantum mechanics

00:01:15.760 | is sort of as a generalization

00:01:17.800 | of the rules of probability themselves.

00:01:20.280 | So, you know, you might say there was a 30% chance

00:01:25.320 | that it was going to snow today or something.

00:01:27.640 | You would never say that there was a negative 30% chance,

00:01:30.480 | right, that would be nonsense.

00:01:32.440 | Much less would you say that there was, you know,

00:01:34.280 | an I% chance, you know, square root of minus 1% chance.

00:01:38.760 | Now, the central discovery that sort of quantum mechanics

00:01:43.760 | made is that fundamentally the world is described by,

00:01:48.960 | or, you know, the sort of, let's say the possibilities

00:01:53.800 | for, you know, what a system could be doing

00:01:56.840 | are described using numbers called amplitudes, okay,

00:02:00.480 | which are like probabilities in some ways,

00:02:04.000 | but they are not probabilities.

00:02:05.780 | They can be positive, for one thing,

00:02:07.400 | they can be positive or negative.

00:02:09.520 | In fact, they can even be complex numbers, okay?

00:02:12.360 | And if you've heard of a quantum superposition,

00:02:14.960 | this just means that some state of affairs

00:02:18.160 | where you assign an amplitude,

00:02:20.360 | one of these complex numbers to every possible

00:02:23.560 | configuration that you could see a system in

00:02:27.260 | on measuring it.

00:02:28.280 | So for example, you might say that an electron

00:02:31.640 | has some amplitude for being here

00:02:34.360 | and some other amplitude for being there, right?

00:02:37.200 | Now, if you look to see where it is,

00:02:39.660 | you will localize it, right?

00:02:41.400 | You will sort of force the amplitudes

00:02:44.060 | to be converted into probabilities.

00:02:47.160 | That happens by taking their squared absolute value, okay?

00:02:50.200 | And then, you know, you can say either the electron

00:02:55.200 | will be here or it will be there.

00:02:57.560 | And, you know, knowing the amplitudes,

00:02:59.160 | you can predict at least the probabilities

00:03:01.520 | that you'll see each possible outcome, okay?

00:03:04.840 | But while a system is isolated

00:03:07.480 | from the whole rest of the universe,

00:03:09.600 | the rest of its environment,

00:03:11.320 | the amplitudes can change in time

00:03:13.760 | by rules that are different

00:03:16.320 | from the normal rules of probability

00:03:19.360 | and that are, you know, alien to our everyday experience.

00:03:22.440 | So anytime anyone ever tells you anything

00:03:25.280 | about the weirdness of the quantum world,

00:03:27.480 | you know, or assuming that they're not lying to you,

00:03:30.560 | right, they are telling you, you know,

00:03:32.600 | yet another consequence of nature

00:03:35.000 | being described by these amplitudes.

00:03:37.960 | So most famously, what amplitudes can do

00:03:40.720 | is that they can interfere with each other, okay?

00:03:42.960 | So in the famous double slit experiment,

00:03:46.160 | what happens is that you shoot a particle,

00:03:48.440 | like an electron, let's say,

00:03:50.120 | at a screen with two slits in it,

00:03:52.400 | and you find that there, you know, on a second screen,

00:03:55.840 | now there are certain places

00:03:57.520 | where that electron will never end up,

00:03:59.880 | you know, after it passes through the first screen.

00:04:04.280 | And yet, if I close off one of the slits,

00:04:07.280 | then the electron can appear in that place, okay?

00:04:10.160 | So by decreasing the number of paths

00:04:12.880 | that the electron could take to get somewhere,

00:04:15.400 | you can increase the chance that it gets there, okay?

00:04:18.120 | Now, how is that possible?

00:04:20.000 | Well, it's because, you know, as we would say now,

00:04:23.440 | the electron has a superposition state, okay?

00:04:26.480 | It has some amplitude for reaching this point

00:04:29.720 | by going through the first slit.

00:04:32.480 | It has some other amplitude for reaching it

00:04:34.560 | by going through the second slit.

00:04:36.360 | But now if one amplitude is positive

00:04:38.680 | and the other one is negative,

00:04:40.400 | then, you know, I have to add them all up, right?

00:04:42.720 | I have to add the amplitudes for every path

00:04:45.480 | that the electron could have taken to reach this point.

