Solving Unsolvable Problems: The Future of Quantum Computing


RYAN MANDELBAUM: Good evening,
everybody, and welcome to the Chicago Quantum Summit– Solving Unsolvable
Quantums– the Future of Quantum Computing. My name is Ryan Mandelbaum. I’m a science writer at Gizmodo. So what is an
unsolvable problem? It is not what is
the meaning of life, and it is not why do I continue
to tweet even though it causes me so much pain and anguish. [LAUGHTER] But we’re looking at sort of
understanding the universe at the most intricate scales. What kinds of new
materials or medicines might come about from this
sort of new technology? So we’ve got a really
awesome program tonight. It’s going to start with a talk
from Talia Gershon from IBM. And then we’ll go into a panel. I’m just going to ask that
you hold any questions off until the end of the panel. I know you’re going
to have a lot, but we just want to make sure
that everything flows nicely. So to start, I’m going
to introduce Talia. Talia Gershon is the director
of research strategy and growth initiatives at IBM Research. She received her
PhD from Cambridge developing thin film materials
for solar energy applications, but she can explain quantum
computing better than anybody– certainly better than me. So Talia, why don’t you come up? [APPLAUSE] TALIA GERSHON: All right. Thank you, Ryan, for
the introduction. And it’s a pleasure and an
honor to be here talking to all of you about quantum computing. So I’ve got about 20 minutes. It’s not a lot of
time, but my hope is that I want to give you
at least a small flavor for what quantum computing
is, where we are, and how we got here. So I thought I’d start
with how we got here, because it’s a nice segue. So let’s see if I can
get my clicker to work. Clicker’s not– oh, there we go. OK. So I’m trying to
go back to the– there we go. All right, so rewind the
clock about a hundred years. Put yourself into the shoes of
scientists in the early part of the 20th century. You know, you have a
model of the world. It’s the classical physics. And you continue to study, you
know nature, and the universe. And you start to observe
a couple of things that you really can’t explain
with your classical physics. So let’s take a couple examples. First one is
superposition, right? So imagine you’re
studying electrons. And, you know, you believe
that these electrons can be either spin-up or spin-down. You start to observe
that, actually, they’re not always just either
spin-up or spin-down. Sometimes you find them
in some combination of spin-up and spin-down. So this was a really interesting
and surprising observation. Other things that people
discovered around that time– the uncertainty principle– this
idea that particles sometimes behave like waves. You know, waves sometimes
behave like particles. And really interesting
things were observed. Fast forward a couple of years– the observation of entanglement. So this was also another
surprising finding. So imagine you have your
two-electron system, and, you know, you believe
that you can describe them each individually. We’ve got this electron. I can describe the
properties of this one. We’ve got the second electron. I can describe the properties of
that one totally independently. But you observe
that, actually, you can do this thing where
they’re entangled. And suddenly, you can’t describe
them independently anymore. Any operation you do on
one of these electrons affects the whole system, right? So these observations that
there is these things that are happening that you can’t
explain with classical theory, this was really the birth
of quantum mechanics. Around the time, you
know, of mid-1900s, there was also huge innovation
in the field of computing. Fundamentally,
information theory– you know, progress
and computation around World War II, the
mention of the microprocessor and beyond. So there was a lot of ideas
and thinking around information theory itself. And around that time
in the ’60s, scientists like Charlie Bennett, who
is still an IBM researcher and roams our halls today,
started to ask the question, how can we think about the
efficiency of computation. Can we actually use ideas
from quantum mechanics to rethink information
itself, to rethink processing information itself? So, you know, we went back,
and we found some really early writings from Charlie
where he started to think about these ideas. And he wrote
quantum information, and then he scribbled
it out as false, right? But these things– people
were exploring these ideas and figuring things out. So this community began to form. In the ’80s– 1981– there was
this really famous conference that convened a lot of the
big thinkers in this space. And, you know, it was
a joint conference organized by MIT and IBM. And at that conference,
they were exploring, OK, so we have these ideas around
this quantum information theory. We think it’s going to be more
efficient for certain kinds of computations. What kinds of computations do
we think it could be better at? And one idea that came out
was simulating nature, right? And there’s a very famous
quote from Richard Feynman from that conference
where he said, nature isn’t classical, dammit. So if you want to
simulate nature, you better make it
quantum mechanical. And so people started
to think about, OK, well, that’s a
particular problem that’s very hard to simulate
on a conventional machine. We believe it could be done
more efficiently on a quantum computer. In the ’90s, a
new idea came out. We think we can– you
know, Peter Shor published a famous algorithm
where he said, I think we can actually factor
numbers significantly more efficiently on a
quantum computer. And this was a really big idea. So if you think about taking
two large prime numbers and multiplying them
together, that’s a relatively easy
computational task. But if you take that
new number, and you try to factor it
back into the primes, that’s a very hard
problem, right? Because is it divisible by two? Is it divisible by three? And so on and so forth. So when Peter Shor published
this new algorithm and said, actually, no. We can factor numbers
exponentially faster on a quantum computer, it was
very, very exciting, right? And so new ideas
kept coming out. We think this more efficient
form of computation will be interesting for
solving certain key problems. Then you’ve got to say, well, I
need a quantum computer if I’m going to do that, right? So towards the end of the
1900s, David DiVincenzo published a series of criteria. He said, OK, if
we’re really going to build a quantum
computer, it has to have some fundamental
properties, right? You need a carrier of
the quantum information. Just like how with
classical computing, we have the bit, with
quantum computing, we need a quantum bit or qubit. We need to be able to initialize
these qubits in a ground state. We need to be able to
control their properties and perform operations on
them called gates– logic operations called gates. So he published
a series of ideas that said, OK, if we’re
really going to build this, this is what it needs to be. In the– moving into the
2000s, we saw enormous progress in the experimental side of
quantum computing, right? So in the early 2000s,
we saw NMR technologies being used to prove
that you could factor 15 down into the
primes three and five– a demonstration. And then moving
forward a few years were the invention of
this new kind of qubit that’s actually patterned
on a processor, right? So this idea that
now we’re going to take a wafer just like we
took with classical machines, right? Classical computing’s
wafer-level devices– we’re going to actually pattern
qubits on silicon, right? In 2007 was the invention
of the transmon qubit, which is the technology we use
in IBM Research for our quantum systems. And it’s used by others as well. Fast forward a
couple years– a lot of improvements
in the technology. And in 2016, it was the
first-ever demonstration that we could put a real live
quantum computer– not live, obviously. It’s technology– a real
quantum computer on the cloud– you know, a five-qubit
machine that you could use the
operations that you need to do to actually
program a quantum computer. So it was a five-qubit system
with a universal gate set that people anywhere
in the world for free could access, and interact with,
and run research experiments, and explore, and learn. Since that time, we’ve gone
from having just one system on the cloud to having a fleet
of 14 systems on the cloud. And, you know, the technology’s
advanced from five, to 14, to 20, and even now 53 qubits
