June 9, 2026 • 39 mins
In Episode 142, Yuping Huang, CEO and Chairman of Quantum Computing Inc and Physics Professor at Stevens Institute of Technology, joins Patrick and Ciprian to make the case for photonics, not just as a modality, but as a philosophy. QCI's approach starts with the physics rather than mapping quantum onto classical architecture, leading to machines that look nothing like what most people expect a quantum computer to look like. Yuping walks through the deceptively hard problem of deterministically generating entangled photon pairs, why you don't need a million entangled photons to build something useful, and why QCI operates entirely at room temperature, if it can't fit in a backpack, it won't end up in users' hands. The conversation closes on Neural Wave, a hybrid photonic-digital system that offloads computation into the optical domain and cuts AI energy consumption by orders of magnitude.
Listen
Watch
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
SPEAKER_02 (00:40):
Hey Sifrien, how are you doing today?
Hey Patrick, I'm doing good,looking forward for another
great episode of UntangledThings.
SPEAKER_04 (00:47):
I don't know how it happens, but our guests keep
getting better.
So today we're joined by Yuping.
You Ping, can you pleaseintroduce yourself to our
audience?
SPEAKER_01 (00:54):
Hey, uh Patrick and Sypia and hey everybody.
You Ping Huang here.
I'm the uh CEO and uh thechairman of uh Quantum Computing
Inc., a public company.
But uh I'm really uh also aphysics professor, so we're
gonna have to be here and uh toshare some of my thoughts and
(01:16):
maybe to learn something fromPatrick and see if he can't.
SPEAKER_04 (01:21):
Well, I mean, as a as a professor, you're in the
right place at the right time.
Quantum's exploding.
There's so much buzz right nowabout things getting uh real.
And you guys are are heavilyinvolved, especially on the
photonic side, which I justdon't see a future where
photonics doesn't play a role.
Other modalities areinteresting, and I'm sure
they're gonna play a role.
(01:42):
But but photonics seems the onlyway we're gonna connect this
stuff together.
SPEAKER_01 (01:47):
You know what, Patrick, is uh I feel very
pleased and honored to haveheard what you said, because uh
this has been what I have beentrying to tell everybody that uh
since uh over 15 years ago, thatuh hey guys, look at photonics
(02:09):
and uh photonics is ininherently quantum mechanical.
Of course, uh there's sometechnology and engineering
challenges to implement uh uhquantum tech in photonics
platform, but uh over the pastuh 10 years and we have seen
(02:34):
lots of progress on both fronts,and now I think um the
opportunity for photonics tomake big buzz and to have a
strong push on the quantumtechnology transport industry
(02:56):
right here.
So so I I I'm actually excitedby the progress that we have
been making at QCI and uh theprogress that uh people are
making uh in all different areasof photonics for quantum.
SPEAKER_02 (03:16):
The other thing that Patrick and I were discussing a
lot and were mentioning a lot,photonics also seems to be the
one and only bridge betweenquantum computing and quantum
communications, which isprobably kind of like the
cornerstone of the future of thefield, right?
(03:36):
So I think that's another veryimportant angle when we talk
about photonics.
SPEAKER_01 (03:41):
This is this is a hundred percent uh correct.
So the only way to connectquantum computers over distance
is by photonics.
And uh we have seen in thedigital era that we have been in
(04:02):
uh over the uh since uh um thelate 90s that uh so the power of
computing is really modifiedwhen you can connect many
computers together.
And uh I have no reason tobelieve that uh this is not true
(04:26):
for quantum.
In fact, uh so I believe that uhthe quantum future really um
hands on the success of quantuminternet as well, where we
connect many, many quantumdevices uh by internet, and the
only way to do it to the best ofmy knowledge as a physics
(04:53):
professor is by photonics.
So this is why um um that uh umprocessely as you just said that
uh well since we are going touse photonics to connect uh uh
quantum machines, quantumcomputers, the easiest way to
(05:17):
make it happen is to usephotonics to build quantum
machines, to build quantumcomputers, so that uh you don't
need to have this uhintermediate step of converting
um the information carrier of umsay from Adam uh to photonics or
(05:41):
vice versa.
But uh on the other hand, so I'mnot saying that uh um uh we have
to use uh photonics to buildquantum computers.
In fact, so I believe in theopposite.
I believe that uh the quantumspace is open, open not just in
(06:07):
the technology approach, butalso in this wide application
space.
