April 28, 2026 • 42 mins
In Episode 139, Patrick and Ciprian are joined by Bert de Jong, senior scientist at Lawrence Berkeley National Laboratory. The team discusses quantum computing's role in material science and energy, exploring industry challenges and strategic partnerships. The conversation emphasizes innovation urgency and national labs' influence on the future.
Bert de Jong is the Director of the Quantum Systems Accelerator, which is part of the National Quantum Initiative. In addition, de Jong is the Team Director of the Accelerated Research for Quantum Computing (ARQC) Team MACH-Q, funded by DOE ASCR, focused on developing software stacks for near-term quantum computing devices. In addition, de Jong has a program in AI and machine learning to understand biomolecular processes, and discover new materials and molecular crystals for gas adsorption. de Jong serves as the Department Head for Computational Sciences, and leads the Applied Computing for Scientific Discovery Group, which advances scientific computing by developing and enhancing applications in key disciplines, as well as developing HPC, quantum and AI tools and libraries for addressing general problems in computational science.
Listen
Watch
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
SPEAKER_02 (00:10):
In episode 139, Patrick Insecret speaker with
senior scientists at theLawrence Berkeley National
Laboratory, Bert De Young, todiscuss quantum advantage for
scientific discovery, chemistryand material science, and the
pursuit of the right modality.
Welcome to Entangled Things,your quantum computing podcast,
(00:34):
hosted by Patrick and Cyprien.
SPEAKER_03 (00:40):
Hey Cyprian, how are you doing?
SPEAKER_00 (00:42):
Hey Patrick.
I'm doing great.
Looking forward for anotherepisode of Entangle Things.
SPEAKER_03 (00:46):
Well, this is a good one today.
So, Bert, do you mindintroducing yourself to our
audience?
SPEAKER_01 (00:52):
Sure.
Well, Patrick and Brittany,thank you for having me here.
So I'm Bert DeYoung.
I'm a senior scientist here atLawrence Berkeley National Lab.
So one of the uh 17 Departmentof Energy national labs.
Um what drives me with quantumis so right now I lead the
Quantum System Accelerator,which is one of the five
(01:12):
national quantum initiativecenters.
Um that's a large investment bythe Department of Energy focused
on uh on delivering kind of thenext generation quantum
technologies and not just thetechnologies, but also starting
to demonstrate that there is aquantum advantage or a
usefulness uh for scientificdiscovery using quantum
(01:34):
computing specifically.
That is our focus.
Uh so the five centers, in caseyou uh you're not familiar with
that, uh the National QuantumInitiative is something that uh
uh Congress has been uh umputting in place for the last
five years.
They are reauthorising it rightnow, um, which really lays out
(01:55):
the drive for the US to be aleader in quantum technologies.
Um so these five centers are theDepartment of Energy's effort to
really show advance thetechnology and show the
opportunities and uh the uhchallenges and how we can solve
those uh in the next, well, forus five years.
(02:15):
We just got renewed as fivecenters.
SPEAKER_03 (02:17):
Excellent.
SPEAKER_01 (02:18):
Um our undersecretary Dario Gill
announced that late November.
Um, so we have been running umto really start this process and
and really making significantadvances.
SPEAKER_03 (02:31):
That's amazing.
SPEAKER_01 (02:32):
Um so I can tell you a little bit about my center.
Please.
Uh my center has beendeveloping, um we have a very
broad view of how we're going toapproach this.
Uh so first of all, I shouldhave said maybe I'm a
computational chemist bytraining, so I'm not a quantum
physicist.
Um I come from a uh a backgroundwhere I want to solve scientific
(02:53):
problems.
So these kind of what you heardat the beginning, we want to
solve problems, this is reallywhat drives anything that I do.
So I lived in a in in for a longtime in the high performance
computing world, so in theexascale world.
And for the last 14 years, I'vebeen focused more on AI and
quantum.
And quantum is really thebiggest portfolio right now.
(03:15):
Um, so I want to solve realchemistry problems in my case,
but there is very interestingproblems for the Department of
Energy to be solved in a verybroad set of um fundamental
science problems.
Uh think (03:27):
can we build a thousand-mile battery?
Can we build a solar cell thatrecovers 80% of the energy
instead of 20%?
Uh there is a lot ofopportunities, and that's the
energy part of the story.
SPEAKER_03 (03:42):
But there is a lot of material science, really.
SPEAKER_01 (03:45):
Material science, that's a very good, but also
chemistry.
Um, everybody that has worked inquantum uh has heard the story
of FOMOCO, which is uh thesystem, the the natural system
that can convert nitrogen andhydrogen into ammonium.
That takes about three percentof the world's energy right now.
(04:06):
Well, if we can replace theso-called Fritz Haver process um
that is used in industry, veryenergy consumption high energy
consumption, if we can replacethat with a natural version of
it, so to speak, or mimickednatural version, because nature
(04:26):
knows how to do it, then wecould potentially save a lot of
energy.
But that's not a trivialsimulation.
So um that's why people havebeen using that as one of the
benchmarks.
Um, that's a very complicatedsystem uh that needs to be
solved, and uh uh we will need avery significant size quantum
computer for that.
SPEAKER_03 (04:45):
So Cyprian's talked about this quite a bit.
He's the one that brought thismolecule to my attention years
ago.
Um, and he's mentioned it in thepast.
Do you have an idea?
I mean, uh we have an idea ofhow big of a quantum computer in
terms of logical bits we'd needfor Schore's algorithm to impact
RSA 2048.
We're looking at five to fifteenthousand logical qubits.
(05:07):
Do you have an idea of how bigof a quantum of a logical
quantum qubits you'd need inorder to do that that molecule?
Is it thousands?
Is it millions?
SPEAKER_01 (05:17):
So the interesting part is it's not always driven
by um the size, but rather thefidelity.
So let's say uh I can doprobably simulate the system
with a couple of hundred logicalqubits.
But unlike what you said, uh wewant to solve Schor's algorithm,
(05:39):
we need to kind of have 10 tothe minus 16, 10 to the minus 18
fidelity on the operations to beable to get a reliable result.
We don't need that as much forthese chemistry simulations.
We are very happy if we get afinal answer that is 10 to the
minus 1, because that is enoughof an accuracy for us to provide
an understanding and get anunderstanding of how the process
(06:02):
works.
So we have always targeted, andand my center is really focused
on targeting uh systems that aremaybe 10 to the minus seven, 10
to the minus eight fidelity peroperation.
Um so if I can get a 200 logicalqubits at 10 to the minus eight,
that means I can do about amillion to 10 million operations
(06:26):
uh before that result gets worsethan 10 to the minus one.
I can do a lot.
SPEAKER_02 (06:32):
Yeah.
SPEAKER_01 (06:33):
The key part with ki with the FOMOCO system, for
example, the chemistry uhmolecule is everybody has to be
focused on let's get the groundstate, get to get to the lowest
energy state of that system.
Completely irrelevant.
It's a very dynamic process, soyou actually need to start to
simulate the dynamics.
(06:53):
And this is actually reallyinteresting because when we say
we want to take a quantumcomputer, a quantum system to
mimic a dynamical evolution of aquantum system, that's exactly
what Feynman had in mind.
Um so it's very Feynman-stylethinking of simulations uh and
computing.
SPEAKER_00 (07:12):
So you you were saying that you're like very
interested in solving chemistryproblems.
Uh, for our audience, like whatwould be your like top three
problems that you would love tosee being solved by computers in
the near future?
SPEAKER_01 (07:29):
So this is the in chemistry specifically.
SPEAKER_00 (07:32):
In chemistry, yeah, yeah, yeah.
SPEAKER_01 (07:34):
Yeah, so that's the hard part.
The chemistry is such a broadfield, right?
Um, so this is why I brought upcan we actually build a battery
that is where the chargetransfer is more easy so that we
can, and the transformation thathave to happen in the battery
(07:54):
are not as extreme so that theylast longer and we can get more
miles out of a single battery.
Uh that to me is a veryimportant one.
I again I'm in California, so wewe like to think of green energy
to some extent, right?
SPEAKER_03 (08:08):
Uh and gas prices aren't hurting that that
thinking either.
SPEAKER_01 (08:12):
Uh yeah, there is some some gas stations here that
are reaching seven or eightdollars right now.
So, yes, it's a little bit of achallenging situation.
Uh but in general, um, there isa lot of problems here too.
Let think of fresh water.
This is something that is amajor problem in the nation uh
going forward.
Uh less snow, less fresh water.
(08:33):
So we need to extract it fromseawater.
That's very energy intensive.
So um, can we use solar?
Recover energy, recover energyat a low energy gradient is is
something that you need to haveorganic uh uh organic materials
to actually extract the energyback out of something that has
(08:57):
been boiling.
It's so you need to actuallyhave uh thermoelectrics that are
of a very low energy gradient,and that is something that we
cannot really get.
We have a very low percentage ofreturn on that.
It's like we can't recover 10%of the energy.