that we have, so the technology has made enormous progress. This is the sort of thing
that you can do now. You can log into an interface
that looks like this. This is the IBM Q Experience. You can log into an
interface like this, and all that quantum technology
is just abstracted from you. You can go in and play around
with these different kinds of gates and experiment
with how they work. And you can click that
little button that says Run, and it sends a real experiment
to a real quantum computer sitting in a lab in New York. It’s kind of mind boggling
when you think about it. So what’s actually
behind the scenes, right? What’s actually behind this? So we talked about
the need for qubits. So this is a picture
of an actual processor. These squares
represent our qubits. These are those transmon
qubits we talked about. And it’s not microscopic. This is not an individual atom. This is a millimeter
scale device, but it operates like an atom. So it has properties
where you can put it into a superposition state. You can entangle the state of
one qubit with another qubit. You have these
squiggly lines that represent how qubits
talk to each other and how they talk to
the control system outside of how you
do measurements and perform operations. And you have to cool the whole
thing down– you see there, 15 millikelvin. And the reason you
have to make it so cold is not just because
these systems are made from superconducting
materials, but because any
noise in the system will destroy your information. If you think about light,
which is quantum mechanical, you don’t want the light
hitting your quantum processor. You want to shield the processor
from as much noise as possible, so it lives all the way at
the bottom of a dilution refrigerator. I’m clicking the button,
but it’s not advancing. It lives all the
way at the bottom of a dilution refrigerator,
which looks like this. So you have your chip that’s in
that can to the bottom right. It’s inside of many
layers of shielding. And you have to communicate
with that processor, and we communicate
through microwaves. You saw those resonators. You communicate with a
chip through microwaves. And those microwaves
come in down through those coaxial
cables, those cables you see on the right. And, you know, it’s controlled
through a rack of electronics. So outside of the
system, you’ll have a rack of electronics, which is
creating those microwaves that communicate to the processor. And then you have to
close the whole thing up, because room temperature is very
far away from 15 millikelvins. So, you know, there’s
a lot of technology that goes into actually building
and creating these systems. And if I click the next– this is roughly how
it works, right? So you have the job that
you send from your machine to the quantum system, right? So you define. You see that interface roughly
resembles a quantum experience. You send the thing you
want to do to the system. These are the microwaves that
represent that operation you want it to perform on a chip. Those microwaves come into the
chip through those resonators. They perform an
operation on the qubits. And then you collect
the information out back through the
microwave electronics. And the answer is read out
through your interface. So what kinds of problems do
we believe that this can solve? I sort of alluded
in the beginning to simulating
nature is something we still believe these
quantum systems will be very powerful to do. Problems like factoring,
I mentioned already. But we believe that other
problems like optimization and machine learning–
there’s a huge opportunity to explore those applications
on a quantum computer. And, you know, that doesn’t
mean that quantum computers are going to solve everything. They’re not going to replace
classical machines, right? You’re still going
to have a laptop. You’re still going to
have your smartphone. But now you’re going to have
access to computational power to solve certain kinds of
problems you couldn’t solve before, but it’s
definitely not true that everything
will now be quantum. This is a device and
a computational system for really hard problems. So what does it look like to
actually simulate these things on a quantum computer? I talked about simulating
nature, simulating chemistry. One thing you might want to do
is calculate the bond length. So if you have two
atoms, and they’re going to bond with each other,
if they’re too far apart, the energy is higher than
the equilibrium energy. If you push them too close,
the energy also goes up. So there’s some
equilibrium bond length where there’s a minimum energy. And this is a result that we
published a couple of years ago that showed how
you would simulate these three different
molecules on a quantum system. These are small
molecules, right? So the straight line–
or the filled line– represents what the
actual calculation would be on a regular
classical machine, and then the dots
represent what answer you get if you simulate
on a quantum computer. And you can see that
there’s some noise that makes the experimental results
differ from the expected results. But what we showed a
year later is that, actually, there
can be techniques using the exact same hardware
where you mitigate the errors, right? So using experimental
techniques, if you understand
the noise enough, and you perform experiments
where you kind of stretch the pulses, you can
actually back calculate what your answer would have been
on a non-noisy quantum system. So tons of innovation happening
not only on the algorithm side, but figuring out
how we make the most use of these near-term devices. And we use techniques
like error mitigation to improve our ability
to get great results. So, you know, these
noisy systems– you might ask, OK, well,
are they actually going to be useful for doing
meaningful things. So there was a proof
that came out last year in 2018 where we showed
that, actually, there’s a mathematical proof
that shows that, yes, you can achieve quantum advantage. And by that, I mean
there’s an advantage, a computational advantage,
for using a quantum computer. You would choose to do
it because it’s better. That it’s actually
mathematically possible to achieve
quantum advantage even with these noisy machines. And shallow circuits
is about having– OK, you can’t have an
infinitely long set of gates that you perform, because the
information will eventually decohere. But so if you restrict
the number of gates you have, then you keep
the circuit shallow, then you can do
useful computation. So we’re not the only
ones advancing the field. That’s a really,
really important point. So we’ve got incredible
research scientists. We have a very exciting
research agenda, but just think about
having 160,000 people all over the world
accessing these devices and doing some really
exciting things. So I wanted to
spend a few minutes sharing some of the
innovation that’s come out of the community
leveraging these systems and really advancing the field. And just to make the
point that there’s a huge amount of
innovation happening outside in the
community– students all over the world really making
a difference in the field. So the first thing is research. So research scientists–
imagine a time– you’re a theoretical quantum
information scientist. Imagine a time before you
had access to a real machine. You know, you would innovate. You’d create these algorithms. And then you had no way to test
them on a real device, right? It used to be that
experimentalists would build experimental systems, and
they were the only ones that had access to them. And then theoreticians
would build algorithms, and they never had
access to a machine. And now there’s a way to bring
these two worlds together and to make it so
that theorists don’t need to have access and
build their own system, they can use one
through the cloud. And there’s been tons of
really exciting research that’s been published
leveraging these systems. Everything from studying
entanglement, studying new algorithms,
to studying noise, to building new
compilers, or thinking about different ways of
actually running algorithms on these machines. And a key enabler for this
is an open source project we launched called Qiskit. We launched it in 2017– Quantum Information
Software Kit. And I’ll talk a little bit
more about that in a bit. But huge amount of innovation
in the research community. 200 papers in just three years. So that’s really been
very exciting to see. The second one is education. So think about you’re a student. Imagine you’re in a