I believe the future of quantumwill consist of uh say those
atom-based, uhsuperconductor-based, ion-based,
and photonic-based quantumcomputers, and each of them will
(06:32):
um take care of uh um some nicheapplications in its own areas
where uh it makes most of sensein terms of the cost of
operation, in terms of the speedto solution, for example.
SPEAKER_04 (06:52):
We've said that.
We we've discussed the fact thatthere's we don't there there's
no winning modality, thatthere's so many modalities, and
we may have others to discoverthat we haven't seen.
It seems like any binaryproperty at the quantum level
can be turned into a qubit ifyou if you engineer it right.
SPEAKER_01 (07:10):
Indeed.
Precisely.
Patrick, uh do you have aphysics PhD?
SPEAKER_04 (07:18):
I do not.
No.
I but I hang out with peoplelike you way too often, so it's
it ribs off.
SPEAKER_01 (07:24):
Impressed.
Yes, indeed.
So in fact, uh any these greatquantum state system can
potentially turn into aninformation processing machine.
And uh depending on whatapplications, it could be
computer, right?
It could be encoder, right?
(07:46):
It could be sensor.
SPEAKER_04 (07:49):
Amazing.
Uh I I do have one question thatmight might might veer on to the
professor side versus the CEOside of your uh personality,
which is we've we've seen somearticles in the past that said
there was a hypothesis thatmaybe instead of entangling uh
instead of measuring and andentangling that that you might
(08:11):
take, especially in photonics,entangle millions of photons and
measure along the way.
So as a way to basically,because it's it's the fragility
of the state that's our biggestchallenge.
And so, like normally you youtake and you you operate and
then you entangle and then youmeasure.
(08:32):
But but someone posited, and Ididn't really get enough about
the article, um, I didn't readthe underlying paper, um, that
maybe you could entangle a hugeamount of photons at creation
and have like this herd of themthat you could measure some of
them along the way to make sureyou're going the right
direction.
Is that something that that QCIis looking at?
(08:54):
Or are you more theold-fashioned way of like
manipulating entangled states asyou go?
Let me answer your question.
Uh and if that's an unfairquestion, I'm I'm happy to move
on because it's a weirdquestion.
SPEAKER_01 (09:11):
No, it is a a very good and a deep question.
So I can tell you this.
Um many physicists, includingmyself, have spent over 10 years
trying to find a way toefficiently create entangled two
(09:32):
photons.
Just two.
And it's already very hard.
And the big issue there is theuh standard way for us to create
entanglement so the processitself is uh render, is uh
stochastic.
(09:54):
Meaning that uh so you don'thave much control of when the
two entangled photons can becreated.
But uh say maybe on average umeach each time you try uh you
can you can say have uh uh uhsay eight eight percent five
(10:16):
percent probability of gettingthe entangled payout.
This is actually an issue forusing the photonics uh for
quantum computing because uhjust as you said, Patrick, for
many applications you do need toprepare uh many many photons
(10:42):
that are entangled.
And uh now if you only have uhfive or eight percent
probability of uh to create uhone pair, so the probability of
being able to get many many pairwould be very very slim.
So we don't like that.
(11:03):
And uh I actually myself spentuh geez studying studying from
2010, so really looked into thefundamental physics uh first and
the uh uh engineering aspect ofhow to make it happen.
(11:26):
So in fact, we find a way tomake it happen, although right
now, so there are still some uhengineering challenge uh
actually on the manufacturingside to make the device.
But hey, so we find a way toallow you to determine
(11:47):
deterministically createentangled photons so that uh
when you say that give me onepair of entangled photons, you
will get one pair.
SPEAKER_04 (12:02):
Yeah, we we assume it's that's a step that
everybody simplifies.
Since the first time we Ciprianand I started talking about,
well, you entangle a couplethings and then and we skip
that.
And so it's fascinating how howin-depth it is and how hard it
we we don't take it, we take itfor granted that it's it's not
as hard.
SPEAKER_01 (12:21):
And uh and so now uh so this is actually what's uh
what's happening now at QCI now.
Uh for uh for that to happen forthe deterministic generation of
entangled photons, uh you willneed to make very specific uh
(12:43):
photonic integrated circuitswith some crazy parameters for
you to hit.
And uh well, so uh since Iproposed this approach, say I
think uh there are a total ofeight pretty crazy parameters
(13:05):
for us to hit, and we have hitseven.
unknown (13:08):
Wow.
SPEAKER_01 (13:09):
We are trying to hit the eighth one uh in the QSAI
fab now.