We need to recover 80% of theenergy.
SPEAKER_03 (09:15):
Exactly.
SPEAKER_01 (09:15):
So being able to model that is again, it's a very
dynamical process, uh, a veryFeynman-style process.
And so impossible to do inclassical computers.
And so we need to do that.
SPEAKER_03 (09:29):
Forget about the fact that efficiency is the is
better than invention, becauseyou already it's a burden hand
if you can make it moreefficient.
You mentioned solar.
Uh perovskite seems to be thethe thing that's getting us to
40%, perhaps by layering.
Uh is that the future, or arethere other things that we might
be on uh thinking about that'llget us to 80%?
SPEAKER_01 (09:51):
Who knows, right?
So so yeah, my center is focusedon doing some of the chemical
dynamics, but also on exactlythat.
Can we look at the high energy,uh sorry, the high TC uh
superconductors, right?
Exactly like that is related tothese perovskites.
Can we actually reallyunderstand how these systems
(10:14):
work?
Again, we have models onclassical computers, we have
some understanding, but they arenot of a level that are
predictive, so that we couldactually do the simulation and
design on a quantum computer andthen build.
SPEAKER_03 (10:29):
When you say TC, I do you mean temperature High
temperature, yeah.
Okay, so basically somethingthat might even work at at
ambient temperatures with theTrevor Burrus, Jr.
SPEAKER_01 (10:37):
That's I have a hard time believing.
SPEAKER_03 (10:41):
You don't think that's possible at the moment?
SPEAKER_01 (10:43):
I would have a hard time believing that because if
that would be feasible, I woulduh I would assume nature has
already done it.
Okay.
They have had what?
Um billions of years of ofevolution, and then they must
have figured it out somehow.
SPEAKER_03 (10:59):
Okay, I hear you.
SPEAKER_01 (11:01):
So but there is a lot of uh uh opportunities along
that way, along the way to makethat happen.
SPEAKER_00 (11:07):
So speaking about uh you mentioned the topic of
quantum advantage, and I knowthat's a topic that sparked a
lot of debate and back andforth.
Where do you think we are umright now in terms of uh being
to able to prove actual quantumadvantage?
(11:29):
Because we've seen it in thepast, right?
And it was kind of like a backand forth.
So I would love to hear yourtake on where are we now as of
2026 with with this 2026?
SPEAKER_01 (11:44):
I would say in the next couple of years, we
actually will demonstrate anadvantage in the sense that we
can do simulations that are notfeasible in classical computers
anymore.
I I don't think we're that faraway.
You see that in a lot ofcompanies uh and and research
institutions right now.
They are actually working onsimulations, demonstrating
(12:05):
simulations that are gettingclose.
Um we're actually working onsomething that uh we're trying
to prove actually might be atrue scientific advantage using
a quantum computer.
Um that and that's whereactually your statement actually
becomes very interesting, right?
So um you have that dynamicwhere the quantum computing side
(12:30):
will say, hey, we have anadvantage.
And then uh the classicalcomputing side comes back and
says, No, we know how to fixthis now.
We learn from you.
But the interesting part is thatis a good dynamic.
Uh, it keeps us honest uh in thequantum computing world, um, but
it also actually advances theclassical computing side too.
(12:51):
Yeah there is a lot ofquantum-inspired algorithms that
have come been developed as aresult of what we learned of
using a quantum computer.
And that means we can actuallyget more bang for a buck out of
a classical computer.
So honestly, to me, that is awin on its own.
Uh that dynamic, that that notfight, but kind of that that
tension between classical andquantum, pushing each other
(13:13):
exactly, driving this forward,and and jointly get to be able
to get to points where we cansimulate things we couldn't do a
couple of years ago, eitherclassically or quantumly.
And the reality is, in the longrun, I would expect the
classical computers and thequantum computers to be very
closely working together.
(13:34):
A quantum computer is not goodat everything.
No.
It has its quantum nature, ithas powers uh that can be
harnessed for specific parts ofsimulations.
There is other parts of thesimulations that are way better
on classical computers.
Classical computers are good atadding one plus one, a quantum
computers not.
Um so you need to kind of startlooking at how we can use the
(13:57):
strength of both.
And that's what you start to seethese days is that um now that
the quantum computers get ascale, there is that need to
integrate with HPC.
Um and that's in many differentways.
Uh we have simulations that arejointly run on quantum and
classical.
The the key components that areneed to be run on quantum
(14:19):
computers are run on a quantumcomputer, and vice versa.
Um and the reality is quantumcomputing needs to rely on HPC
going forward.
Um we need to go to errorcorrection.
The only way to get errorcorrection and and know where to
correct is to analyze the data.
How do you do that?
You need to use classicalcomputers.
(14:41):
Um we are going to be relying onclassical computers for a long
time.
Quantum is not just plainlyreplacing everything.
SPEAKER_00 (14:50):
Well, I think the the perfect example, right, is
we've been mentioning on theshow many, many times Schorr's
algorithm, right, and the theproblem of uh factorization.
Quantum actually solves a veryspecific part of that, like the
order-finding part.
Um and and the rest of it isthis is classical.
SPEAKER_01 (15:09):
So um, yeah, exactly.
And a lot of you actually haveto take the data from that you
want to actually go and and anduh and push through Schure's
algorithm.
You have to transform it firstanyway.
You have to get it into a shapethat you can feed into a quantum
computer, yeah, do the in the uhrun the simulation on a quantum
(15:31):
computer, get the data out, andhave to do pro-processing again.
So that's that synergy is goingto go and stay there for the
long run.
And and honestly, in the future,what I see is an uh a more it's
not a complete black box, butkind of a an ecosystem where you
have quantum computers,classical computers, and AI
(15:53):
working together in any way,shape, or form um to advance the
knowledge that they can produce.
SPEAKER_03 (16:02):
I the way I like to think about it is or explain it
to people is it's like you'retrying to peel the paint off of
something.
It's hard to get that firstlittle corner.
That's where quantum gives yousomething.
It does something that's notpossible to do otherwise, but
you still need the normalclassical to do the the heavy
lifting.
And that that shores kind oftaught us that, and as did uh
(16:24):
Grover's as well.
SPEAKER_01 (16:26):
Yeah.
Actually, sometimes it might bethe other way around.
The classical might give usenough of appeal than the rest
of the quantum and get it.
I've got actually converting alot of information that we have
in the classical space onto aquantum computer is is a very
challenging problem.
SPEAKER_03 (16:42):
I assume AI is going to help us quite a bit with
that.
That AI will will that the twowill help each other move
forward in a way that uh I meanI'm starting to see articles
about quantum memory, which uhCyprian you know pointed out
very long ago was was a problem.
And and the the advancements inerror correction, um at least
(17:04):
the buzz about it in the lastfew weeks has been quite big.
I uh Quantinium came out with aum a system they're saying that
they can do a two to one logicalto physical qubits.
I I haven't dug into the papersor anything, but even if it's
ten times worse than that,that's still better than I
expected as far as errorcorrection goes.
SPEAKER_01 (17:23):
Oh yeah.
There is a very I would say thelast couple of years have been
very, very exciting in in thequantum computing ecosystem in
general.
It's like, one, we are nowgetting systems that are at
scale.
Um two, we are actually figuringout how we can harness the
errors one way or another, andthat is error correction.
(17:45):
And we have seen literally inthe last year, so many
demonstrations where they haveat least a break-even, are
starting to see advantage ofusing error correction.
And this is also the first timethat we can actually start to
work on these kinds of things,because error correction
requires enough qubits.
Right.
So, yeah, you can develop allthe theories you want, you don't
(18:08):
know how they actually are goingto work in practice.
And this is the cool part again,is that now that we have the
systems at scale, you see moreand more practitioners push the
boundaries on what errorcorrection can do, test error
correction, and as a result,start to figure out new ways of
doing error correction.
And so, yes, the quantitium is agreat example.
(18:30):
Um, quantinium and uh theneutral atom companies have a
little bit of an advantage.
They all use atoms, and atomsare all the same.
So they have um a little biteasier than, for example, the
superconducting qubit guys thatwhere every qubit is different.
Um so a lot of the uh importantthings where we want to go to
(18:52):
with error correction is thingslike transversal operations,
transversal gates.
Um, that is very hard to doclassic uh with superconducting
qubits, but it's very easy to dowith neutral atoms and trap
diodes.
SPEAKER_03 (19:05):
So do you have a modality that that you think is
serving it?
Well, I I want to put this in ain the right way without shading
it too much.
There are multiple modalities tochoose from.
Um it in the beginning there wassuperconducting qubits, and that
seemed to be the whole field.
Microsoft went off and decidedto try something that could
promise a very rapid expansionif they can get it right, but
(19:29):
they're starting from furtherback.
And then we've got the theneutral atoms and the trapped
ion people with their littlelasers.
Those modalities, they I don'tknow that I would ever fear that
there's going to be a winner.