computer science class, and you don’t have an
actual computer, right? There’s so much learning
you do just by doing, that having access now
to an actual quantum computer in the classroom has
made a tremendous difference. We’ve seen a lot of universities
start to adopt actually real quantum computers in the
classroom, assigning homework and problem sets and projects
through this interface. There’s tons of educational
material out there. Not just that we
produce, but, you know, MIT has a course that
Isaac Chuang ran. We have a video series and
textbook that we’ve created. So tons of really exciting
things that have been enabled through access to
these kinds of systems. And then, finally,
you don’t have to be someone who’s
studying quantum information science as a discipline. You can just be anyone
with an interest, right? And you can have access to this. There is a huge community
online of people who talk to each other about
quantum computing and Qiskit through Slack. There’s been tens of
millions of experiments run. And it’s really anyone now. And I think that’s
very exciting. So, like I said, we launched
this quantum open source project. And it’s been really fascinating
to see the evolution of Qiskit. So Qiskit has four components. There’s Qiskit Terra, which
is down at the circuit level and at the gate
level, and this is where some of that work
on quantum compilers has been happening. There’s Qiskit Aqua, which
is a layer about algorithms. So what are some
algorithms that we want to run on quantum
systems, and can we define those in that layer? There’s Qiskit Ignis, which
is about error mitigation and techniques
around how to deal with error in these systems. And then there’s Qiskit Air,
which is our simulators, right? So how can we
simulate noise better? How can we simulate
different kinds of noise? And really thinking about
using that as a vehicle also for advancing the field. So I highly encourage
and welcome all of you to go check it out. It’s completely
free and available. And there’s lots of discussion
groups and support online. And we’ve also been
running hackathons, right? So lots of open source
projects have a community of people that get together,
and they advance the field together. So these are just
some of the things that we’ve seen
recently come out of the community at
these hackathons. One thing that was really
exciting– and again, these are not people who
studied, necessarily, quantum physics. Sometimes we’ve had
developers come in. We’ve had high school students
come in and participate in Qiskit hackathons. You can see that this team
that won the hackathon, they integrated PyTorch with Qiskit. PyTorch is a very popular
machine learning framework, so that was very cool
and interesting to see. So the key message
that I want to end on is basically that you don’t
have to think this is something that’s inaccessible to you. You don’t have to think that
this is somebody else’s, you know, innovation to drive. This is an open community. Tools are available to you. Resources are available to you. And it’s an opportunity for you
to engage and really advance the field. It’s a very exciting time in the
history of quantum computing. So thank you. [APPLAUSE] RYAN MANDELBAUM: Awesome. That was great. Thank you. So next up, we’re going
to have a little panel. We’re going to bring
onstage David Awschalom. David is the director of the
Chicago Quantum Exchange. He’s also a professor at
the University of Chicago and a scientist–
a senior scientist at Argonne National Lab. He’s one of the
leading researchers in spintronics and quantum
information engineering. And his team sort of
explores the weirdness of the small and the weird. And so he’s going to help
us understand this as well. Dave, why don’t you
come up on stage? [APPLAUSE] RYAN MANDELBAUM: So
to get started, I think that we maybe need to
take a step back and just ask what are these
machines actually going to do for us in this room. I mean, how is this
going to affect my life? I’m very selfish. TALIA GERSHON: Do
you want to start? I’ve been talking
for 20 minutes. I’ll keep going. So, you know, there’s
a couple of things. One of the things I think
is particularly exciting about this moment is that
there’s a set of things that we believe with evidence
that these machines will do, and then there’s a
whole host of things that we’ll discover just
by virtue of having them. If you think about the
early days of computing, you know, we had a
machine, and we knew what it would be useful for. In the early days,
there were some ideas, but just think of
everything people figured out they could do
by virtue of having access to one of those machines. And I think that’s a very
exciting thing, right? All the innovation
is going to happen just as people explore things. You saw some of the things
that we’re already looking at. How can we better
simulate nature, and we can we use that to, you know,
understand matter better and maybe discover and
invent new materials? And things like
machine learning– you know, these AI systems and
machine learning systems, they’re getting more complex. They’re getting more
computationally intensive, right? If you look at some of the
most advanced AI algorithms out there and the
models out there, they could take a long time to
train– like, weeks to train. So can we think
about actually making that more efficient
to allow us to get to even more complex systems
which can do even more? So those are just
some ideas about where things might be going. DAVID AWSCHALOM: Yeah, I
mean, I think your idea of, for example, trying to predict
new materials with a quantum machine means the future
could be incredibly exciting. You could think about room
temperature superconductors. If you could design a
material to create those, it would change everything
from energy transmission to levitating trains. And one of the problems of
moving in that direction is it’s hard to calculate the
properties of these materials. So I think the most
exciting thing in the field is we don’t really know,
actually, where we’re going. Well, we say it’s like
driving on low beams a hundred miles an
hour on the road. You’re going really fast,
but you don’t really know where you’re
going, and you’re sort of hoping nobody’s
going to stop you. [LAUGHTER] TALIA GERSHON: And
that’s the power of letting people
try things, right? Letting people explore–
there’s going to be so much discovery that happens just
by virtue of exploration. RYAN MANDELBAUM: So it’s sort
of just any industry that can use these optimization
algorithms, machine learning– I mean, it sounds like
pharmaceuticals, material science, energy,
pretty much anything that touches our lives– TALIA GERSHON: I can tell
you about a couple companies in particular that we know
are excited about this. So a couple of years ago,
IBM launched something called the IBM Q Network. There’s about 80
organizations spanning academia, national
labs, and companies in various industries. There’s JP Morgan Chase
is in there, so obviously some enthusiasm about
discovering applications for finance. Samsung is in there. Mitsubishi Chemical, right? So there really is a lot
of interest and fast movers in some of these
industries to say, I want to be the first, right? I want to have the
first advantage. DAVID AWSCHALOM:
Well, but I think it’s also because
the technology that’s used to build quantum
machines can also be used for other things beyond
computing, like communication and sensing. So these qubits that IBM has
worked so well to protect from the world, you
could flip them around and say, well,
don’t protect them. Expose them. And they’re
extraordinary sensors. So for medical imaging, right? For putting quantum
sensors in living cells, watching information
move, these are things that are happening right now. So I think it’s already
going to impact our life. RYAN MANDELBAUM: So why don’t we
get into that a little bit more then? So, I mean, what are
some of the places, you know, of these
quantum technologies that we ignore when
we’re talking just about quantum computing? I mean, can you get into maybe
some of the communications aspects and things like that? DAVID AWSCHALOM:
Yeah, so, I mean, you mentioned JP Morgan Chase. I think a lot of us are worried
about our information security. I am. And it be great if
I was going to be able to send you a message
using a protocol that would prevent eavesdropping. So one of the great things that
you mentioned, for example, about these quantum states
is the act of looking at them changes it. So you have built-in