So I tell you what, so I feelgood about it, but uh after have
pushed this like for for forover 10 years, I'm kind of
getting old, so I'm pushing mymy engineers that okay guys, I
(13:31):
told you the recipe, right?
You can just make it happen andthen come back.
SPEAKER_04 (13:36):
Oh I'll be I'll be on the beach.
Just let me know when it's done.
SPEAKER_01 (13:41):
So so so so the real requirement is that uh so you
will have to make this a veryhigh quality, what we call micro
ring optical cavity.
And uh in order to get the Q, uhthe Q factor of the cavity.
(14:02):
Okay, guys, stop me if I startedto talk too much into physics.
Um, that's fine.
We we like that.
We needed to we needed to cut umsome uh micron size bridge uh
(14:23):
reach waveguide on uh uh thinfew lithium niggas wafer with
the surface roughness like lessthan 300 picometer.
Wow.
It it is crazy, right?
(14:44):
Um but I can tell you that uhthat that's what we have uh we
have achieved last year, butthis is just the first step, and
there are multiple steps, and uhthe team's working hard on that.
While uh I I probably can findone one or two days uh to go to
(15:07):
the beach, as Patrick just did.
Um but but now getting back toyour original question, Patrick,
because you really are asked aastronomical question that can
we entangle one million photons?
(15:32):
I guess uh then after my longanswer, the short answer is that
yes in principle but yes inprinciple, but uh I don't see
how in my lifetime, butfortunately we don't need that.
(15:52):
We don't need it to entangle onemillion photons for us to build
a quantum computer usingphotonics.
As long as say we can make uhhundreds of thousands of uh
pairs of entangled photons wheretwo photons are entangled in
(16:16):
each pair, so we can alreadybuild some very useful and very
powerful quantum computers.
SPEAKER_04 (16:24):
That that makes perfect sense because I think
the context of the entangle amillion was the fact that there
was so much loss intransmission.
But you're you're talking aboutcalculation as well, you're
talking about gates.
Yeah, and and um so you guyshave built a universal quantum
computer gate-based um withphotonics.
SPEAKER_01 (16:45):
Uh no, so we have walked on this and we figured
out how to build, um, but uh sothat this work is still ongoing.
What we have now on shelf um uhfor customers uh to buy as of
now is a specialized uh quantumoptimization machine that
(17:11):
definitely utilizes quantumeffects, but it is not
gate-based.
SPEAKER_04 (17:17):
Um okay.
SPEAKER_01 (17:18):
And they are very good uh uh for solving a pretty
uh large portfolio ofoptimization problems, uh but it
is not a general-purpose quantumcomputers.
SPEAKER_04 (17:35):
Is that a goal down the line for you guys, or is
that not an area that you'regoing into?
SPEAKER_01 (17:40):
We are working on building the gate-based uh uh
quantum computers.
Yes, so this is our goal.
But um the approach that uh weare taking at QCI is that uh so
we don't want to define uh whatkind of quantum computers or
(18:03):
what quantum computers shouldlook like by ourselves.
In fact, uh so we go to thecustomers, we talk with them and
we understand what they reallyneed.
And we build quantum machines uhto meet their need.
(18:24):
In fact, uh very interestingly,so while physicists, including
myself, um were very fascinatedon building the gate-based uh uh
general purpose quantumcomputers, actually.
So many customers uh they thethe feedback that we got from
(18:48):
them uh um um has been that uhuh so do you have a quantum
machine that can solve myproblems now?
And what we find is that uh sofor many of those problems,
gate-based machine actually doesnot give us the edge, does not
(19:10):
give us the advantage.
It's that some specially uhspecial purposed machines can
solve their problems.
So this is how we get startedwith the quantum optimization
machine.
Cool.
SPEAKER_02 (19:27):
That's I think that's very interesting because
one of the themes that that uhoccurs quite a lot, and we
discuss about it quite a lot, isthe difficulty in essentially
embedding real-world problemsinto the space of quantum,
right?
Defining a problem in a way thatcan actually be addressed,
(19:50):
right, with a quantum device ora quantum computer or however we
want to name it.
And I think besides the specifictask of building these devices,
the theoretical problem ofdefining the problem in a way
that's suitable, I think is iskind of the other big big goal
(20:10):
here for everyone in the in thefield.
SPEAKER_01 (20:13):
And uh do you know the fundamental reason for such
difficulty for cus uh for peoplein in needs uh to formulate the
problem to run on quantumcomputers?
What is what is it?