00:04:48.480 | And those amplitudes,

00:04:50.400 | if they're pointing in different directions,

00:04:52.500 | they can cancel each other out.

00:04:54.440 | That would mean the total amplitude is zero

00:04:56.800 | and the thing never happens at all.

00:04:58.940 | I close off one of the possibilities,

00:05:01.060 | then the amplitude is positive or it's negative

00:05:03.520 | and now the thing can happen.

00:05:05.160 | Okay, so that is sort of the one trick of quantum mechanics.

00:05:08.760 | And now I can tell you what a quantum computer is, okay?

00:05:11.760 | A quantum computer is a computer that tries to exploit,

00:05:16.760 | you know, these, exactly these phenomena,

00:05:20.320 | superposition, amplitudes, and interference

00:05:23.800 | in order to solve certain problems much faster

00:05:26.920 | than we know how to solve them otherwise.

00:05:29.120 | So it's the basic building block of a quantum computer

00:05:31.620 | is what we call a quantum bit or a qubit.

00:05:34.800 | That just means a bit that has some amplitude

00:05:37.200 | for being zero and some other amplitude for being one.

00:05:40.720 | So it's a superposition of zero and one states, right?

00:05:44.080 | But now the key point is that if I've got,

00:05:47.160 | let's say a thousand qubits,

00:05:49.580 | the rules of quantum mechanics are completely unequivocal

00:05:52.920 | that I do not just need one, you know,

00:05:55.320 | I don't just need amplitudes for each qubit separately.

00:05:58.400 | Okay, in general, I need an amplitude

00:06:00.960 | for every possible setting of all thousand of those bits.

00:06:05.120 | Okay, so that what that means

00:06:06.720 | is two to the 1000 power amplitudes.

00:06:09.720 | Okay, if I had to write those down,

00:06:12.480 | let's say in the memory of a conventional computer,

00:06:15.280 | if I had to write down two to the 1000 complex numbers,

00:06:18.600 | that would not fit within the entire observable universe.

00:06:22.240 | Okay, and yet, you know, quantum mechanics is unequivocal

00:06:25.720 | that if these qubits can all interact with each other,

00:06:28.720 | and in some sense, I need two to the 1000 parameters,

00:06:32.880 | you know, amplitudes to describe what is going on.

00:06:36.040 | Now, you know, now I can tell you know,

00:06:38.400 | where all the popular articles, you know,

00:06:41.040 | about quantum computing go off the rails is that they say,

00:06:44.340 | you know, they sort of say what I just said,

00:06:46.800 | and then they say, oh, so the way a quantum computer works

00:06:49.520 | is just by trying every possible answer in parallel.

00:06:52.760 | So, you know, that sounds too good to be true.

00:06:55.920 | And unfortunately, it kind of is too good to be true.

00:06:59.540 | The problem is I could make a superposition

00:07:02.880 | over every possible answer to my problem,

00:07:05.880 | you know, even if there were two to the 1000 of them,

00:07:08.640 | right, I can easily do that.

00:07:10.680 | The trouble is for a computer to be useful,

00:07:13.080 | you've got at some point, you've got to look at it

00:07:15.200 | and see an output, right?

00:07:17.280 | And if I just measure a superposition

00:07:19.620 | over every possible answer,

00:07:21.640 | then the rules of quantum mechanics tell me

00:07:23.560 | that all I'll see will be a random answer.

00:07:26.280 | You know, if I just wanted a random answer,

00:07:27.920 | well, I could have picked one myself

00:07:29.360 | with a lot less trouble, right?

00:07:31.200 | So the entire trick with quantum computing,

00:07:34.840 | with every algorithm for a quantum computer,

00:07:37.560 | is that you try to choreograph a pattern

00:07:40.560 | of interference of amplitudes.

00:07:43.160 | And you try to do it so that for each wrong answer,

00:07:46.240 | some of the paths leading to that wrong answer

00:07:48.840 | have positive amplitudes,

00:07:50.520 | and others have negative amplitudes.

00:07:52.680 | So on the whole, they cancel each other out.

00:07:55.160 | Okay, whereas all the paths leading to the right answer

00:07:58.140 | should reinforce each other, you know,

00:08:00.340 | should have amplitudes pointing the same direction.

00:08:02.800 | - So the design of algorithms in the space

00:08:05.360 | is the choreography of the interferences.