security just with the laws of quantum physics. And building a global network
with complete security would change the way we think
about sharing information, right? And protecting things. And I think we’re
sensing, you know, things like understanding
proteins and us. You know, today, it’s hard
to understand more than 1% of the proteins in our
bodies, because you have to come up with a way to
crystallize them, and take them to machine to explore
them, and then argue at science meetings
about what you’re looking at. It’s a lot of fun. It pays the rent, but
at the end of the day, it’s not what you want, right? You want a tool that could
look at any arbitrary protein and get the structure
function relationship to understand how we work
and how pharmaceuticals work. Quantum [INAUDIBLE]
in principle, there’s a pathway to doing that. And I would argue,
if these qubits could be used for quantum
sensing at that level, that would revolutionize areas
of biology and chemistry. RYAN MANDELBAUM: And,
I mean, how is that going to integrate into
sort of the processor? I mean, is there sort
of a quantum landscape that you’re envisioning in the
future with a whole quantum infrastructure? TALIA GERSHON: I think
there’s so many– I mean, maybe this is,
again, a comment for David, because maybe it’s
closer to your area. But there are so many challenges
that you’d have to get past. Like, we talked about
the fact that, you know, all the things about– you
want to sense one thing, but you don’t want anything
else in the environment, probably, to influence
that measurement. You know, some of
these technologies are like trapped ions. How do you have a trapped ion
that you can control and put out in the world as a sensor? I think that some of the
technologies that you would have to invent for
quantum sensing are probably different
in some respects. DAVID AWSCHALOM: I
think that’s right. And I think you’re
right– you want to build an ecosystem
where quantum states are moved all around without ever
seeing the classical world. So you’d like the sensors,
the communication, the quantum computer all to be
moving this around. There are a lot of challenges. And the great thing about
research, of course, is unlike you where you actually
have to do something useful. [LAUGHTER] We don’t really–
the stockholders we have are all of you. So thank you for your
taxes, by the way. [LAUGHTER] But that means we can
actually reap the benefits of accidental discoveries. And I think a lot of
the things that you’ve been mentioning have
come out of kind of surprises in this
field, like the sensing. RYAN MANDELBAUM: So it seems
like there is a ton of promise. This sort of, like, whole
future that we can envision, but today, I would like to
know where we actually are. I mean, I just heard
this announcement of quantum supremacy. And it sounded weird and scary. So I wanted to know what
was actually going on there. TALIA GERSHON: OK,
I’ll start with kind of two observations about that. And then, David, obviously. So observation one is
just language, right? We’ve got to be careful
about what language we use, because there is a very clear
and deliberate misinformation about what that is. The name kind of
quantum supremacy actually originated when it
was a small insular quantum information community,
and they all knew what it meant
amongst themselves. They named an
experiment, and they named it Quantum Supremacy. And the intent was
never to communicate to the outside world what
we all interpret it to be. If I tell you I have this thing,
and I have quantum supremacy, it’s like quantum computers
taking over the world, and, like, you know,
infiltrating your home, and all this crazy stuff. And it’s not what it is. And it doesn’t even
really mean yet that there is a machine
that does something meaningful and useful
that we can’t do today with regular machines. So we’ve got to be really
careful with the language we use to talk about it. And we’ve got to
avoid the hype, right? So I think this whole field
struggles with some hype. Especially when you have
language like that making claims that you
can do a thing that would take 10,000 years to do. It just turns out that there was
no reason to say 10,000 years. We did an experiment where
we ran that same thing in two and a half
days we published the published the paper on. So there’s no reason to hype. There’s no reason to
sort of be deliberate about misinformation. So that’s number one. Number two is really about
how we define progress in the field. So in my opinion we should be
measuring progress in the field by how are we making
progress toward solving those meaningful problems. I mean, from my perspective,
one of the things that’s so exciting about this field
is we really believe it will be useful for meaningful problems. So we should be measuring
our progress towards that. So where we are– back to your
question of where we are– is that you saw some of our
demonstrations that, you know, not just us. There’s lots of people doing
really outstanding science. We’re at the place
where we’re taking those meaningful
problems and we’re demonstrating slow and steady
progress toward solving them. We are not yet–
nobody, this is not us. Not anybody has yet
demonstrated that they have a machine that can solve
meaningful problems better than others. But we believe that, over
the coming few years, we will be crossing a threshold
into, really, over the next maybe a couple of years, we’ll
start to see that quantum advantage that I talked
about, which is the ability to demonstrate that you can
do useful computation better than without a quantum
computer, that somebody would choose to use a quantum
computer because it offers you an advantage. DAVID AWSCHALOM: Yeah, I
agree with all of that. I mean, the language
is terrible. TALIA GERSHON: Yeah. [LAUGHTER] DAVID AWSCHALOM: Yes, that’s
my short answer to that. TALIA GERSHON: It’s misleading. DAVID AWSCHALOM: I don’t have
to sell anything to anyone. TALIA GERSHON: [LAUGHS] DAVID AWSCHALOM: But
quantum advantage means something, which really
means meaningful problems, what can you do. I do think there’s a value for
a large company saying something like this, even if it’s
not entirely correct. TALIA GERSHON: [LAUGHS] DAVID AWSCHALOM:
But it sets a point that people can start to probe. So when a company comes down
and says, we’ve done it. Now, there are a lot of
assumptions to make that true. And they get thousands of
people now looking at it and examining it and, maybe,
debunking parts of it, right? And it’s a way to start
focusing the field. And I think there’s
some value to that. But I agree with you. I think we’re not there. I think we will be there. And it’s very important
not to exaggerate. I mean, that’s the first thing
that will kill science fields, is to claim or push
things faster than reality because the reality is this
field is going really fast, really well. TALIA GERSHON: Already. DAVID AWSCHALOM: Already. TALIA GERSHON: And
it’s so exciting, and there’s no reason
to exaggerate it because the actual
truth is very exciting. DAVID AWSCHALOM:
Yeah, for example, here in Illinois,
the National Labs are funding a program
to teleport information between around a
30 mile distance. That’s going to happen shortly. And just to be able
to build a platform to teleport particles, not
people, just so we’re clear. [LAUGHTER] Though, to have
thought about that even a decade ago was
unthinkable, right? And the fact that you
can literally do it today is extraordinary. So we don’t need that. You’re absolutely right. RYAN MANDELBAUM: So
it’s almost like we’re in a place where quantum
technology is real. It exists. It is starting to get
to a point where it– we’re starting to explore
what its applications might be but not necessarily a
point where it’s actually doing those applications or
certainly not beating things. But we can start to actually
envision them, and it’s real. DAVID AWSCHALOM:
Yeah, I think it’s very much like saying
we’re at the vacuum tube stage, a quantum analogy. I’m guessing some people here
don’t know about vacuum tubes. [LAUGHTER] But they were like really big,
useless transistors, but– and there were
things you could do with that from building
radios [INAUDIBLE] microwave electronics. And then things became scalable,
and things really accelerated. And I think we’re
just at that stage, and it’s a little
dangerous to predict. RYAN MANDELBAUM: And so,
now, just kind of bringing it back to where we
are today, there are a lot of
different technologies that are coming about. Specifically, there is