This is because in many cases,say quantum computers uh have
(20:40):
been designed based on theunderstanding of scientists
criticists.
Right.
So so it is that okay, so forexample, just to use this
gate-based.
So so in fact, uh so there is norule set that uh so we have to
(21:01):
use the gate-based.
It was that okay, so when weopen a classical computer,
right?
So you will have the memory,right?
You have the processor, and uhthere that's a lot of
transistors, right?
So uh the easiest way to picturewhat uh a future quantum
(21:28):
computer would look like is thatokay, how about just uh we
replace all the parts of aclassical computer with a quanta
equivalent and to it's what weknow.
Yes.
If you wanted to build a newtechnology or build a new type
of hardware uh by harnessing atotally different uh physics,
(21:58):
right?
So you're starting.
point is should not be try tobox the new physics into the old
architecture.
You should uh go deep into thenew physics itself okay and then
try to find the most logical wayto build the hardware according
(22:23):
to the physics itself.
SPEAKER_04 (22:25):
This is why in every sci-fi movie all the aliens look
like humans.
They just have different likenose ridges and things like
that.
It's it's the lack ofimagination is what you're
getting at.
SPEAKER_01 (22:36):
Yes and uh and actually so I but that um um
actually so I have been I mean Ihad uh uh been struggled uh in
in the same puzzle and but thenum um since uh about 10 years
(22:58):
ago I started to tell myselfthat okay let's uh really try to
uh let quantum be quantum and uhtry to make friends with quantum
instead of trying to box thequantum in a classical
architecture so let's build uh aquantum machine that uh um
(23:25):
observe the unique property ofquantum physics itself and it's
and then see how we can leverageit.
How can we can leverage it.
So this is why we are buildingsome quantum machines at QCI
that is very different from whatanybody has done.
But hey so it works and uh uh wehave find that uh so they can
(23:50):
indeed solve um problems veryefficiently and uh give us
better solutions and uh I infact so when we first announced
this uh machine many people saidoh this is not a quantum
computer because in people'smind it is that for a quantum
(24:11):
computer you have to have thegate base you have to have the
quantum memory and we are doingsomething very different but my
definition of a quantum computeris that a machine that utilized
the quantum effects to solveproblems with advantage that is
(24:34):
unmatched by any classical andthat's working for you because
you guys just announced um uh anAI edge product I believe didn't
you yes uh this is a another uhpretty exciting development uh
in fact uh so uh we know that uhAI is here to change the world
(24:58):
uh for good or of or for bad Idon't want it to it's like the
internet it's gonna changethings but it's up to us the
jury's still out on that oneright and uh now so but uh one
thing that we know for sure isthat uh AI is going to consume
(25:20):
more and more electric power anduh it's not a sustainable mode
uh to uh further uh develop uhthe AI architecture by by
building uh nuclear power uhplants next to data centers I
(25:43):
don't like that idea and I don'tthink uh that idea can carry us
much further.
So instead so at QCI we'reasking ourselves can we offload
some complex and uh powerconsumption intense calculations
(26:06):
from the digital to the analogoptical domain because the
moment that we can do that wecan for example reduce the um
power consumption percalculation by many orders of
(26:28):
magnitude.
SPEAKER_02 (26:28):
In fact we have uh we have uh done some research in
the lab and we find that uh uhwe can use a single photon for
machine learning task and uh asingle photon has energy of 10
to the negative uh 19 joule soto just uh to uh put this uh in
(26:55):
the context uh project so uh thelight bulb uh in your room now
is probably emitting if you haveit on is probably emitting like
10 to the uh twenty two uh yes10 to the 21 photons per second
(27:18):
so yeah we just need uh almostno energy to to perform that
project particular ai so nowthis is why at qci we are
looking at uh photonics approachand by incorporating uh
photonics neural netlock uhneural network layer with the
(27:45):
digital layer so that we canhave this hybrid uh machine
architecture that uh can reducethe energy consumption by i mean
10 times i mean that's 100 timesthat's amazing i mean that's so
much bigger than what we've seenin other things so um i don't
(28:06):
know if you saw this in the newsa while back but google came out
with a way to compress memoryusage by uh the the leading
models and training by sixfoldthis is a much bigger savings
and so that kind of thing Ithink it was called turboquant
or something on that name uhfrom Google that came out memory
(28:27):
management and and it it itchanged the metrics of uh the
cost the the data centers theywere gonna build so this this
kind of thing could also have abig impact is it's it's
something you just announced uhin the last last month I believe
right yeah so we our first uhcommercial product was uh
released and uh and uh lastmonth and it's called neuron
(28:51):
wave and uh what it does is thatuh so inside the box uh so it is
a hybrid uh uh neural networkthat uh processfully does what I
said so a part of thecalculation is done still done
uh in digital so uh FPGA and apart of that is done in the
(29:15):
optics and uh by combining thetwo actually so we have already
seen significant uh um savingsin the power consumption but
this is is just a start yes welland and machine learning
underlies everything I I learnedthat from Cyprian right if it's
weren't for the machine learningwe wouldn't be where we are in
(29:37):
the in the LLMs right no but Ithink this this comes at the
perfect moment because whatwe're seeing there is the end of
the free tokens for everybodyparty um I think people are
starting to realize that thesehuge models that are being used
out there right are are likehuge energy consumers and moving
(30:02):
even kind of forward with theseit's just going to be not
sustainable.