We've said there'd probably bemultiple modalities based on
what you're trying to do.
Is there one that's right nowlooking more promising for the
(19:49):
energy solutions that you'relooking for, the chemistry
solutions you're looking for?
Or is are you guys just stillplaying the field?
SPEAKER_01 (19:56):
So my center has taken a very unique approach
there.
So um uh the first we have beenfor the last five years, and
we'll continue to do that in thenext five years, for we have
focused on neutral atoms, traptions, and superconducting
qubits.
SPEAKER_03 (20:12):
So no, you haven't made a choice, really.
SPEAKER_01 (20:14):
We haven't made a choice, but actually you're
exactly right.
What we're seeing is thatdifferent technologies might
have different advantages in thedifferent science domains of the
Department of Energy.
SPEAKER_03 (20:27):
Could we could could we have a single quantum
computer with qubits of allthree types?
Possible.
SPEAKER_01 (20:33):
Uh so uh you could do that with atoms, but it's
it's superconducting qubitsbecause of the difference in
speed, but also the type offrequencies that they work on.
Uh um there's a lot of moredistributed computing.
SPEAKER_03 (20:46):
You have one of one type doing one kind of
operation.
Uh we we've we've talked aboutuh networking.
There's uh Cisco came on andtalked about their their
ability, they're trying tonetwork quantum computers.
So you could take 1050 logicalqubit quantum computers and have
a 500 quantum computer, uh uhcubic quantum computer.
Um maybe we'll see that whereyou could network together
(21:08):
disparate topologies or mmodalities, but but uh you
you're probably right.
It's the they're not gonna be inthe same box.
SPEAKER_01 (21:16):
Uh and that's fine, right?
So um we're doing a little bitof that in our center too.
Uh we have been coupling uhtrapped ions uh systems
together.
Um we at Berkeley uh at BerkeleyLab, we actually also have a
team uh that is focusedspecifically on how we can
connect quantum computerstogether.
So they are actually doing itbetween um NV centers and and um
(21:41):
and trapped ions.
But there's a reason for that.
They're in the same frequencyregime, they can actually go in
the telecom bands uh rallyrelatively easily, and so it's
easy to communicate.
But if you have to take uminformation, let's say you have
a quantum computer.
That is neutral atom and thenone that's superconducting
(22:02):
qubits.
And you need to kind of transferquantum information from the
neutral atoms to thesuperconducting qubits, there is
a very big change in frequenciesthat have to be done.
So that that conversion processis actually something that is
very hard to do, and people arestarting to still figure out.
(22:23):
Although there's always peoplethat are creative and see if
they can actually put themphysically together and just
based on space and separation,maybe you have them interact
with.
SPEAKER_03 (23:27):
So we know them quite well.
Yeah.
SPEAKER_01 (23:30):
And that's the I think the key part.
Our center is not a, ah, let'sdo fundamental science, we'll
see where it goes.
We are actually looking atunderstanding the technology
holes, the technology advancesthat need to have a have to
happen, uh, and do thefundamental science to the
engineering to make that happen.
(23:50):
And a great example is, forexample, first five years, we've
been working a lot on integratedphotonics.
This is done uh uh with ourpartners at Cindilla and uh and
some work also in Colorado.
That technology has reached apoint where we actually have
chips that are desired by theNeutral Atoms and Ions
(24:10):
companies.
So now not only Quera, but alsoAtom Computing, uh INQ and uh um
and quantinium are actually nowpartners in the center.
SPEAKER_03 (24:22):
Oh wow.
SPEAKER_01 (24:23):
And I think this is the a real powerful part of a
center like ours.
This is we are closely workingwith industry.
And this is what DUE wants.
This is also, it wants to see atech transfer.
Um kind of us de-risking some ofthese companies because
integrated photonics issomething that is on the roadmap
(24:45):
for companies, but further out.
They have to deliver two VCsright now, um, in the next
couple of years.
Um, so if we can actually uhde-risk them by actually
providing some technologiesgoing forward, then they can
take advantage of it.
They they can deliver theirproducts in the future more more
efficiently and and morecost-effectively.
So uh this is really a veryclose partnership that uh our
(25:09):
center has with a lot of thesekind of companies.
Um and we're even doing it inthe superconducting qubit world.
Uh so um uh we are we'repartnering with Colab.
Um so Colab uh is founded byJohn Martinez, our Nobel Prize
winner.
So I'm I'm always excited tosay, yeah, we have a Nobel Prize
winner.
SPEAKER_03 (25:29):
It's nice to have those in the in the back room,
right?
SPEAKER_01 (25:32):
But he is really thinking about how to scale
these quantum computers on uhwith superconducting qubits and
understands a lot of thechallenges.
And this is where uh thenational labs, these centers can
actually help a company, helpthe industry accelerate.
Um, so we're actually going touse uh an entity called the
(25:52):
Molecular Foundry, which is auser facility here in Berkeley
Lab.
Um we will have do a lot ofrapid testing of chip designs
and understanding where theerrors come from so that we can
actually help them build thatnext generation technology.
And that's again a partnershipthat has to happen.
SPEAKER_03 (26:11):
That's very exciting.
Yeah, and and I mean that's beensince the end of World War II,
that or since before World WarII, really, that's been a big
hallmark of private-publiccooperation.
You know, DARPA goes and figuresout what the military needs, the
DOE figures out what the energyneeds are, and energy needs uh
are really going to be top ofmind for the next decade.
SPEAKER_01 (26:34):
Oh, for sure.
unknown (26:35):
Yeah.
SPEAKER_00 (26:36):
One of the things, in addition to the actual
building, right, and the errorcorrecting, one of the things
that we've been struggling forever since the quantum computing
was proposed as a computingmodel, is correctly embedding
the problems or defining theproblems, the real-world
problems, right, uh, for uhquantum, because it's it's
completely different,fundamentally different from
(26:58):
what we used to do forclassical.
So I'd like to hear your opinionon what kind of progress are we
seeing on that front?
Are we getting better of framingthe problems, uh kind of
defining them in a way that theycan benefit from these
advancements?
Or we're still at the statewhere we are really struggling
with that.
SPEAKER_01 (27:20):
For I would say the Department of Energy problems,
we definitely seeing thattransition into we know what we
need to simulate, but justtrying to make it more and more
efficient.
Okay.
Um, so my center, uh I'll talk alot about the technology that
our centers focus on, but wehave uh uh a very significant
effort actually in uh doing theapplication side.
(27:43):
So we have research in high-NGphysics, in, as I said, the
chemical dynamics, and then themany body physics, which is
really um these uh high TCsuperconductor type uh problems.
They all have a commonality, andthat is the dynamical nature of
these problems.
And we're actually believingthat in the next couple of years
(28:06):
we can demonstrate an advantagethere.
So we understand what needs tobe done, and we also understand
the problem we want to run andthe outcome, because that
outcome is something thatcommunity will be very excited
about.
And that's how I drive itanyway.
I want to, we're not just goingto run a simulation and say, oh,
(28:28):
we can run it at something thatthat could not be done
classically.
We see that a lot.
Now we're starting to see moreand more people understanding
this is a problem we want tosolve, and if we get an answer,
we have some new insight.
SPEAKER_02 (28:41):
Right.
SPEAKER_01 (28:41):
And that's where we want to go, and that's where our
center wants to go.
It's not just advancing thetechnology in thousand decks,
but also starting to demonstratethat using this quantum
computer, we got some newscientific insight, we
discovered something new, uh,because that's really what we
are about.
That's what we should be about.
SPEAKER_03 (29:02):
Solving real problems, yeah.
SPEAKER_01 (29:03):
Real problems.
Now, so that is for theDepartment of Energy.
I think there is uh a broaderquestion of how this could how
quantum technology is going tobe brought be utilized or uh
given advantage to industry atlarge, right?
So I would say there is someopportunities.
Uh so the optimization communityhas demonstrated there is
(29:25):
definitely an advantage there.
Uh the hard part is encoding theproblem into a uh a good
representation that you can runon a quantum computer
efficiently.
But I think that community isgetting there.
Um and we know that because thebanks have figured those kind of
things out already.
Uh JP Morgan Chase is doing alot in that area.
(29:46):
Uh and then another area that wenow start to see progress in is
more the kind of CFD typecommunity, right?
Um nonlinear PDEs that areparticle different partial
differential equations uh thatare kind of solved by companies
(30:06):
like Boeing, like car companies.
Uh that's a nascent area, butpeople are getting really into
trying to understand how theycan encode that kind of
simulation, that kind of problemon a quantum computer.
And we see slowly more and morepapers and more ideas and and
approaches that are beingdeveloped there.
(30:28):
And it's again, that's it'sbecause we are now at scale and
we can actually start to tacklethese problems.
SPEAKER_03 (30:33):
Aaron Powell That's felt like the domain of um of
the uh annealing computers likeD-Wave for a while.
The the the if if I'm thinkingif I'm understanding what you're
talking about, the optimizationproblems and things like that.