00:08:07.800 | - Precisely, that's precisely what it is.

00:08:09.800 | - Can we take a brief step back?

00:08:11.400 | And you mentioned information.

00:08:14.520 | - Yes.

00:08:15.360 | - So in which part of this beautiful picture

00:08:17.920 | that you've painted is information contained?

00:08:21.360 | - Oh, well, information is at the core of everything

00:08:24.240 | that we've been talking about, right?

00:08:25.720 | I mean, the bit is, you know,

00:08:27.700 | the basic unit of information.

00:08:30.000 | Since, you know, Claude Shannon's paper in 1948,

00:08:33.720 | you know, and you know, of course,

00:08:34.880 | people had the concept even before that,

00:08:36.960 | you know, he popularized the name, right?

00:08:39.920 | But I mean--

00:08:40.740 | - But a bit is zero or one.

00:08:42.520 | - That's right.

00:08:43.360 | - So that's the basic element of information.

00:08:44.180 | - That's right, and what we would say

00:08:45.120 | is that the basic unit of quantum information is the qubit,

00:08:49.280 | is, you know, the object, any object

00:08:51.740 | that can be maintained in a,

00:08:54.480 | manipulated in a superposition of zero and one states.

00:08:58.760 | Now, you know, sometimes people ask,

00:09:00.760 | well, but what is a qubit physically, right?

00:09:03.780 | And there are all these different, you know,

00:09:07.000 | proposals that are being pursued in parallel

00:09:09.440 | for how you implement qubits.

00:09:11.640 | There is, you know, superconducting quantum computing

00:09:14.400 | that was in the news recently

00:09:16.160 | because of Google's quantum supremacy experiment, right?

00:09:19.680 | Where you would have some little coils

00:09:24.200 | where a current can flow through them

00:09:26.960 | in two different energy states,

00:09:29.000 | one representing a zero, another representing a one,

00:09:32.240 | and if you cool these coils to just slightly

00:09:35.080 | above absolute zero, like a hundredth of a degree,

00:09:38.560 | then they superconduct, and then the current

00:09:41.440 | can actually be in a superposition

00:09:43.000 | of the two different states.

00:09:44.440 | So that's one kind of qubit.

00:09:47.000 | Another kind would be, you know,

00:09:49.160 | just an individual atomic nucleus, right?

00:09:52.400 | It has a spin.

00:09:53.640 | It could be spinning clockwise,

00:09:55.700 | it could be spinning counterclockwise,

00:09:57.840 | or it could be in a superposition of the two spin states.

00:10:00.700 | That is another qubit.

00:10:02.240 | But see, just like in the classical world, right,

00:10:04.920 | you could be a virtuoso programmer

00:10:07.500 | without having any idea of what a transistor is, right,

00:10:11.080 | or how the bits are physically represented

00:10:13.680 | inside the machine,

00:10:15.040 | even that the machine uses electricity, right?

00:10:17.920 | You just care about the logic.

00:10:19.680 | It's sort of the same with quantum computing, right?

00:10:21.920 | Qubits could be realized

00:10:23.620 | by many, many different quantum systems,

00:10:26.020 | and yet all of those systems will lead to the same logic,

00:10:29.140 | you know, the logic of qubits,

00:10:31.940 | and how you measure them, how you change them over time.

00:10:36.140 | And so, you know, the subject of how qubits behave

00:10:40.300 | and what you can do with qubits,

00:10:42.100 | that is quantum information.

00:10:43.900 | - So just to linger on that.

00:10:45.580 | - Sure.

00:10:46.420 | - So the physical design implementation of a qubit

00:10:50.200 | does not interfere with that next level of abstraction

00:10:55.200 | that you can program over it.

00:10:57.120 | So it truly is, the idea of it is, okay.

00:11:01.720 | - Well, to be honest with you,

00:11:03.460 | today they do interfere with each other.

00:11:05.440 | That's because all the quantum computers

00:11:07.580 | we can build today are very noisy, right?

00:11:10.080 | And so sort of the qubits are very far from perfect,

00:11:15.080 | and so the lower level sort of does affect the higher levels,

00:11:18.680 | and we sort of have to think about all of them at once.

00:11:21.240 | Okay, but eventually where we hope to get

00:11:23.920 | is to what are called error-corrected quantum computers,

00:11:27.240 | where the qubits really do behave

00:11:29.280 | like perfect abstract qubits for as long as we want them to.