I think uh a few months ago oreven perhaps a year there were a
few very interesting papers thatwere published that were using
the laws of thermodynamics toprove that there is not enough
energy on this planet to achieveuh like the Skynet level of
(30:24):
artificial intelligence becausethe curve is just exponentially
in in increasing so I think thetiming of your announcement is
is just great because people arestarting to realize that the the
party's over and especially theenergy party is is over and
there is like no unlimitedpossibility of moving forward
(30:45):
with what we have unlesscompanies like you come with
these amazing innovations wherewe literally scale it back down
orders of magnitude the energyconsumption.
SPEAKER_01 (31:01):
Yeah thank you so much I I think uh yeah so and uh
uh we are um we are uh strivingto contribute uh uh but uh uh
say I'm glad to say that uh umthere are um uh other players so
they are making the effort alongthe same lines so I'm actually
(31:24):
pretty optimistic that um um uhfor photonics uh uh will uh uh s
play more and more importantrole in ai so you know even a
couple of years ago Nvidia isreally looking at photonics to
(31:47):
move the data from GPO to GPOright but now we we are we
should be looking at photonicsnot just for transferring
information but also but alsofor computing processing the
information and uh I believethat uh uh there is a a very
(32:10):
bright and um exciting futurethere and one thing that uh I
believe everybody will benefitfrom is that uh the tokens could
uh become cheaper and cheaper aswe adopt more and more photonics
because the energy consumptionwill be will uh will drop down
(32:36):
significantly as we integratemore and more photonics into the
AI architecture.
SPEAKER_02 (32:42):
And we have to say something before we we stop
right which is criticalphotonics and if we talk about
in the context of energyphotonics has one like huge
advantage which is it operatesat room temperature.
SPEAKER_01 (32:57):
Yes that's yeah well so actually uh this is a uh not
a very um very uh interestingtopic that uh um uh if I may I
want to share some uh thoughtsthat I have had for a long time.
(33:17):
Please uh as a physics reresearcher actually so I feel
that uh um there's a uh majormisconcept that quantum can only
happen at room temperature so infact if you open any quantum
(33:40):
mechanics book there's no suchthing called a cryogenic or or
uh or cold or or closed systemin fact so we have seen so many
signs that uh quantum effects doexist at room temperature in
fact so quantum effects arealready with us in the in many
(34:05):
biological systems so and I I'mnot sure if if you guys heard
that uh actually so thephotosynthesis uh processes that
happen uh now in any green plantso there are quantum effects
(34:27):
there right so in fact quantumeffects help uh uh transport in
the the the the energy and tostore at the bottom of the uh
green leaves okay so quantumeffects already exist uh in
(34:47):
nature at a room temperature incomplex uh biological systems we
don't need to put the leaves ina cryogenic fridge in fact we'd
better not do that's right sothen the big question question
is that uh why do we need to goto cryogenic to build quantum
(35:12):
computers and and so nature ishas already told us clearly that
no you don't need to do that andthis is uh also uh a guiding
philosophy at qc i i've beentelling everybody that okay guys
I know that uh it's easier forus to observe some quantum
(35:35):
effects at cryogenic temperaturebut we don't do it at QCI
everything we do is roomtemperature why because we want
every technology that we areinnovating engineering and
turning that to manufacturingwill end in people's hands right
(35:59):
we don't need uh we don't wantany of our users to buy that big
cryogenic fridge in order to useQSI product instead amen they
should be able to buy a productand put in the backpack and just
use it as another say computingdevice communication device I I
(36:22):
really love the philosophy youhave of like letting the
technology letting the physicsguide how we use it and and uh
not try trying not to resort tocryogenics.