Um but even they are looking atdoing uh classical uh quantum uh
universal gate quantum computingas well.
Yes.
Um so maybe you know we got someadvantage from the annealing
(30:56):
being uh available while wewaited for the universal gates
to mature.
And so so maybe that's that'sthe phase that we're coming
into.
SPEAKER_01 (31:06):
I agree.
And then so the hard theinteresting challenge
opportunity neutral atoms areactually very good at doing
optimization problems too.
We've we've started todemonstrate that now.
They're naturally a collectionof atoms that are going to go
for the most optimalconfiguration, the lowest energy
(31:27):
configuration, which is what youwant to do for an optimizer.
Right.
So there is that natural way todo these kind of things.
Um, it's not really then adigital simulation, right?
So, and that's that's theinteresting part that we're also
looking at a lot.
It's like a lot of theseproblems don't need to be
digital, they could be an analogsimulation.
(31:49):
Um, Feynman's idea was an analogsimulation, not a digital
simulation.
Um Shore came and said, oh, wecan do digital operations.
Uh that's how the community islike, oh, we can do digital
cooperations.
Oh, we can do something thatlooks like a classical computer.
Right.
SPEAKER_00 (32:05):
So it's it's that's just fabulous how we're going
full circle, right?
We essentially simplified ourworld model by introducing the
digitalization, because that'sessentially sampling of reality,
right?
Now we're kind of going backfull circle and we're starting
to look at at analog again justbecause we have better
(32:26):
technology.
SPEAKER_01 (32:27):
Yeah, so digital is an interesting thing to think
about, right?
As we had all our problems if inour community, any of these
equations fundamentally are ananalog equation, right?
But we had to find a way tocompute it.
So early on, we digitized or wediscretized, which automatically
(32:50):
meant digitizing.
Um, and that's how we'veevolved.
We've done this for what, 50, 60years now?
SPEAKER_00 (32:57):
Yeah.
SPEAKER_01 (32:58):
Um now the quantum computing comes around and says,
well, it might not be asclear-cut.
Maybe some of these things needto be revisited.
And and I always push my mystudents, my postdocs also, it's
like, think creatively.
This is your basic equation.
How would you solve it when youhave things like entanglement,
(33:19):
um, uh evolution, naturalevolution of quantum systems, um
and and and uh uh all thequantum features that you have
in your system, how would yousolve this problem differently?
Um so some of them have come upwith the idea of simulations
that are more of the analognature, some are mixtures, um,
(33:41):
because hey, it might be easierto do something in a digital
version uh versus an analogversion.
And so this opportunity of usingdifferent the modality in
different ways is is a uniqueway, but it also changes the
thinking and has to change thethinking of how we solve
problems.
SPEAKER_00 (33:59):
Yeah, that's actually one of the things that
we've been talking also a lot isum how much does it count in the
bias of being basically educatedand uh growing in the world of
classical computing when tryingto have this completely
(34:20):
different mindset.
And we're talking a lot aboutwhat we call the quantum-born
generation, who are youngerfolks who have been basically uh
developing their professionalcareers in a world where quantum
is a thing, it's not just somekind of theory.
SPEAKER_01 (34:39):
As I said, I come from the IHBC world for a lot
for at least 15 years.
Uh and the first time I startedto work on quantum problems,
it's the same mode.
This is how I did it on aclassical computer.
Let's just translate.
SPEAKER_03 (34:56):
Yeah.
SPEAKER_01 (34:57):
And see how how we can simply translate it and run
it.
That's not the most efficientway to do it.
And so we're starting to seemore and more people realizing
you cannot just translate, youhave to rethink the problem.
And that's where I think thenext couple of years will be
interesting when people see newways to use the quantum
(35:18):
technology that's there andactually make significant gains.
SPEAKER_03 (35:22):
So we're coming close to the end of our time
slot, but I did have one otherthing to add to that, and that
is back about 10 years ago at aconference Supreme and I were
at, One Qubit, a company in uhVancouver, north of you, um were
talking about the fact thatbefore there were quantum
computers, they were getting an8% increase in performance by
(35:43):
just taking most algorithms andreimagining them in quantum
space and using simulators.
And um and since then they'veyou know they've taken that to
the next level.
And we've had other guests whowho worry more about the quantum
simulation and uh the quantumimagining, giving them an
advantage even on simulations.
Do you, as a chemist, I I studychemistry, I was a chemistry
(36:05):
major for to the first two yearsof college and then switched.
Um so I I I uh I I chickenedout, I guess you would say.
Um but would you there's beensome talk in the past, mostly in
passing, about a biologicalcomputing model or a chemical
computing model.
I I don't know what that wouldlook like in in um, but I also
(36:26):
had trouble imagining what aquantum computing model would
look like before.
Do you think there's gonna beother competitors in those
spaces, especially as someonewith a chemistry background?
SPEAKER_01 (36:38):
I think there's a lot of creative ways that uh uh
once we really understand how toharness the quantum information,
yeah, there might be differentmodalities.
And the reality is one of theearly demonstrations of a
quantum computer was actuallyusing NMR, nucleomagnetic
resonance of a molecule.
They encoded the informationinto the vibrational modes of a
(37:01):
molecule.
So, yes, early on they have donethat.
Uh interestingly, in my centerwe have the neutral atoms, but
we also are working on neutralmolecules or even charged
molecules.
Um why?
Because now we have fermionicdegrees of freedom, which is
what the atoms are, but we alsohave what's called bosonic
(37:22):
degrees of freedom, which is thecoupling between these atoms.
Uh and we can take advantage ofthat.
Um that's what we're going to dowith the trapped ion quantum
computers too.
It's like we're not going to puthundreds, we actually have build
a trap that can hold 200 ions.
Uh but um we're now much moreinterested in can we actually uh
take 50 trapped ions and couplethe modes between them.
(37:47):
Um so instead of having two tothe n, so two to the fifty,
could we actually do six to thefifty?
That's a computational powerthat's different.
It's again, it's a different wayof thinking of okay, we have a
qubit that is two states.
Well, maybe each bit has sixstates or ten states.
SPEAKER_03 (38:08):
Oh wow.
SPEAKER_01 (38:09):
And that changes the dynamic.
And uh it's actually interestingbecause the first time uh the
superconducting qubit communitystarted to talk about Q-threads,
which means they have threelevels, the high energy physics
community kind came around andsaid, that's exactly what we
need, and we can now encode theinformation a lot more
efficiently than we could dobefore.
SPEAKER_03 (38:30):
So that would be like taking a classical bit and
changing it from a coin to asix-sided dice.
Yeah.
It would explode the memory,explode the possible
permutations.
SPEAKER_01 (38:40):
Yes.
But the way they have to thinkabout encoding their
information, they just don'thave a spin that spin up and
down.
They also have a flavor or theyhave a color.
SPEAKER_02 (38:49):
Yeah.
SPEAKER_01 (38:50):
They want to encode that kind of an information that
allows them to actually do thatusing more levels in a quantum
system.
And so that's those are the kindof things that we are exploring.
It's the not just let's justscale.
We are working on scaling,believe me.
We're going to we're actuallygoing to build a neutral atom
(39:11):
system that is going to have 100logical qubits, maybe a little
more, at 10 to the minus six, 10to the minus 7 in the next two,
three years.
Wow.
So, yes, we are going to drivesome of this.
Um, they all have like 40,000atoms.
We can do a lot more with that.
If you go and think of thehigh-NG physicists, they they
(39:32):
are like, oh, give me 40,000atoms.
I can do some other things withthat too.
SPEAKER_02 (39:36):
Right.
SPEAKER_01 (39:36):
So there's a lot of these interesting dynamics that
are that are starting to happenwhere you get the application
people, they now have access tosystems, they are asking for new
features within these systems tomake them even more flexible,
uh, to have more opportunitiesto kind of mimic the physics
(39:56):
that they want to do onto thehardware that they have.
Um, instead of the other wayaround, which we have been doing
in classical computing.
Hey, we have bits, we have um uhspecific operations, just match
the problem to whatever thehardware is.
There is now still that dynamicwhere could we adjust the
hardware potentially to solvethe problem in a more efficient
(40:18):
way?
And that's that's I think thereal cool dynamic that is
happening right now in thequantum ecosystem, in the
quantum computing ecosystem,where people are trying to
figure out where thatrelationship is.
SPEAKER_03 (40:31):
Well, unless we're gonna make this a two-hour
episode, I think we have to callit at some point.
But this has been an amazingconversation, Bert.
I really hope you're willing tocome back on a more on a regular
basis and talk to us becausethis has been very enlightening.
SPEAKER_00 (40:45):
Thank you.
SPEAKER_03 (40:46):
Thank you very much.
SPEAKER_00 (40:47):
It's been a real pleasure.
SPEAKER_03 (40:48):
So thanks everybody and join us again next time uh
on Entangle Things.
SPEAKER_01 (40:53):
Thank you.