00:11:33.520 | And in that future, you know, a future that we can already,

00:11:37.920 | you know, sort of prove theorems about or think about today,

00:11:41.160 | but in that future, the logic of it

00:11:44.240 | really does become decoupled from the hardware.

00:11:46.700 | - So if noise is currently like the biggest problem

00:11:50.940 | for quantum computing, and then the dream

00:11:54.100 | is error-correcting quantum computers,

00:11:57.780 | can you just maybe describe what does it mean

00:12:01.340 | for there to be noise in this system?

00:12:03.460 | - Absolutely.

00:12:04.420 | So yeah, so the problem is even a little more specific

00:12:07.260 | than noise, so the fundamental problem,

00:12:09.980 | if you're trying to actually build a quantum computer,

00:12:13.100 | you know, of any appreciable size,

00:12:15.940 | is something called decoherence.

00:12:18.460 | Okay, and this was recognized from the very beginning,

00:12:21.020 | you know, when people first started thinking about this

00:12:23.220 | in the 1990s.

00:12:24.700 | Now, what decoherence means is sort of

00:12:27.740 | the unwanted interaction between, you know, your qubits,

00:12:31.620 | you know, the state of your quantum computer

00:12:33.940 | and the external environment.

00:12:35.860 | Okay, and why is that such a problem?

00:12:37.860 | Well, I talked before about how, you know,

00:12:39.940 | when you measure a quantum system,

00:12:42.460 | so let's say if I measure a qubit

00:12:44.700 | that's in a superposition of zero and one states

00:12:46.940 | to ask it, you know, are you zero or are you one?

00:12:49.460 | Well, now I force it to make up its mind, right?

00:12:51.860 | And now probabilistically it chooses one or the other,

00:12:55.620 | and now, you know, it's no longer a superposition,

00:12:58.060 | there's no longer amplitudes,

00:12:59.780 | there's just there's some probability that I get a zero

00:13:02.260 | and there's some that I get a one.

00:13:03.960 | And now the trouble is that it doesn't have to be me

00:13:09.900 | who's looking, okay?

00:13:11.060 | In fact, it doesn't have to be any conscious entity,

00:13:13.700 | any kind of interaction with the external world

00:13:18.940 | that leaks out the information about whether this qubit

00:13:22.940 | was a zero or a one, sort of that causes the zeroness

00:13:27.100 | or the oneness of the qubit to be recorded in, you know,

00:13:30.940 | the radiation in the room, in the molecules of the air,

00:13:35.180 | in the wires that are connected to my device, any of that.

00:13:40.660 | As soon as the information leaks out,

00:13:42.740 | it is as if that qubit has been measured, okay?

00:13:45.780 | It is, you know, the state has now collapsed.

00:13:49.860 | You know, another way to say it is that it's become

00:13:51.740 | entangled with its environment, okay?

00:13:54.100 | But, you know, from the perspective of someone

00:13:56.580 | who's just looking at this qubit,

00:13:58.180 | it is as though it has lost its quantum state.

00:14:01.480 | And so what this means is that if I want to do

00:14:04.460 | a quantum computation, I have to keep the qubits

00:14:08.300 | sort of fanatically well isolated from their environment.

00:14:12.180 | But then at the same time, they can't be perfectly isolated

00:14:15.060 | because I need to tell them what to do.

00:14:17.260 | I need to make them interact with each other, for one thing,

00:14:20.540 | and not only that, but in a precisely choreographed way.

00:14:24.460 | Okay, and, you know, that is such a staggering problem,

00:14:27.660 | right, how do I isolate these qubits from the whole universe

00:14:31.100 | but then also tell them exactly what to do?

00:14:33.500 | I mean, you know, there were distinguished physicists

00:14:36.180 | and computer scientists in the '90s who said,

00:14:39.300 | this is fundamentally impossible, you know?

00:14:41.740 | The laws of physics will just never let you control qubits

00:14:45.260 | to the degree of accuracy that you're talking about.

00:14:48.860 | Now, what changed the views of most of us

00:14:52.300 | was a profound discovery in the mid to late '90s,

00:14:56.900 | which was called the theory of quantum error correction

00:15:00.060 | and quantum fault tolerance, okay?