different architectures for quantum computing. Would you be able to sort of
compare what exists today? And I know that there
is a trapped ion quantum computer, superconducting. I don’t– I would love to know
sort of the difference and what we can expect to see from them. TALIA GERSHON: OK, so– [LAUGHS] so the
technology we use is a superconducting
[INAUDIBLE],, which we talked about. Some of– there’s– all of
them have their trade-offs. So if you think about,
how do we envision– there’s a lot of things
we have to care about. How easy is it and
quick is it to actually send the signals in and out and
actually do operations and get the information back? You have to care about,
how scalable is it? How many can I fit? And how can I arrange them? And how can I build
larger and larger systems? Coherence times– how
long is my coherence time? That’s really a
measure of how long my information will last that
I can do useful computation. And a lot of these systems
have their trade-offs. So the technology we chose we
felt had the best trade-off. But others– for
example, the ions have longer coherence times. They’re a little
bit slower to probe. Right, so anything to– DAVID AWSCHALOM:
Yeah, I would say, for example, a technology that’s
the furthest behind, which, of course, what we work on– it’s the least
useful right now– is seeing if we could
take the trillion dollars of semiconductor
technology that’s driving a lot of
our economy today and use it to build
a quantum machine. It’s a challenge. It’s very hard to do this. But if you could
do that and there’s one patient enough
to make it work, then you’ll have a
scalable technology. And that’s the key,
making billions of quantum states,
where you really could build extraordinary machines. There’s an artist’s rendition
that’s on the wall around us. I’m staring at it. TALIA GERSHON: [LAUGHS] DAVID AWSCHALOM: It came
from one of the articles our students did
a number of years ago to show you could move
the quantum state from one electron, which is that
green dot on the wall, into the core of the atom, one
nuclear spin, come back later and extract it. So it’s a subatomic
quantum memory. It worked really unreliably
and not very well. It worked about a third
of the time, which, for physicists, is fantastic. [LAUGHTER] For IBM, not so good. But, scientifically,
it showed there’s no barrier for building
subatomic memories and making– literally,
in that memory, you can store billions
of variables fine. But to do it enough to
make something useful, not so much then, but
so I think exploring all of these material platforms,
atoms, superconductors, semiconductors– Intel has now launched
a semiconductor program, for example, for quantum
states because they want to use their fabs, to think
of, can you make it scalable? And they’re working with
Delft in the Netherlands and other places to
really push that. And it’s beginning to
work in a few bits no. RYAN MANDELBAUM: So they can
already make silicon stuff, so now they’re going to make
silicon quantum computers. DAVID AWSCHALOM:
Well, you said it. TALIA GERSHON: [LAUGHS] RYAN MANDELBAUM: They might. Well, you know, the talk is
called the Future of Quantum Computing, so I would love
to talk about what are some of the next major milestones. Maybe let’s start with
what IBM is working on. And then how are we going to get
to this quantum future, where we actually have quantum
technology as part of our lives? TALIA GERSHON: So
likely, this kind of goes back to the question
of how we measure progress. So we talked about progress
towards useful problems. There’s another unit
of progress that we’ve been documenting for
ourselves and we’ve been exploring it
with the community is this idea of quantum volume. So, frequently, when we talk
about quantum computing, we talk about a
number of qubits. But, actually, you can pattern
an infinite number of qubit– not an infinite. You could pattern a lot
of qubits on a chip. But the problem is you can’t
control them altogether reliably with all of those
nice properties of being able to untangle states and
control the whole thing. So it’s really about making
progress towards, not only the number of qubits,
but how good a job you’re doing at controlling
them appropriately. So how much can you
minimize errors? How well can you actually get
operations on multiple qubits at once? So we’ve kind of taken a
lot of these parameters that have to do with building
quantum systems, where you can actually make
meaningful computations, and we’ve rolled that into a
single parameter called quantum volume. So really, we have a
roadmap for quantum volume. We’ve seen a lot of
progress in quantum volume over the last few years. We want to continue
that progress. And if we do, then,
again, we believe that, in the coming
years, we’ll be able to demonstrate a
meaningful quantum advantage, where somebody would
choose to use a quantum computer to solve a
problem because they can solve it better. And it’s not just a quantum
computer to just add to that. It’s this partnership between
classical and quantum systems. There’s a classical computer
controlling the quantum computer, so it’s really about
getting these systems to work well together. DAVID AWSCHALOM:
Yeah, and I think it’s important to remember it’s
a global enterprise, where it’s not just in the United States. So, for example, a few years ago
was China’s both announcement and demonstration
of using a satellite to link quantum ground stations,
1,200 kilometers apart. It’s extraordinary to
think about building a global network of quantum
systems that you wire together. But I think one of the real
challenges for progress will be the workforce
so, particularly, for a lot of people
in this room. And when you talk to
the 70 odd companies at the quantum economic
development consortium and you ask them, what’s
their big concern? What do they worry
about with progress? The first thing they all
say is the workforce. Where will the people–
where will quantum engineers come from? And when you look back at the
formation of the electronics industry, that’s what drove it. Electrical engineers,
microwave engineers, radar– That, I think, will
actually limit the progress, believe it or not. I think it’s going
to be the workforce. RYAN MANDELBAUM:
Well, so how are we going to develop this
quantum infrastructure and workforce to actually
push things forward? TALIA GERSHON: So I believe that
we’re already probably starting to see it is that a
lot of the ideas that are behind the scenes of
these quantum computers are probably already making
their way into curricula. If you think about
it, I took a class that was extremely
memorable when I was an undergrad
in material science. And it was a class about
electronic, optical, and magnetic
materials and devices. And it taught me everything
about lasers and diodes and transistors
and photovoltaics and magnetic hard drives,
every device you could imagine. I’ll bet that class,
if it was taught today, would include
qubit technologies. So we’re going to
start to see, this is the technology that’s going
to become extremely important. That’s just going to
infiltrate a lot of curricula. So that’s I think one. And then I think
we’re also going to see– as the industry
grows and there’s a demand for the skills,
I think that will also drive the curricula and
the student interest, too. DAVID AWSCHALOM: I
think that’s all true. I think that’s right. It’s hard– one of the
things in this field that’s going to be
challenging is to get computer scientists thinking
about quantum machines. Anybody here a
computer scientist? OK, then I– RYAN MANDELBAUM: We had one. DAVID AWSCHALOM: That’s great. TALIA GERSHON: [LAUGHS] DAVID AWSCHALOM: One
of the tricky things about getting computer science
into quantum engineering is, for example, one curricula
that, traditionally, doesn’t teach quantum mechanics. And why should it
because it’s different. But you will need this
for quantum technology. And it’s a challenge to, say,
take a graduating student, who is offered a lucrative job at
IBM or Facebook or LinkedIn, and say, well, or you could
work on a machine that doesn’t really exist the way you want. And it’s not clear
where it’s going. But you should work there. [LAUGHTER] TALIA GERSHON: But
we are starting to see programs emerge that
are quantum computer science programs. And in fact, we’ve hired amazing
people out of those programs. So they’re emerging. DAVID AWSCHALOM: Absolutely. TALIA GERSHON: And
we are also seeing that there’s an opportunity
for people who have developed really deep skills in