SPEAKER_02 (41:28):
Visit PulsarSecurity.com and let's
secure your digital futuretogether.
In episode 139, Patrick Insecret speaker with
senior scientists at theLawrence Berkeley National
Laboratory, Bert De Young, todiscuss quantum advantage for
scientific discovery, chemistryand material science, and the
pursuit of the right modality.
Welcome to Entangled Things,your quantum computing podcast,
(00:34):
hosted by Patrick and Cyprien.
SPEAKER_03 (00:40):
Hey Cyprian, how are you doing?
SPEAKER_00 (00:42):
Hey Patrick.
I'm doing great.
Looking forward for anotherepisode of Entangle Things.
SPEAKER_03 (00:46):
Well, this is a good one today.
So, Bert, do you mindintroducing yourself to our
audience?
SPEAKER_01 (00:52):
Sure.
Well, Patrick and Brittany,thank you for having me here.
So I'm Bert DeYoung.
I'm a senior scientist here atLawrence Berkeley National Lab.
So one of the uh 17 Departmentof Energy national labs.
Um what drives me with quantumis so right now I lead the
Quantum System Accelerator,which is one of the five
(01:12):
national quantum initiativecenters.
Um that's a large investment bythe Department of Energy focused
on uh on delivering kind of thenext generation quantum
technologies and not just thetechnologies, but also starting
to demonstrate that there is aquantum advantage or a
usefulness uh for scientificdiscovery using quantum
(01:34):
computing specifically.
That is our focus.
Uh so the five centers, in caseyou uh you're not familiar with
that, uh the National QuantumInitiative is something that uh
uh Congress has been uh umputting in place for the last
five years.
They are reauthorising it rightnow, um, which really lays out
(01:55):
the drive for the US to be aleader in quantum technologies.
Um so these five centers are theDepartment of Energy's effort to
really show advance thetechnology and show the
opportunities and uh the uhchallenges and how we can solve
those uh in the next, well, forus five years.
(02:15):
We just got renewed as fivecenters.
SPEAKER_03 (02:17):
Excellent.
SPEAKER_01 (02:18):
Um our undersecretary Dario Gill
announced that late November.
Um, so we have been running umto really start this process and
and really making significantadvances.
SPEAKER_03 (02:31):
That's amazing.
SPEAKER_01 (02:32):
Um so I can tell you a little bit about my center.
Please.
Uh my center has beendeveloping, um we have a very
broad view of how we're going toapproach this.
Uh so first of all, I shouldhave said maybe I'm a
computational chemist bytraining, so I'm not a quantum
physicist.
Um I come from a uh a backgroundwhere I want to solve scientific
(02:53):
problems.
So these kind of what you heardat the beginning, we want to
solve problems, this is reallywhat drives anything that I do.
So I lived in a in in for a longtime in the high performance
computing world, so in theexascale world.
And for the last 14 years, I'vebeen focused more on AI and
quantum.
And quantum is really thebiggest portfolio right now.
(03:15):
Um, so I want to solve realchemistry problems in my case,
but there is very interestingproblems for the Department of
Energy to be solved in a verybroad set of um fundamental
science problems.
Uh think (03:27):
can we build a thousand-mile battery?
Can we build a solar cell thatrecovers 80% of the energy
instead of 20%?
Uh there is a lot ofopportunities, and that's the
energy part of the story.
SPEAKER_03 (03:42):
But there is a lot of material science, really.
SPEAKER_01 (03:45):
Material science, that's a very good, but also
chemistry.
Um, everybody that has worked inquantum uh has heard the story
of FOMOCO, which is uh thesystem, the the natural system
that can convert nitrogen andhydrogen into ammonium.
That takes about three percentof the world's energy right now.
(04:06):
Well, if we can replace theso-called Fritz Haver process um
that is used in industry, veryenergy consumption high energy
consumption, if we can replacethat with a natural version of
it, so to speak, or mimickednatural version, because nature
(04:26):
knows how to do it, then wecould potentially save a lot of
energy.
But that's not a trivialsimulation.
So um that's why people havebeen using that as one of the
benchmarks.
Um, that's a very complicatedsystem uh that needs to be
solved, and uh uh we will need avery significant size quantum
computer for that.
SPEAKER_03 (04:45):
So Cyprian's talked about this quite a bit.
He's the one that brought thismolecule to my attention years
ago.
Um, and he's mentioned it in thepast.
Do you have an idea?
I mean, uh we have an idea ofhow big of a quantum computer in
terms of logical bits we'd needfor Schore's algorithm to impact
RSA 2048.
We're looking at five to fifteenthousand logical qubits.
(05:07):
Do you have an idea of how bigof a quantum of a logical
quantum qubits you'd need inorder to do that that molecule?
Is it thousands?
Is it millions?
SPEAKER_01 (05:17):
So the interesting part is it's not always driven
by um the size, but rather thefidelity.
So let's say uh I can doprobably simulate the system
with a couple of hundred logicalqubits.
But unlike what you said, uh wewant to solve Schor's algorithm,
(05:39):
we need to kind of have 10 tothe minus 16, 10 to the minus 18
fidelity on the operations to beable to get a reliable result.
We don't need that as much forthese chemistry simulations.
We are very happy if we get afinal answer that is 10 to the
minus 1, because that is enoughof an accuracy for us to provide
an understanding and get anunderstanding of how the process
(06:02):
works.
So we have always targeted, andand my center is really focused
on targeting uh systems that aremaybe 10 to the minus seven, 10
to the minus eight fidelity peroperation.
Um so if I can get a 200 logicalqubits at 10 to the minus eight,
that means I can do about amillion to 10 million operations
(06:26):
uh before that result gets worsethan 10 to the minus one.
I can do a lot.
SPEAKER_02 (06:32):
Yeah.
SPEAKER_01 (06:33):
The key part with ki with the FOMOCO system, for
example, the chemistry uhmolecule is everybody has to be
focused on let's get the groundstate, get to get to the lowest
energy state of that system.
Completely irrelevant.
It's a very dynamic process, soyou actually need to start to
simulate the dynamics.
(06:53):
And this is actually reallyinteresting because when we say
we want to take a quantumcomputer, a quantum system to
mimic a dynamical evolution of aquantum system, that's exactly
what Feynman had in mind.
Um so it's very Feynman-stylethinking of simulations uh and
computing.
SPEAKER_00 (07:12):
So you you were saying that you're like very
interested in solving chemistryproblems.
Uh, for our audience, like whatwould be your like top three
problems that you would love tosee being solved by computers in
the near future?
SPEAKER_01 (07:29):
So this is the in chemistry specifically.
SPEAKER_00 (07:32):
In chemistry, yeah, yeah, yeah.
SPEAKER_01 (07:34):
Yeah, so that's the hard part.
The chemistry is such a broadfield, right?
Um, so this is why I brought upcan we actually build a battery
that is where the chargetransfer is more easy so that we
can, and the transformation thathave to happen in the battery
(07:54):
are not as extreme so that theylast longer and we can get more
miles out of a single battery.
Uh that to me is a veryimportant one.
I again I'm in California, so wewe like to think of green energy
to some extent, right?
SPEAKER_03 (08:08):
Uh and gas prices aren't hurting that that
thinking either.
SPEAKER_01 (08:12):
Uh yeah, there is some some gas stations here that
are reaching seven or eightdollars right now.
So, yes, it's a little bit of achallenging situation.
Uh but in general, um, there isa lot of problems here too.
Let think of fresh water.
This is something that is amajor problem in the nation uh
going forward.
Uh less snow, less fresh water.
(08:33):
So we need to extract it fromseawater.
That's very energy intensive.
So um, can we use solar?
Recover energy, recover energyat a low energy gradient is is
something that you need to haveorganic uh uh organic materials
to actually extract the energyback out of something that has
(08:57):
been boiling.
It's so you need to actuallyhave uh thermoelectrics that are
of a very low energy gradient,and that is something that we
cannot really get.
We have a very low percentage ofreturn on that.
It's like we can't recover 10%of the energy.
We need to recover 80% of theenergy.
SPEAKER_03 (09:15):
Exactly.
SPEAKER_01 (09:15):
So being able to model that is again, it's a very
dynamical process, uh, a veryFeynman-style process.
And so impossible to do inclassical computers.
And so we need to do that.
SPEAKER_03 (09:29):
Forget about the fact that efficiency is the is
better than invention, becauseyou already it's a burden hand
if you can make it moreefficient.
You mentioned solar.
Uh perovskite seems to be thethe thing that's getting us to
40%, perhaps by layering.
Uh is that the future, or arethere other things that we might
be on uh thinking about that'llget us to 80%?
SPEAKER_01 (09:51):
Who knows, right?
So so yeah, my center is focusedon doing some of the chemical
dynamics, but also on exactlythat.
Can we look at the high energy,uh sorry, the high TC uh
superconductors, right?
Exactly like that is related tothese perovskites.
Can we actually reallyunderstand how these systems
(10:14):
work?