SPEAKER_04 (36:31):
I think we're out of time but I really hope you'll
come back and talk to us again.
SPEAKER_01 (36:35):
It's been a great conversation is there any
anything uh at that you stillwant to mention before we wrap
up uh I I just want to say thatI really appreciate the for the
this opportunity um uh toconnect uh with uh with uh you
and uh I hope that as uh myhonest uh uh opinion uh actually
(37:02):
so maybe as a physics uhprofessor and so so so uh uh can
be perceived uh as a scientificand and uh scientific insights
instead of uh um uh uh uh mycomments as the CEO because I
(37:26):
really shared uh what I havebeen thinking and uh what my
true belief is uh yeah I thinkwe'd be all better off if all
the CEOs were also physicsprofessors.
SPEAKER_00 (37:39):
I'm not sure about that thanks again for joining us
and uh and hopefully we'll seeyou again soon thanks again bye
thank you very much it's been apleasure all right bye everybody
see you next time thank youthank you bye bye cybercrime is
one of the biggest threats tobusinesses of all sizes and
industries with almost half amillion open cyber positions the
(38:01):
problem is compounded by thelack of available talent in the
marketplace at Pulsar Securityour elite team of highly
credentialed experts collaboratewith you to assess your current
defenses and develop solutionstailored to your specific needs.
With services ranging fromcybersecurity education to
advanced penetration testing andred teaming you can start
(38:22):
reducing your risks today visitpulsarsecurity.com and let's
secure your digital futuretogether
Hey Sifrien, how are you doing today?
Hey Patrick, I'm doing good,looking forward for another
great episode of UntangledThings.
SPEAKER_04 (00:47):
I don't know how it happens, but our guests keep
getting better.
So today we're joined by Yuping.
You Ping, can you pleaseintroduce yourself to our
audience?
SPEAKER_01 (00:54):
Hey, uh Patrick and Sypia and hey everybody.
You Ping Huang here.
I'm the uh CEO and uh thechairman of uh Quantum Computing
Inc., a public company.
But uh I'm really uh also aphysics professor, so we're
gonna have to be here and uh toshare some of my thoughts and
(01:16):
maybe to learn something fromPatrick and see if he can't.
SPEAKER_04 (01:21):
Well, I mean, as a as a professor, you're in the
right place at the right time.
Quantum's exploding.
There's so much buzz right nowabout things getting uh real.
And you guys are are heavilyinvolved, especially on the
photonic side, which I justdon't see a future where
photonics doesn't play a role.
Other modalities areinteresting, and I'm sure
they're gonna play a role.
(01:42):
But but photonics seems the onlyway we're gonna connect this
stuff together.
SPEAKER_01 (01:47):
You know what, Patrick, is uh I feel very
pleased and honored to haveheard what you said, because uh
this has been what I have beentrying to tell everybody that uh
since uh over 15 years ago, thatuh hey guys, look at photonics
(02:09):
and uh photonics is ininherently quantum mechanical.
Of course, uh there's sometechnology and engineering
challenges to implement uh uhquantum tech in photonics
platform, but uh over the pastuh 10 years and we have seen
(02:34):
lots of progress on both fronts,and now I think um the
opportunity for photonics tomake big buzz and to have a
strong push on the quantumtechnology transport industry
(02:56):
right here.
So so I I I'm actually excitedby the progress that we have
been making at QCI and uh theprogress that uh people are
making uh in all different areasof photonics for quantum.
SPEAKER_02 (03:16):
The other thing that Patrick and I were discussing a
lot and were mentioning a lot,photonics also seems to be the
one and only bridge betweenquantum computing and quantum
communications, which isprobably kind of like the
cornerstone of the future of thefield, right?
(03:36):
So I think that's another veryimportant angle when we talk
about photonics.
SPEAKER_01 (03:41):
This is this is a hundred percent uh correct.
So the only way to connectquantum computers over distance
is by photonics.
And uh we have seen in thedigital era that we have been in
(04:02):
uh over the uh since uh um thelate 90s that uh so the power of
computing is really modifiedwhen you can connect many
computers together.
And uh I have no reason tobelieve that uh this is not true
(04:26):
for quantum.
In fact, uh so I believe that uhthe quantum future really um
hands on the success of quantuminternet as well, where we
connect many, many quantumdevices uh by internet, and the
only way to do it to the best ofmy knowledge as a physics
(04:53):
professor is by photonics.