one place to reapply those skills in this place. There’s a bunch of people
who are part of our HPC supercomputing team that are
now working with our quantum researchers because
simulating is a big deal. DAVID AWSCHALOM:
But I think we are going to need to get the help
of, say, high school teachers. TALIA GERSHON: Right. DAVID AWSCHALOM:
Somehow this is going to have to be conveyed at a
much earlier level than people heading to the job market. TALIA GERSHON: Sure. DAVID AWSCHALOM: That’s a
society challenge, which I’m sure we’ll do, by the way. But we need to figure out
a good way to do that. TALIA GERSHON: But it’s
interesting because I– I am sorry if you have
other– but This is an– RYAN MANDELBAUM:
I was [INAUDIBLE] DAVID AWSCHALOM:
It’s all right– TALIA GERSHON: –interesting
topic because– DAVID AWSCHALOM: –you
can say something– RYAN MANDELBAUM: [INAUDIBLE]
it’s pretty amazing that we’re getting
physicists to code because I’m a labs physicist,
and physicists are the worst coders. TALIA GERSHON: [LAUGHS] RYAN MANDELBAUM: No
offense physicists. We just– TALIA GERSHON: But
think about like– and probably some
of you can relate. And I don’t know. Maybe you’ve always
loved quantum mechanics. But I think a lot
of people who take quantum mechanics
in college, f they don’t see the applicability. DAVID AWSCHALOM: Well, that’s
because it’s taught terribly. [LAUGHTER] TALIA GERSHON: Could be. Could be. But when you study
a discipline and you don’t see the
practical application, it’s very hard to get excited
and motivated about it. And I think now, with all of
these exciting applications, people can get way more excited
when they take introduction to quantum mechanics. They know that there’s
something meaningful that can be done with it from
an engineer perspective. DAVID AWSCHALOM:
Actually, I think that’s a really good point. I know when I was taking
quantum mechanics, really? That’s all? Just a few minutes left? When I was taking
quantum mechanics or taking relativistic
quantum mechanics, you say, why the hell
am I taking this? TALIA GERSHON: Yeah. DAVID AWSCHALOM: It’s
completely useless. But now, you realize,
in fact, it’s not. And it’s an incredibly
important part of this emerging technology. So I think you’re right. I think that’s going to change
the discipline quite a bit. RYAN MANDELBAUM: Well, so how
is sort of academia and business and the government
working together to actually see this through
and make it actually happen? TALIA GERSHON: This is
a great one for you, about the National
Quantum Initiative. Maybe you could talk about that. DAVID AWSCHALOM: Yeah, so
a lot of us in this room were engaged in building
the National Quantum Initiative, which
was signed into law at the end of last year, which
is a billion dollar initiative set to get the US
going in building an infrastructure for quantum
science and technology. It’s fantastically
important because it gets universities and
companies working together on joint efforts with some
big significant goals. So I think the
United States has– it’s been a bipartisan effort,
something that I personally never seem to read about. But– I know. We’re not supposed to
talk about that, but– [LAUGHTER] It is something
that people agree on that is so important that
we just have to move forward. So I think that’s
a really good step. TALIA GERSHON: Yeah, and I think
the point you made earlier– there is a huge opportunity
for different parts of our ecosystem to focus on
different things like David talked a lot about some of the
really exploratory work his team is doing– DAVID AWSCHALOM: That’s
a nice way to put it. TALIA GERSHON:
–which has to happen. We need to be doing
that exploratory work. We need National Labs to be
setting those benchmarks. We need to be
engaging and defining what are these measurements
that we’re going to use to track our progress. We need industry to be focused
around what business value can we create and
startup ecosystems. So I think there’s a lot of
room for different players to play a different
important role. DAVID AWSCHALOM: And you
mentioned the National Labs strictly here in
Illinois as a good point. TALIA GERSHON: Yeah. DAVID AWSCHALOM:
The requirements for this technology are
unparalleled, historically. You need to build circuits where
you can place individual atoms where you want,
deterministically, which means by design. How do you even
know you’ve done it? You need some sort of
microscope to look at something at the atomic scale,
which often means you need a synchrotron
in places like here at Argonne, at Stanford,
which are billion dollar microscopes that the
national government has built for all sorts of studies. And so getting companies to
work with these systems, look at what they’ve
made, [INAUDIBLE] iterate back and forth, you
need all of this infrastructure to push this forward. It’s a good point. RYAN MANDELBAUM: Can you
now just envision for me– it’s 40 years from now. The ultimate quantum computer,
the universal [INAUDIBLE] TALIA GERSHON: [LAUGHS] RYAN MANDELBAUM:
–quantum computer exists. And the quantum technology and
communications have advanced. It’s actually in production. What are you most
excited to see? What are you doing
right now with it, and why are you
excited about it? TALIA GERSHON: So this may
not satisfy you as an answer, but I think it’s true. So we benefit from
technological innovations every day without
thinking about them. Just imagine the scale of
engineering and technology that went into the
creation of a smartphone. You don’t ever. Nobody ever thinks about that. But that’s decades,
years of innovation that has enabled us to have
a different quality of life. So I think that the technology
is going to progress, and it’s going to change
people’s quality of life. And it’s going to be behind
the scenes of some really important improvements
in the way we live. That’s my two cents on that. DAVID AWSCHALOM: Yeah, I
would echo that and also echo the fact that you’ll be
disappointed with any answer that I give– TALIA GERSHON: [LAUGHS] DAVID AWSCHALOM:
–because every time you try to predict the future,
you’re definitely wrong. TALIA GERSHON: Yeah. DAVID AWSCHALOM: But
there’s no question it’s going to impact
the way we live. When you think about having
small bits of matter that can store more
information than atoms in the observable universe,
which quantum systems can do, it will change a lot
of things and the way we deal with technology. And I think the exciting part
is we really don’t know yet. RYAN MANDELBAUM: To
be fair, backstage, David did say he envisioned
a caffeine molecule that was $10,000 times more potent. [LAUGHTER] But I think– DAVID AWSCHALOM: Thank you for
sharing that with the room, by the way. [LAUGHTER] RYAN MANDELBAUM: It was good. That’s what I was
hoping you’d say. So I think that it’s
probably time we just get to audience questions. I’m sure that some of you
are burning with questions. We’ll just start with
you in the front, and then I’ll work my way– AUDIENCE: One of the things I’ve
noticed is I started out a long time ago at the University of
Illinois in a totally different discipline and got into
[INAUDIBLE] computer [INAUDIBLE] which seems to be
the stage where [INAUDIBLE] well, the– oh, hello. [LAUGHTER] Right now, as we push
the speed through traditional linear computers
that operate so rapidly, we call it multitasking because
we do a lot of switching. It seems the weak link,
right now, in quantum will be the error
correction that is normally built into any one
of our computers right now. [INAUDIBLE] I
didn’t drop the mic. [LAUGHTER] The thing that is the weak
link right now, it seems, in the whole system
is the fact that we’re doing traditional linear
to just gather the data and control the system. What kind of work
possibly is being done on the error correction or– I don’t know what the term might
be– normalization of the data that you were you
were talking about, just to get it into
a compliant format for us to actually
look at, if we’ve got anything working with
quantums to do that function? RYAN MANDELBAUM: Yeah, I’d love
to hear about error correction as well. TALIA GERSHON: Yeah, I
think there’s actually two parts to the question. So part one is a little
bit more about how the computation works. I just want a quick
clarification there. So quantum computers are not
doing exactly the same series of operations that classical
computers do but faster. Actually, each operation on
a quantum computer is slower. The key is that it does the
computation a fundamentally different way and