Again, we have models onclassical computers, we have
some understanding, but they arenot of a level that are
predictive, so that we couldactually do the simulation and
design on a quantum computer andthen build.
SPEAKER_03 (10:29):
When you say TC, I do you mean temperature High
temperature, yeah.
Okay, so basically somethingthat might even work at at
ambient temperatures with theTrevor Burrus, Jr.
SPEAKER_01 (10:37):
That's I have a hard time believing.
SPEAKER_03 (10:41):
You don't think that's possible at the moment?
SPEAKER_01 (10:43):
I would have a hard time believing that because if
that would be feasible, I woulduh I would assume nature has
already done it.
Okay.
They have had what?
Um billions of years of ofevolution, and then they must
have figured it out somehow.
SPEAKER_03 (10:59):
Okay, I hear you.
SPEAKER_01 (11:01):
So but there is a lot of uh uh opportunities along
that way, along the way to makethat happen.
SPEAKER_00 (11:07):
So speaking about uh you mentioned the topic of
quantum advantage, and I knowthat's a topic that sparked a
lot of debate and back andforth.
Where do you think we are umright now in terms of uh being
to able to prove actual quantumadvantage?
(11:29):
Because we've seen it in thepast, right?
And it was kind of like a backand forth.
So I would love to hear yourtake on where are we now as of
2026 with with this 2026?
SPEAKER_01 (11:44):
I would say in the next couple of years, we
actually will demonstrate anadvantage in the sense that we
can do simulations that are notfeasible in classical computers
anymore.
I I don't think we're that faraway.
You see that in a lot ofcompanies uh and and research
institutions right now.
They are actually working onsimulations, demonstrating
(12:05):
simulations that are gettingclose.
Um we're actually working onsomething that uh we're trying
to prove actually might be atrue scientific advantage using
a quantum computer.
Um that and that's whereactually your statement actually
becomes very interesting, right?
So um you have that dynamicwhere the quantum computing side
(12:30):
will say, hey, we have anadvantage.
And then uh the classicalcomputing side comes back and
says, No, we know how to fixthis now.
We learn from you.
But the interesting part is thatis a good dynamic.
Uh, it keeps us honest uh in thequantum computing world, um, but
it also actually advances theclassical computing side too.
(12:51):
Yeah there is a lot ofquantum-inspired algorithms that
have come been developed as aresult of what we learned of
using a quantum computer.
And that means we can actuallyget more bang for a buck out of
a classical computer.
So honestly, to me, that is awin on its own.
Uh that dynamic, that that notfight, but kind of that that
tension between classical andquantum, pushing each other
(13:13):
exactly, driving this forward,and and jointly get to be able
to get to points where we cansimulate things we couldn't do a
couple of years ago, eitherclassically or quantumly.
And the reality is, in the longrun, I would expect the
classical computers and thequantum computers to be very
closely working together.
(13:34):
A quantum computer is not goodat everything.
No.
It has its quantum nature, ithas powers uh that can be
harnessed for specific parts ofsimulations.
There is other parts of thesimulations that are way better
on classical computers.
Classical computers are good atadding one plus one, a quantum
computers not.
Um so you need to kind of startlooking at how we can use the
(13:57):
strength of both.
And that's what you start to seethese days is that um now that
the quantum computers get ascale, there is that need to
integrate with HPC.
Um and that's in many differentways.
Uh we have simulations that arejointly run on quantum and
classical.
The the key components that areneed to be run on quantum
(14:19):
computers are run on a quantumcomputer, and vice versa.
Um and the reality is quantumcomputing needs to rely on HPC
going forward.
Um we need to go to errorcorrection.
The only way to get errorcorrection and and know where to
correct is to analyze the data.
How do you do that?
You need to use classicalcomputers.
(14:41):
Um we are going to be relying onclassical computers for a long
time.
Quantum is not just plainlyreplacing everything.
SPEAKER_00 (14:50):
Well, I think the the perfect example, right, is
we've been mentioning on theshow many, many times Schorr's
algorithm, right, and the theproblem of uh factorization.
Quantum actually solves a veryspecific part of that, like the
order-finding part.
Um and and the rest of it isthis is classical.
SPEAKER_01 (15:09):
So um, yeah, exactly.
And a lot of you actually haveto take the data from that you
want to actually go and and anduh and push through Schure's
algorithm.
You have to transform it firstanyway.
You have to get it into a shapethat you can feed into a quantum
computer, yeah, do the in the uhrun the simulation on a quantum
(15:31):
computer, get the data out, andhave to do pro-processing again.
So that's that synergy is goingto go and stay there for the
long run.
And and honestly, in the future,what I see is an uh a more it's
not a complete black box, butkind of a an ecosystem where you
have quantum computers,classical computers, and AI
(15:53):
working together in any way,shape, or form um to advance the
knowledge that they can produce.
SPEAKER_03 (16:02):
I the way I like to think about it is or explain it
to people is it's like you'retrying to peel the paint off of
something.
It's hard to get that firstlittle corner.
That's where quantum gives yousomething.
It does something that's notpossible to do otherwise, but
you still need the normalclassical to do the the heavy
lifting.
And that that shores kind oftaught us that, and as did uh
(16:24):
Grover's as well.
SPEAKER_01 (16:26):
Yeah.
Actually, sometimes it might bethe other way around.
The classical might give usenough of appeal than the rest
of the quantum and get it.
I've got actually converting alot of information that we have
in the classical space onto aquantum computer is is a very
challenging problem.
SPEAKER_03 (16:42):
I assume AI is going to help us quite a bit with
that.
That AI will will that the twowill help each other move
forward in a way that uh I meanI'm starting to see articles
about quantum memory, which uhCyprian you know pointed out
very long ago was was a problem.
And and the the advancements inerror correction, um at least
(17:04):
the buzz about it in the lastfew weeks has been quite big.
I uh Quantinium came out with aum a system they're saying that
they can do a two to one logicalto physical qubits.
I I haven't dug into the papersor anything, but even if it's
ten times worse than that,that's still better than I
expected as far as errorcorrection goes.
SPEAKER_01 (17:23):
Oh yeah.
There is a very I would say thelast couple of years have been
very, very exciting in in thequantum computing ecosystem in
general.
It's like, one, we are nowgetting systems that are at
scale.
Um two, we are actually figuringout how we can harness the
errors one way or another, andthat is error correction.
(17:45):
And we have seen literally inthe last year, so many
demonstrations where they haveat least a break-even, are
starting to see advantage ofusing error correction.
And this is also the first timethat we can actually start to
work on these kinds of things,because error correction
requires enough qubits.
Right.
So, yeah, you can develop allthe theories you want, you don't
(18:08):
know how they actually are goingto work in practice.
And this is the cool part again,is that now that we have the
systems at scale, you see moreand more practitioners push the
boundaries on what errorcorrection can do, test error
correction, and as a result,start to figure out new ways of
doing error correction.
And so, yes, the quantitium is agreat example.
(18:30):
Um, quantinium and uh theneutral atom companies have a
little bit of an advantage.
They all use atoms, and atomsare all the same.
So they have um a little biteasier than, for example, the
superconducting qubit guys thatwhere every qubit is different.
Um so a lot of the uh importantthings where we want to go to
(18:52):
with error correction is thingslike transversal operations,
transversal gates.
Um, that is very hard to doclassic uh with superconducting
qubits, but it's very easy to dowith neutral atoms and trap
diodes.
SPEAKER_03 (19:05):
So do you have a modality that that you think is
serving it?
Well, I I want to put this in ain the right way without shading
it too much.
There are multiple modalities tochoose from.
Um it in the beginning there wassuperconducting qubits, and that
seemed to be the whole field.
Microsoft went off and decidedto try something that could
promise a very rapid expansionif they can get it right, but
(19:29):
they're starting from furtherback.
And then we've got the theneutral atoms and the trapped
ion people with their littlelasers.
Those modalities, they I don'tknow that I would ever fear that
there's going to be a winner.
We've said there'd probably bemultiple modalities based on
what you're trying to do.
Is there one that's right nowlooking more promising for the
(19:49):
energy solutions that you'relooking for, the chemistry
solutions you're looking for?
Or is are you guys just stillplaying the field?
SPEAKER_01 (19:56):
So my center has taken a very unique approach
there.
So um uh the first we have beenfor the last five years, and
we'll continue to do that in thenext five years, for we have
focused on neutral atoms, traptions, and superconducting
qubits.
SPEAKER_03 (20:12):
So no, you haven't made a choice, really.
SPEAKER_01 (20:14):
We haven't made a choice, but actually you're
exactly right.
What we're seeing is thatdifferent technologies might
have different advantages in thedifferent science domains of the
Department of Energy.
SPEAKER_03 (20:27):
Could we could could we have a single quantum
computer with qubits of allthree types?
Possible.
SPEAKER_01 (20:33):
Uh so uh you could do that with atoms, but it's
it's superconducting qubitsbecause of the difference in
speed, but also the type offrequencies that they work on.