So this is why um um that uh umprocessely as you just said that
uh well since we are going touse photonics to connect uh uh
quantum machines, quantumcomputers, the easiest way to
(05:17):
make it happen is to usephotonics to build quantum
machines, to build quantumcomputers, so that uh you don't
need to have this uhintermediate step of converting
um the information carrier of umsay from Adam uh to photonics or
(05:41):
vice versa.
But uh on the other hand, so I'mnot saying that uh um uh we have
to use uh photonics to buildquantum computers.
In fact, so I believe in theopposite.
I believe that uh the quantumspace is open, open not just in
(06:07):
the technology approach, butalso in this wide application
space.
I believe the future of quantumwill consist of uh say those
atom-based, uhsuperconductor-based, ion-based,
and photonic-based quantumcomputers, and each of them will
(06:32):
um take care of uh um some nicheapplications in its own areas
where uh it makes most of sensein terms of the cost of
operation, in terms of the speedto solution, for example.
SPEAKER_04 (06:52):
We've said that.
We we've discussed the fact thatthere's we don't there there's
no winning modality, thatthere's so many modalities, and
we may have others to discoverthat we haven't seen.
It seems like any binaryproperty at the quantum level
can be turned into a qubit ifyou if you engineer it right.
SPEAKER_01 (07:10):
Indeed.
Precisely.
Patrick, uh do you have aphysics PhD?
SPEAKER_04 (07:18):
I do not.
No.
I but I hang out with peoplelike you way too often, so it's
it ribs off.
SPEAKER_01 (07:24):
Impressed.
Yes, indeed.
So in fact, uh any these greatquantum state system can
potentially turn into aninformation processing machine.
And uh depending on whatapplications, it could be
computer, right?
It could be encoder, right?
(07:46):
It could be sensor.
SPEAKER_04 (07:49):
Amazing.
Uh I I do have one question thatmight might might veer on to the
professor side versus the CEOside of your uh personality,
which is we've we've seen somearticles in the past that said
there was a hypothesis thatmaybe instead of entangling uh
instead of measuring and andentangling that that you might
(08:11):
take, especially in photonics,entangle millions of photons and
measure along the way.
So as a way to basically,because it's it's the fragility
of the state that's our biggestchallenge.
And so, like normally you youtake and you you operate and
then you entangle and then youmeasure.
(08:32):
But but someone posited, and Ididn't really get enough about
the article, um, I didn't readthe underlying paper, um, that
maybe you could entangle a hugeamount of photons at creation
and have like this herd of themthat you could measure some of
them along the way to make sureyou're going the right
direction.
Is that something that that QCIis looking at?
(08:54):
Or are you more theold-fashioned way of like
manipulating entangled states asyou go?
Let me answer your question.
Uh and if that's an unfairquestion, I'm I'm happy to move
on because it's a weirdquestion.
SPEAKER_01 (09:11):
No, it is a a very good and a deep question.
So I can tell you this.
Um many physicists, includingmyself, have spent over 10 years
trying to find a way toefficiently create entangled two
(09:32):
photons.
Just two.
And it's already very hard.
And the big issue there is theuh standard way for us to create
entanglement so the processitself is uh render, is uh
stochastic.
(09:54):
Meaning that uh so you don'thave much control of when the
two entangled photons can becreated.
But uh say maybe on average umeach each time you try uh you
can you can say have uh uh uhsay eight eight percent five
(10:16):
percent probability of gettingthe entangled payout.
This is actually an issue forusing the photonics uh for
quantum computing because uhjust as you said, Patrick, for
many applications you do need toprepare uh many many photons
(10:42):
that are entangled.
And uh now if you only have uhfive or eight percent
probability of uh to create uhone pair, so the probability of
being able to get many many pairwould be very very slim.
So we don't like that.
(11:03):
And uh I actually myself spentuh geez studying studying from
2010, so really looked into thefundamental physics uh first and
the uh uh engineering aspect ofhow to make it happen.
(11:26):
So in fact, we find a way tomake it happen, although right
now, so there are still some uhengineering challenge uh
actually on the manufacturingside to make the device.
But hey, so we find a way toallow you to determine
(11:47):
deterministically createentangled photons so that uh
when you say that give me onepair of entangled photons, you
will get one pair.
SPEAKER_04 (12:02):
Yeah, we we assume it's that's a step that
everybody simplifies.
Since the first time we Ciprianand I started talking about,
well, you entangle a couplethings and then and we skip
that.