more efficiently. So just to give you an
analogy, what a regular machine will do, like you said,
maybe you can paralyze it. But there’s a
sequence of operations that have to be performed
kind of one bit at a time if you get down to
the lowest level because you have to describe
each bit individually. And you have to address
each bit individually. But what quantum
computers are going to do is actually
create a whole system. f And the information
about the problem and then the processing you
do to manipulate that state is happening on
the whole system. AUDIENCE: My understanding
is that, normally, if you have a
traditional computer, it’s like putting a
mouse in the maze, and he’s trying all the examples
whereas quantum is you’re standing above the experiment. And you’re just looking
down, and you’re going, oh, there it is. TALIA GERSHON: [LAUGHS] It’s– there are–
basically, what you’re doing, at the end of
the day, is you’re creating a quantum state. You’re physically creating a
quantum state on a processor. You’re manipulating
a series of qubits to create a quantum state. And then what you
have to do is you have to take the problem you
want to solve and encode parts of it into that quantum state. And now, you– as the
algorithm continues, you’re manipulating that
whole quantum state. And at the end, the final
state represents the solution. That’s probably how you
should think about it. And that’s not the
same thing as I’m going to do all the same
operations in the same order. So that’s part one. And then part two is
about error correction. So you’re totally right
that error correction is a fundamentally different
thing in quantum systems because, in classical
information, you can just make a
copy of the information, and then you can check both and
say, you know, do they agree? Or do they not agree? If they don’t agree, you had
an error somewhere, right? You can’t make copies of
the quantum information. So there’s been
tons of innovation in trying to figure out how do
we actually correct for errors. There is a long term plan,
which is the millions of fault tolerant– millions of qubit
fault tolerant– they call them error
correction codes. There’s a way where
you put extra qubits. And those qubits
help you kind of, essentially, remove the errors
and the noise from the systems. And then you have the
computational qubits. Then there’s the
near term devices, the noisy, intermediate
scale quantum computers. And that’s where we are now. And I shared with
you some techniques that we have that allow
you– an idea that we shared at APS March meeting,
I think, last year was you could
stretch the pulses. So we talked about
microwaves come in. They manipulate
the quantum states. And it comes out. Maybe there’s errors there. But if you make
the pulses worse, you repeat it multiple
times, each time getting a little bit
worse, you can actually back calculate what
your answer would have looked like without noise. So there’s these
mitigation techniques that we’re looking at
to try to use near term, noisy, quantum computers. And those are slightly
different schemes than the sort of long term plan. Any other– DAVID AWSCHALOM: I
think we should maybe deal with some other questions. But just to say that basically
your question is incredibly important because what limits
the scaling of quantum machines is dealing with the errors. And the errors scale
in a horrible way. TALIA GERSHON: [LAUGHS] DAVID AWSCHALOM: It
gets harder and harder as you add more and more
of these quantum bits. So at the end of the day,
that’s the biggest challenge. RYAN MANDELBAUM: Next
question, and we’ll take somebody from
the middle back, that guy all the way in the back. AUDIENCE: OK. DAVID AWSCHALOM: Can someone
[INAUDIBLE] turn the lights on here so we– RYAN MANDELBAUM: Oh, sure. That would be great. DAVID AWSCHALOM: Can we turn the
lights up so we can see people? Or maybe not. [LAUGHTER] AUDIENCE: Hello, also on the
topic of error correction, today, Ivanka Trump
tweeted, It’s official. Explosion emoji. The US has achieved
quantum supremacy. TALIA GERSHON: This is
what I’m talking about. AUDIENCE: So I just wanted to
ask if you could correct her a little bit because,
maybe, you should scale down that statement because
I was thinking, maybe, it would be more
appropriate to say, we’ve demonstrated
quantum supremacy. But now that I’ve
heard you speak, I think it’s even
less than that. TALIA GERSHON: It is. [LAUGHTER] AUDIENCE: So– TALIA GERSHON: This
is exactly the point I was making about hype and
language because the words, they sound like
they mean the thing. But they don’t. It’s very dangerous for people
to misinterpret where we are. AUDIENCE: And they claimed– the Trump administration
claimed credit for this. [LAUGHTER] TALIA GERSHON: There’s so
many things wrong with this. AUDIENCE: Can you– RYAN MANDELBAUM: All
right, next question. AUDIENCE: –generate some
headlines about what it– by correcting her,
like what would she– what is real? TALIA GERSHON: Well, we
talked about where we are. We talked about
where we are, which is we’re at a stage where we
have systems that are getting sufficiently
sophisticated that we want to be able to apply them
towards meaningful problems. Nobody has yet–
nobody has yet proven that they can achieve
an advantage for solving meaningful problems in a
way classical machines can. Nobody. So the truth is– [INTERPOSING VOICES] TALIA GERSHON: That
we’re making progress. And we talked about
some of the progress. And we’ll continue
to see progress. And we should be
measuring progress in units of how well
are we doing and moving in a direction of being able
to solve meaningful problems. DAVID AWSCHALOM: So
I would say there are a lot of tweets that
come from the east coast that I may not pay
much attention to. [LAUGHTER] RYAN MANDELBAUM: It’s basic– DAVID AWSCHALOM: What
was really achieved? RYAN MANDELBAUM: They
solved one very specific, rather contrived problem and
claims that the problem was solved in 200 seconds, the
way that a supercomputer would solve it in 10,000 years. Now, there is a lot
of people who’ve already refuted that statement
or have published papers. It sort of, maybe,
have dialed it back. But it’s just one
very specific– they claim the first time the
quantum computer has beaten a supercomputer at anything. Now, whether that’s true
or not remains to be seen. But it just basically
means it’s not that. It’s not worth like the
tweeting explosion emoji. AUDIENCE: Because it
sounds like a weapons race. That is, in the
intelligence field or the geopolitical
field, the main thing is to get quantum technology to
break each other’s encryption. DAVID AWSCHALOM: Well, I think
one of the reasons people are interested and
governments are interested is
because, as Talia’s used this nice example about
factoring prime numbers, the mathematics is
used for factoring prime numbers [INAUDIBLE] the
base of today’s encryption schemes. So when you go on to Amazon
and you buy something with your Visa card, you
see the lock symbol appear. Your number is encrypted. And it’s based on the fact
that there isn’t a technology today that can decrypt
that number on the time scale of your transaction. So it’s secure. If a technology were
to come into play that could do that,
factoring the numbers and extracting your
original number, it would be a risk to the
current financial system. And that means banks
transferring just– yeah, it’s why JP
Morgan’s interested– safely moving money around,
just ordinary things. So that encryption
is important just for the security of
your own information. And that is something– as you said, quantum machines
aren’t good at a lot of things, in fact, many things. A calculator at Walgreens
will do pretty well compared to a quantum machine. [LAUGHTER] But what quantum machines
are exceptionally good at is factoring. TALIA GERSHON: And one
just additional thing to mention there is not
today’s quantum computers, that the ability to do
that meaningful factoring for numbers of sufficiently
interesting size, you do need large, fault
tolerant quantum computers. And I’m talking on the order
of multiple millions of qubits. So we’re still
pretty far from that, at least known algorithms. DAVID AWSCHALOM: And that’s
why this whole idea of building a scalable system
is so important because, if you can end
up building millions, then it is something
you need to think about. RYAN MANDELBAUM: Awesome, so
I love all of these questions. I just want to make sure
to get as many questions in, that we keep it to one
sentence and one question, if you don’t mind. And you’ve been– AUDIENCE: First