Uh um there's a lot of moredistributed computing.
SPEAKER_03 (20:46):
You have one of one type doing one kind of
operation.
Uh we we've we've talked aboutuh networking.
There's uh Cisco came on andtalked about their their
ability, they're trying tonetwork quantum computers.
So you could take 1050 logicalqubit quantum computers and have
a 500 quantum computer, uh uhcubic quantum computer.
Um maybe we'll see that whereyou could network together
(21:08):
disparate topologies or mmodalities, but but uh you
you're probably right.
It's the they're not gonna be inthe same box.
SPEAKER_01 (21:16):
Uh and that's fine, right?
So um we're doing a little bitof that in our center too.
Uh we have been coupling uhtrapped ions uh systems
together.
Um we at Berkeley uh at BerkeleyLab, we actually also have a
team uh that is focusedspecifically on how we can
connect quantum computerstogether.
So they are actually doing itbetween um NV centers and and um
(21:41):
and trapped ions.
But there's a reason for that.
They're in the same frequencyregime, they can actually go in
the telecom bands uh rallyrelatively easily, and so it's
easy to communicate.
But if you have to take uminformation, let's say you have
a quantum computer.
That is neutral atom and thenone that's superconducting
(22:02):
qubits.
And you need to kind of transferquantum information from the
neutral atoms to thesuperconducting qubits, there is
a very big change in frequenciesthat have to be done.
So that that conversion processis actually something that is
very hard to do, and people arestarting to still figure out.
(22:23):
Although there's always peoplethat are creative and see if
they can actually put themphysically together and just
based on space and separation,maybe you have them interact
with.
SPEAKER_03 (23:27):
So we know them quite well.
Yeah.
SPEAKER_01 (23:30):
And that's the I think the key part.
Our center is not a, ah, let'sdo fundamental science, we'll
see where it goes.
We are actually looking atunderstanding the technology
holes, the technology advancesthat need to have a have to
happen, uh, and do thefundamental science to the
engineering to make that happen.
(23:50):
And a great example is, forexample, first five years, we've
been working a lot on integratedphotonics.
This is done uh uh with ourpartners at Cindilla and uh and
some work also in Colorado.
That technology has reached apoint where we actually have
chips that are desired by theNeutral Atoms and Ions
(24:10):
companies.
So now not only Quera, but alsoAtom Computing, uh INQ and uh um
and quantinium are actually nowpartners in the center.
SPEAKER_03 (24:22):
Oh wow.
SPEAKER_01 (24:23):
And I think this is the a real powerful part of a
center like ours.
This is we are closely workingwith industry.
And this is what DUE wants.
This is also, it wants to see atech transfer.
Um kind of us de-risking some ofthese companies because
integrated photonics issomething that is on the roadmap
(24:45):
for companies, but further out.
They have to deliver two VCsright now, um, in the next
couple of years.
Um, so if we can actually uhde-risk them by actually
providing some technologiesgoing forward, then they can
take advantage of it.
They they can deliver theirproducts in the future more more
efficiently and and morecost-effectively.
So uh this is really a veryclose partnership that uh our
(25:09):
center has with a lot of thesekind of companies.
Um and we're even doing it inthe superconducting qubit world.
Uh so um uh we are we'repartnering with Colab.
Um so Colab uh is founded byJohn Martinez, our Nobel Prize
winner.
So I'm I'm always excited tosay, yeah, we have a Nobel Prize
winner.
SPEAKER_03 (25:29):
It's nice to have those in the in the back room,
right?
SPEAKER_01 (25:32):
But he is really thinking about how to scale
these quantum computers on uhwith superconducting qubits and
understands a lot of thechallenges.
And this is where uh thenational labs, these centers can
actually help a company, helpthe industry accelerate.
Um, so we're actually going touse uh an entity called the
(25:52):
Molecular Foundry, which is auser facility here in Berkeley
Lab.
Um we will have do a lot ofrapid testing of chip designs
and understanding where theerrors come from so that we can
actually help them build thatnext generation technology.
And that's again a partnershipthat has to happen.
SPEAKER_03 (26:11):
That's very exciting.
Yeah, and and I mean that's beensince the end of World War II,
that or since before World WarII, really, that's been a big
hallmark of private-publiccooperation.
You know, DARPA goes and figuresout what the military needs, the
DOE figures out what the energyneeds are, and energy needs uh
are really going to be top ofmind for the next decade.
SPEAKER_01 (26:34):
Oh, for sure.
unknown (26:35):
Yeah.
SPEAKER_00 (26:36):
One of the things, in addition to the actual
building, right, and the errorcorrecting, one of the things
that we've been struggling forever since the quantum computing
was proposed as a computingmodel, is correctly embedding
the problems or defining theproblems, the real-world
problems, right, uh, for uhquantum, because it's it's
completely different,fundamentally different from
(26:58):
what we used to do forclassical.
So I'd like to hear your opinionon what kind of progress are we
seeing on that front?
Are we getting better of framingthe problems, uh kind of
defining them in a way that theycan benefit from these
advancements?
Or we're still at the statewhere we are really struggling
with that.
SPEAKER_01 (27:20):
For I would say the Department of Energy problems,
we definitely seeing thattransition into we know what we
need to simulate, but justtrying to make it more and more
efficient.
Okay.
Um, so my center, uh I'll talk alot about the technology that
our centers focus on, but wehave uh uh a very significant
effort actually in uh doing theapplication side.
(27:43):
So we have research in high-NGphysics, in, as I said, the
chemical dynamics, and then themany body physics, which is
really um these uh high TCsuperconductor type uh problems.
They all have a commonality, andthat is the dynamical nature of
these problems.
And we're actually believingthat in the next couple of years
(28:06):
we can demonstrate an advantagethere.
So we understand what needs tobe done, and we also understand
the problem we want to run andthe outcome, because that
outcome is something thatcommunity will be very excited
about.
And that's how I drive itanyway.
I want to, we're not just goingto run a simulation and say, oh,
(28:28):
we can run it at something thatthat could not be done
classically.
We see that a lot.
Now we're starting to see moreand more people understanding
this is a problem we want tosolve, and if we get an answer,
we have some new insight.
SPEAKER_02 (28:41):
Right.
SPEAKER_01 (28:41):
And that's where we want to go, and that's where our
center wants to go.
It's not just advancing thetechnology in thousand decks,
but also starting to demonstratethat using this quantum
computer, we got some newscientific insight, we
discovered something new, uh,because that's really what we
are about.
That's what we should be about.
SPEAKER_03 (29:02):
Solving real problems, yeah.
SPEAKER_01 (29:03):
Real problems.
Now, so that is for theDepartment of Energy.
I think there is uh a broaderquestion of how this could how
quantum technology is going tobe brought be utilized or uh
given advantage to industry atlarge, right?
So I would say there is someopportunities.
Uh so the optimization communityhas demonstrated there is
(29:25):
definitely an advantage there.
Uh the hard part is encoding theproblem into a uh a good
representation that you can runon a quantum computer
efficiently.
But I think that community isgetting there.
Um and we know that because thebanks have figured those kind of
things out already.
Uh JP Morgan Chase is doing alot in that area.
(29:46):
Uh and then another area that wenow start to see progress in is
more the kind of CFD typecommunity, right?
Um nonlinear PDEs that areparticle different partial
differential equations uh thatare kind of solved by companies
(30:06):
like Boeing, like car companies.
Uh that's a nascent area, butpeople are getting really into
trying to understand how theycan encode that kind of
simulation, that kind of problemon a quantum computer.
And we see slowly more and morepapers and more ideas and and
approaches that are beingdeveloped there.
(30:28):
And it's again, that's it'sbecause we are now at scale and
we can actually start to tacklethese problems.
SPEAKER_03 (30:33):
Aaron Powell That's felt like the domain of um of
the uh annealing computers likeD-Wave for a while.
The the the if if I'm thinkingif I'm understanding what you're
talking about, the optimizationproblems and things like that.
Um but even they are looking atdoing uh classical uh quantum uh
universal gate quantum computingas well.
Yes.
Um so maybe you know we got someadvantage from the annealing
(30:56):
being uh available while wewaited for the universal gates
to mature.
And so so maybe that's that'sthe phase that we're coming
into.
SPEAKER_01 (31:06):
I agree.
And then so the hard theinteresting challenge
opportunity neutral atoms areactually very good at doing
optimization problems too.
We've we've started todemonstrate that now.
They're naturally a collectionof atoms that are going to go
for the most optimalconfiguration, the lowest energy
(31:27):
configuration, which is what youwant to do for an optimizer.
Right.
So there is that natural way todo these kind of things.
Um, it's not really then adigital simulation, right?
So, and that's that's theinteresting part that we're also
looking at a lot.
It's like a lot of theseproblems don't need to be
digital, they could be an analogsimulation.
(31:49):
Um, Feynman's idea was an analogsimulation, not a digital
simulation.
Um Shore came and said, oh, wecan do digital operations.