00:15:02.300 | And the upshot of that theory is that if I want to build

00:15:05.980 | a reliable quantum computer and scale it up

00:15:09.380 | to an arbitrary number of as many qubits as I want,

00:15:12.940 | you know, and doing as much on them as I want,

00:15:15.660 | I do not actually have to get the qubits

00:15:18.220 | perfectly isolated from their environment.

00:15:20.740 | It is enough to get them really, really,

00:15:22.620 | really well isolated, okay?

00:15:24.700 | And even if every qubit is sort of leaking its state

00:15:29.700 | into the environment at some rate,

00:15:32.500 | as long as that rate is low enough, okay,

00:15:35.140 | I can sort of encode the information that I care about

00:15:39.780 | in very clever ways across the collective states

00:15:43.060 | of multiple qubits, okay, in such a way that even if,

00:15:46.580 | you know, a small percentage of my qubits leak,

00:15:49.740 | well, I'm constantly monitoring them

00:15:51.580 | to see if that leak happened.

00:15:53.100 | I can detect it and I can correct it.

00:15:55.660 | I can recover the information I care about

00:15:58.020 | from the remaining qubits, okay?

00:16:00.100 | And so, you know, you can build a reliable quantum computer

00:16:04.940 | even out of unreliable parts, right?

00:16:07.420 | Now, in some sense, you know, that discovery

00:16:11.460 | is what set the engineering agenda

00:16:13.780 | for quantum computing research

00:16:15.740 | from the 1990s until the present, okay?

00:16:18.360 | The goal has been, you know, engineer qubits

00:16:21.380 | that are not perfectly reliable, but reliable enough

00:16:25.300 | that you can then use these error-correcting codes

00:16:28.500 | to have them simulate qubits

00:16:30.900 | that are even more reliable than they are, right?

00:16:33.700 | The error correction becomes a net win

00:16:35.740 | rather than a net loss, right?

00:16:37.420 | And then once you reach that sort of crossover point,

00:16:40.540 | then, you know, your simulated qubits

00:16:42.940 | could in turn simulate qubits

00:16:44.860 | that are even more reliable and so on

00:16:47.140 | until you've just, you know, effectively,

00:16:49.260 | you have arbitrarily reliable qubits.

00:16:51.740 | So long story short, we are not at that break-even point yet.

00:16:55.380 | We're a hell of a lot closer than we were

00:16:57.300 | when people started doing this in the '90s,

00:16:59.420 | like orders of magnitude closer.

00:17:01.020 | - But the key ingredient there is the more qubits,

00:17:03.740 | the better, because--

00:17:05.700 | - Well, the more qubits,

00:17:06.780 | the larger the computation you can do, right?

00:17:09.820 | I mean, qubits are what constitute the memory

00:17:13.520 | of your quantum computer, right?

00:17:15.060 | - But also for the, sorry, for the error-correcting mechanism.

00:17:18.060 | - Yes, so the way I would say it

00:17:20.660 | is that error correction imposes an overhead

00:17:23.480 | in the number of qubits.

00:17:24.980 | And that is actually one of the biggest practical problems

00:17:28.060 | with building a scalable quantum computer.

00:17:30.260 | If you look at the error-correcting codes,

00:17:32.740 | at least the ones that we know about today,

00:17:35.140 | and you look at, you know, what would it take

00:17:37.320 | to actually use a quantum computer

00:17:39.220 | to, you know, hack your credit card number,

00:17:43.900 | which is, you know, maybe, you know,

00:17:45.260 | the most famous application people talk about, right?

00:17:47.960 | Let's say to factor huge numbers

00:17:49.960 | and thereby break the RSA cryptosystem.

00:17:52.660 | Well, what that would take would be thousands of,

00:17:56.220 | several thousand logical qubits,

00:17:58.800 | but now with the known error-correcting codes,

00:18:01.220 | each of those logical qubits would need to be encoded itself

00:18:04.740 | using thousands of physical qubits.

00:18:07.040 | So at that point, you're talking about

00:18:08.580 | millions of physical qubits.

00:18:10.580 | And in some sense, that is the reason why quantum computers

00:18:13.660 | are not breaking cryptography already.

00:18:15.580 | It's because of these immense overheads involved.

00:18:18.540 | - So that overhead is additive or multiplicative?

00:18:21.060 | - Well, it's multiplicative.