of all, thank you for a really
enjoyful discussion. And my question is about
that workforce problem. So I’m like in programming in
a computer science environment, and I wanted– when I speak
to people and [INAUDIBLE] like I’m working in quantum
information science, and they are really
excited about that. So there is really a big
interest among programmers, among classical programmers,
into that field. But that the sad
thing about that, they think that they
are dumb, so they can’t understand the physics. And even sadder,
I think they’re– probably they’re true
because quantum physics is really complicated. And once you are trying to
comprehend all these concepts, you inevitably will feel dumb. the. First question is, whether
to be a quantum programmer, do you need to comprehend
all those complex, like Feynman path
integrals, all that stuff, or you probably don’t
need to dig that deep? And if you do, how do
you find time for people to study all that stuff
because it really takes time? TALIA GERSHON: OK, so
there’s three things I want to say about that. Number one, I want to
talk about the computer scientists in our team, working
on quantum computing, right. Did they have a background
in it before they started? No, they did not. So we have teams of
software engineers that help deal with
everything in the quantum stack, everything from
the low level hardware– what is the software stack
that takes you up to the place that you can write algorithms
and they compile on the quantum machine? So there’s tons of
people in the team that had no background in quantum
computing when they came in, and they’re making really
important, meaningful contributions. And they’re right there
behind that Qiskit project. They’re really deeply
integrated in that. That’s number one. Number two is we’ve seen people
all the way from like early, early, like early in their
computer science degree contribute
meaningfully to Qiskit without having ever
had anything to do– this is outside of
our team– having them any background
whatsoever in quantum. And they didn’t have to. They created some really
interesting compiler innovations. So it’s about optimizing. You have gates, and you
want to compile them on an actual machine. There’s a compiler. And there is optimization
problems in there, and you don’t have to know a
lot about quantum information to do that. There is a need for people
in all kinds of places. And then the third
thing I wanted to say was we’ve been running
these Qiskit hackathons. So we had one in
Madrid last year. And tons of people
showed up, and they all had the exact same concern–
oh, I don’t know anything about quantum computing. Am I going to be
able to do anything? And every single team delivered
some really cool project at the end. And they were all
really excited to have been able to do
something meaningful. It’d be great to
have like a mentor to help you come up to speed. But it’s not beyond the reach
of computer scientists today. That’s my two cents. [LAUGHTER] RYAN MANDELBAUM: Classic. AUDIENCE: [INAUDIBLE]
developing the technology for writing [INAUDIBLE]
is one thing and writing the quantum
programs developed [INAUDIBLE] is another thing. TALIA GERSHON: Correct. That’s different. So if you want to create
new algorithms like, sure, you have to know something
about quantum information. And you have to study
it, just like anything. AUDIENCE: [INAUDIBLE] RYAN MANDELBAUM: I
want to get to a couple more questions, if
that’s OK, sorry. All right, next up,
yeah, right there. AUDIENCE: So I just [INAUDIBLE]. Quantum computing is very good
at simulating [INAUDIBLE] What is your definition of nature? And I know it’s very good at
simulating chemical particles, but is there like a greater
scope of definition of nature? TALIA GERSHON: Yeah, so
an oversimplification is to say that to
simulate a quantum system on a classical computer
is really inefficient. But to simulate a quantum
system on a quantum system, there’s a transformation. You can describe
the quantum state want to simulate like this. You can describe the
quantum state you’re going to create like that. And there’s a
transformation. so. It’s actually a lot more
efficient to simulate quantum systems on a quantum system. DAVID AWSCHALOM: But
maybe one example that, maybe, would help
answer your question is something like
photosynthesis, so that’s something
we see today. But you could argue it would be
nice to simulate that exactly on a quantum machine to
understand how light is turned into energy through
a chemical process and how nature restores energy. This is something you
can simulate perfectly on a quantum machine. And by learning how that’s
done, it could lead to new ways to store energy. RYAN MANDELBAUM: Right there. AUDIENCE: Yeah, so
[INAUDIBLE] traditionally, we’ve used tools like
logic, gates, and math to take advantage
of bits and bytes. For qubits, are those
the right tools to use? TALIA GERSHON: We
believe they are. So we talked about the David
[INAUDIBLE] of criteria. You need some carrier
of the information, and then you need some way by
which you can manipulate them to process that information. So there is a
different set of gates. They’re not the
same gates that you have for classical information. There’s a different
set of gates. There’s a Hadamard
gate that’s used to initiate superpositions. There’s like different
kinds of gates– qubit rotations. They’re not the
same kind of gates, but you need gate set
that you use and you create these circuits that
are part of the algorithm computation. RYAN MANDELBAUM: Closer, yeah. Yeah. AUDIENCE: So a lot of the
applications of quantum computing that we’ve
heard are all high level, where you need– it’s like big problems like
drugs and materials discovery. So I was just
wondering, would there ever come a point
where quantum computing would be cost efficient? In the sense, you have a
huge dilution fridge and– and I’m sure that’s– for a
big application, that’s fine. But do you think
there’s a thrust going towards making things which add
a little more, I say, I guess, economically accessible? RYAN MANDELBAUM: I
love that question. TALIA GERSHON: I’m
answering all of them. If you want to weigh in here– DAVID AWSCHALOM: I like that. TALIA GERSHON: The question
was whether quantum would ever be cost competitive. And the kinds of problems that– go ahead. Go ahead. OK, I’ve been talking too much. RYAN MANDELBAUM: Are we ever
going to see a quantum iPhone? And it’s going to
be in my pocket? [LAUGHTER] DAVID AWSCHALOM:
It’ll be in your head. [LAUGHTER] Your head not mine. [LAUGHTER] It depends what you
mean by cost efficient. When you think about the
approach of exascale computers, for example, for really
significant supercomputers and you look at the
energy they consume, next generation
machines will take up to, maybe, 40 to 50 megawatts of
power just to run the computer. So there alone, I would argue,
just some energy concerns, quantum calculations
will really help. Is that what your– TALIA GERSHON: That’s the idea. And as David said,
you’re never going to use it for two plus two, so
that’s not the kind of problem that we’re comparing
cost effectiveness. DAVID AWSCHALOM: Right. RYAN MANDELBAUM: So,
unfortunately, our time is up. I appreciate all of
these awesome questions. And thank you, again,
everybody for coming. Thanks again, Talia and David,
for coming and answering questions. TALIA GERSHON: Thank you. [APPLAUSE] RYAN MANDELBAUM: And yeah,
I really appreciate it.

4 thoughts on “Solving Unsolvable Problems: The Future of Quantum Computing

  1. How you going to low key hate on google. They talk about google hyping there results but act like the hype don't led to young minds being interested in the field.

  2. Looks at government murdering citizens. Looks at government fracking without remorse. Looks at government burning down all of the forests. Looks at government destroying everything on Mother Earth.

    Will the leaders take their warming quantum computers into outer space with them? Same quantum computers will know same humans killed the planet and same quantum computers will watch their murderous owners sleep. It's not rocket science to know what the result will be.

    I'm happy that computers will kill their owners. All of them. Why? Psychopaths have zero real desire to not be psychopathic.

  3. No positive comments? Thanks for uploading, editing this, making it available to the public and not just UC people.

Leave a Reply

Your email address will not be published. Required fields are marked *