Uh that's how the community islike, oh, we can do digital
cooperations.
Oh, we can do something thatlooks like a classical computer.
Right.
SPEAKER_00 (32:05):
So it's it's that's just fabulous how we're going
full circle, right?
We essentially simplified ourworld model by introducing the
digitalization, because that'sessentially sampling of reality,
right?
Now we're kind of going backfull circle and we're starting
to look at at analog again justbecause we have better
(32:26):
technology.
SPEAKER_01 (32:27):
Yeah, so digital is an interesting thing to think
about, right?
As we had all our problems if inour community, any of these
equations fundamentally are ananalog equation, right?
But we had to find a way tocompute it.
So early on, we digitized or wediscretized, which automatically
(32:50):
meant digitizing.
Um, and that's how we'veevolved.
We've done this for what, 50, 60years now?
SPEAKER_00 (32:57):
Yeah.
SPEAKER_01 (32:58):
Um now the quantum computing comes around and says,
well, it might not be asclear-cut.
Maybe some of these things needto be revisited.
And and I always push my mystudents, my postdocs also, it's
like, think creatively.
This is your basic equation.
How would you solve it when youhave things like entanglement,
(33:19):
um, uh evolution, naturalevolution of quantum systems, um
and and and uh uh all thequantum features that you have
in your system, how would yousolve this problem differently?
Um so some of them have come upwith the idea of simulations
that are more of the analognature, some are mixtures, um,
(33:41):
because hey, it might be easierto do something in a digital
version uh versus an analogversion.
And so this opportunity of usingdifferent the modality in
different ways is is a uniqueway, but it also changes the
thinking and has to change thethinking of how we solve
problems.
SPEAKER_00 (33:59):
Yeah, that's actually one of the things that
we've been talking also a lot isum how much does it count in the
bias of being basically educatedand uh growing in the world of
classical computing when tryingto have this completely
(34:20):
different mindset.
And we're talking a lot aboutwhat we call the quantum-born
generation, who are youngerfolks who have been basically uh
developing their professionalcareers in a world where quantum
is a thing, it's not just somekind of theory.
SPEAKER_01 (34:39):
As I said, I come from the IHBC world for a lot
for at least 15 years.
Uh and the first time I startedto work on quantum problems,
it's the same mode.
This is how I did it on aclassical computer.
Let's just translate.
SPEAKER_03 (34:56):
Yeah.
SPEAKER_01 (34:57):
And see how how we can simply translate it and run
it.
That's not the most efficientway to do it.
And so we're starting to seemore and more people realizing
you cannot just translate, youhave to rethink the problem.
And that's where I think thenext couple of years will be
interesting when people see newways to use the quantum
(35:18):
technology that's there andactually make significant gains.
SPEAKER_03 (35:22):
So we're coming close to the end of our time
slot, but I did have one otherthing to add to that, and that
is back about 10 years ago at aconference Supreme and I were
at, One Qubit, a company in uhVancouver, north of you, um were
talking about the fact thatbefore there were quantum
computers, they were getting an8% increase in performance by
(35:43):
just taking most algorithms andreimagining them in quantum
space and using simulators.
And um and since then they'veyou know they've taken that to
the next level.
And we've had other guests whowho worry more about the quantum
simulation and uh the quantumimagining, giving them an
advantage even on simulations.
Do you, as a chemist, I I studychemistry, I was a chemistry
(36:05):
major for to the first two yearsof college and then switched.
Um so I I I uh I I chickenedout, I guess you would say.
Um but would you there's beensome talk in the past, mostly in
passing, about a biologicalcomputing model or a chemical
computing model.
I I don't know what that wouldlook like in in um, but I also
(36:26):
had trouble imagining what aquantum computing model would
look like before.
Do you think there's gonna beother competitors in those
spaces, especially as someonewith a chemistry background?
SPEAKER_01 (36:38):
I think there's a lot of creative ways that uh uh
once we really understand how toharness the quantum information,
yeah, there might be differentmodalities.
And the reality is one of theearly demonstrations of a
quantum computer was actuallyusing NMR, nucleomagnetic
resonance of a molecule.
They encoded the informationinto the vibrational modes of a
(37:01):
molecule.
So, yes, early on they have donethat.
Uh interestingly, in my centerwe have the neutral atoms, but
we also are working on neutralmolecules or even charged
molecules.
Um why?
Because now we have fermionicdegrees of freedom, which is
what the atoms are, but we alsohave what's called bosonic
(37:22):
degrees of freedom, which is thecoupling between these atoms.
Uh and we can take advantage ofthat.
Um that's what we're going to dowith the trapped ion quantum
computers too.
It's like we're not going to puthundreds, we actually have build
a trap that can hold 200 ions.
Uh but um we're now much moreinterested in can we actually uh
take 50 trapped ions and couplethe modes between them.
(37:47):
Um so instead of having two tothe n, so two to the fifty,
could we actually do six to thefifty?
That's a computational powerthat's different.
It's again, it's a different wayof thinking of okay, we have a
qubit that is two states.
Well, maybe each bit has sixstates or ten states.
SPEAKER_03 (38:08):
Oh wow.
SPEAKER_01 (38:09):
And that changes the dynamic.
And uh it's actually interestingbecause the first time uh the
superconducting qubit communitystarted to talk about Q-threads,
which means they have threelevels, the high energy physics
community kind came around andsaid, that's exactly what we
need, and we can now encode theinformation a lot more
efficiently than we could dobefore.
SPEAKER_03 (38:30):
So that would be like taking a classical bit and
changing it from a coin to asix-sided dice.
Yeah.
It would explode the memory,explode the possible
permutations.
SPEAKER_01 (38:40):
Yes.
But the way they have to thinkabout encoding their
information, they just don'thave a spin that spin up and
down.
They also have a flavor or theyhave a color.
SPEAKER_02 (38:49):
Yeah.
SPEAKER_01 (38:50):
They want to encode that kind of an information that
allows them to actually do thatusing more levels in a quantum
system.
And so that's those are the kindof things that we are exploring.
It's the not just let's justscale.
We are working on scaling,believe me.
We're going to we're actuallygoing to build a neutral atom
(39:11):
system that is going to have 100logical qubits, maybe a little
more, at 10 to the minus six, 10to the minus 7 in the next two,
three years.
Wow.
So, yes, we are going to drivesome of this.
Um, they all have like 40,000atoms.
We can do a lot more with that.
If you go and think of thehigh-NG physicists, they they
(39:32):
are like, oh, give me 40,000atoms.
I can do some other things withthat too.
SPEAKER_02 (39:36):
Right.
SPEAKER_01 (39:36):
So there's a lot of these interesting dynamics that
are that are starting to happenwhere you get the application
people, they now have access tosystems, they are asking for new
features within these systems tomake them even more flexible,
uh, to have more opportunitiesto kind of mimic the physics
(39:56):
that they want to do onto thehardware that they have.
Um, instead of the other wayaround, which we have been doing
in classical computing.
Hey, we have bits, we have um uhspecific operations, just match
the problem to whatever thehardware is.
There is now still that dynamicwhere could we adjust the
hardware potentially to solvethe problem in a more efficient
(40:18):
way?
And that's that's I think thereal cool dynamic that is
happening right now in thequantum ecosystem, in the
quantum computing ecosystem,where people are trying to
figure out where thatrelationship is.
SPEAKER_03 (40:31):
Well, unless we're gonna make this a two-hour
episode, I think we have to callit at some point.
But this has been an amazingconversation, Bert.
I really hope you're willing tocome back on a more on a regular
basis and talk to us becausethis has been very enlightening.
SPEAKER_00 (40:45):
Thank you.
SPEAKER_03 (40:46):
Thank you very much.
SPEAKER_00 (40:47):
It's been a real pleasure.
SPEAKER_03 (40:48):
So thanks everybody and join us again next time uh
on Entangle Things.
SPEAKER_01 (40:53):
Thank you.
SPEAKER_02 (41:28):
Visit PulsarSecurity.com and let's
secure your digital futuretogether.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
Betrayal Weekly
Betrayal Weekly is back for a new season. Every Thursday, Betrayal Weekly shares first-hand accounts of broken trust, shocking deceptions, and the trail of destruction they leave behind. Hosted by Andrea Gunning, this weekly ongoing series digs into real-life stories of betrayal and the aftermath. From stories of double lives to dark discoveries, these are cautionary tales and accounts of resilience against all odds. From the producers of the critically acclaimed Betrayal series, Betrayal Weekly drops new episodes every Thursday. If you would like to share your story, you can reach out to the Betrayal Team by emailing them at betrayalpod@gmail.com and follow us on Instagram at @betrayalpod and @glasspodcasts. Please join our Substack for additional exclusive content, curated book recommendations, and community discussions. Sign up FREE by clicking this link Beyond Betrayal Substack. Join our community dedicated to truth, resilience, and healing. Your voice matters! Be a part of our Betrayal journey on Substack.