00:18:22.180 | I mean, it's like you take the number of logical qubits

00:18:26.580 | that you need in your abstract quantum circuit,

00:18:29.120 | you multiply it by a thousand or so.

00:18:31.440 | So, you know, there's a lot of work on, you know,

00:18:33.560 | inventing better,

00:18:34.620 | trying to invent better error-correcting codes.

00:18:37.120 | Okay, that is the situation right now.

00:18:38.900 | In the meantime, we are now in what physicist John Preskill

00:18:43.900 | called the noisy intermediate scale quantum or NISQ era.

00:18:48.260 | And this is the era, you can think of it

00:18:50.140 | as sort of like the vacuum, you know,

00:18:51.760 | we're now entering the very early vacuum tube era

00:18:55.020 | of quantum computers.

00:18:56.480 | The quantum computer analog of the transistor

00:18:59.380 | has not been invented yet, right?

00:19:01.200 | That would be like true error correction, right?

00:19:03.780 | Where, you know, we are not, or something else

00:19:06.200 | that would achieve the same effect, right?

00:19:08.140 | We are not there yet.

00:19:10.100 | And, but where we are now, let's say as of a few months ago,

00:19:14.860 | you know, as of Google's announcement of quantum supremacy,

00:19:18.300 | you know, we are now finally at the point

00:19:20.420 | where even with a non-error-corrected quantum computer,

00:19:23.960 | with, you know, these noisy devices,

00:19:25.980 | we can do something that is hard

00:19:28.480 | for classical computers to simulate, okay?

00:19:31.140 | So we can eke out some advantage.

00:19:33.320 | Now, will we in this noisy era be able to do something

00:19:36.880 | beyond what a classical computer can do

00:19:38.900 | that is also useful to someone?

00:19:41.100 | That we still don't know.

00:19:42.380 | People are going to be racing over the next decade

00:19:45.180 | to try to do that by people, I mean, Google, IBM,

00:19:49.260 | you know, a bunch of startup companies, you know--

00:19:52.180 | - And research labs.

00:19:53.220 | - Yeah, and research labs and governments

00:19:55.860 | and yeah, so--

00:19:57.220 | - You just mentioned a million things.

00:19:58.700 | Well, I'll backtrack for a second and ask.

00:20:00.300 | - Yeah, sure, sure.

00:20:01.820 | - So we're in these vacuum tube days.

00:20:04.260 | - Yeah, just entering, though.

00:20:05.740 | - And just entering, wow, okay.

00:20:07.860 | So how do we escape the vacuum?

00:20:11.060 | So how do we get to where we are now with the CPU?

00:20:16.060 | Is this a fundamental engineering challenge?

00:20:19.380 | Is there breakthroughs on the physics side

00:20:23.820 | that are needed on the computer science side?

00:20:27.340 | Is it a financial issue where a much larger

00:20:32.380 | just sheer investment and excitement is needed?

00:20:35.820 | - So, you know, those are excellent questions.

00:20:38.460 | My guess-- - No answers.

00:20:39.860 | - Well, no, no, my guess would be all of the above.

00:20:44.060 | I mean, my guess, you know, I mean,

00:20:46.820 | you know, you could say fundamentally

00:20:48.140 | it is an engineering issue, right?

00:20:49.860 | The theory has been in place since the '90s,

00:20:52.780 | you know, at least, you know,

00:20:55.100 | this is what, you know, error correction would look like.

00:20:58.100 | You know, we do not have the hardware that is at that level.

00:21:01.500 | But at the same time, you know,

00:21:03.140 | so you could just, you know, try to power through,

00:21:07.020 | you know, maybe even like, you know,

00:21:08.980 | if someone spent a trillion dollars

00:21:11.260 | on some quantum computing Manhattan project, right,

00:21:14.180 | then conceivably they could just, you know,

00:21:17.740 | build an error-corrected quantum computer

00:21:21.220 | as it was envisioned back in the '90s, right?

00:21:24.300 | I think the more plausible thing to happen

00:21:27.140 | is that there will be further theoretical breakthroughs

00:21:30.180 | and there will be further insights

00:21:32.060 | that will cut down the cost of doing this.

00:21:34.620 | [BLANK_AUDIO]

00:21:42.980 | [BLANK_AUDIO]

00:21:52.980 | you

Scott Aaronson: What is a Quantum Computer? | AI Podcast Clips

Chapters