And so it's fascinating how howin-depth it is and how hard it
we we don't take it, we take itfor granted that it's it's not
as hard.
SPEAKER_01 (12:21):
And uh and so now uh so this is actually what's uh
what's happening now at QCI now.
Uh for uh for that to happen forthe deterministic generation of
entangled photons, uh you willneed to make very specific uh
(12:43):
photonic integrated circuitswith some crazy parameters for
you to hit.
And uh well, so uh since Iproposed this approach, say I
think uh there are a total ofeight pretty crazy parameters
(13:05):
for us to hit, and we have hitseven.
unknown (13:08):
Wow.
SPEAKER_01 (13:09):
We are trying to hit the eighth one uh in the QSAI
fab now.
So I tell you what, so I feelgood about it, but uh after have
pushed this like for for forover 10 years, I'm kind of
getting old, so I'm pushing mymy engineers that okay guys, I
(13:31):
told you the recipe, right?
You can just make it happen andthen come back.
SPEAKER_04 (13:36):
Oh I'll be I'll be on the beach.
Just let me know when it's done.
SPEAKER_01 (13:41):
So so so so the real requirement is that uh so you
will have to make this a veryhigh quality, what we call micro
ring optical cavity.
And uh in order to get the Q, uhthe Q factor of the cavity.
(14:02):
Okay, guys, stop me if I startedto talk too much into physics.
Um, that's fine.
We we like that.
We needed to we needed to cut umsome uh micron size bridge uh
(14:23):
reach waveguide on uh uh thinfew lithium niggas wafer with
the surface roughness like lessthan 300 picometer.
Wow.
It it is crazy, right?
(14:44):
Um but I can tell you that uhthat that's what we have uh we
have achieved last year, butthis is just the first step, and
there are multiple steps, and uhthe team's working hard on that.
While uh I I probably can findone one or two days uh to go to
(15:07):
the beach, as Patrick just did.
Um but but now getting back toyour original question, Patrick,
because you really are asked aastronomical question that can
we entangle one million photons?
(15:32):
I guess uh then after my longanswer, the short answer is that
yes in principle but yes inprinciple, but uh I don't see
how in my lifetime, butfortunately we don't need that.
(15:52):
We don't need it to entangle onemillion photons for us to build
a quantum computer usingphotonics.
As long as say we can make uhhundreds of thousands of uh
pairs of entangled photons wheretwo photons are entangled in
(16:16):
each pair, so we can alreadybuild some very useful and very
powerful quantum computers.
SPEAKER_04 (16:24):
That that makes perfect sense because I think
the context of the entangle amillion was the fact that there
was so much loss intransmission.
But you're you're talking aboutcalculation as well, you're
talking about gates.
Yeah, and and um so you guyshave built a universal quantum
computer gate-based um withphotonics.
SPEAKER_01 (16:45):
Uh no, so we have walked on this and we figured
out how to build, um, but uh sothat this work is still ongoing.
What we have now on shelf um uhfor customers uh to buy as of
now is a specialized uh quantumoptimization machine that
(17:11):
definitely utilizes quantumeffects, but it is not
gate-based.
SPEAKER_04 (17:17):
Um okay.
SPEAKER_01 (17:18):
And they are very good uh uh for solving a pretty
uh large portfolio ofoptimization problems, uh but it
is not a general-purpose quantumcomputers.
SPEAKER_04 (17:35):
Is that a goal down the line for you guys, or is
that not an area that you'regoing into?
SPEAKER_01 (17:40):
We are working on building the gate-based uh uh
quantum computers.
Yes, so this is our goal.
But um the approach that uh weare taking at QCI is that uh so
we don't want to define uh whatkind of quantum computers or
(18:03):
what quantum computers shouldlook like by ourselves.
In fact, uh so we go to thecustomers, we talk with them and
we understand what they reallyneed.
And we build quantum machines uhto meet their need.
(18:24):
In fact, uh very interestingly,so while physicists, including
myself, um were very fascinatedon building the gate-based uh uh
general purpose quantumcomputers, actually.
So many customers uh they thethe feedback that we got from
(18:48):
them uh um um has been that uhuh so do you have a quantum
machine that can solve myproblems now?
And what we find is that uh sofor many of those problems,
gate-based machine actually doesnot give us the edge, does not
(19:10):
give us the advantage.
It's that some specially uhspecial purposed machines can
solve their problems.
So this is how we get startedwith the quantum optimization
machine.
Cool.
SPEAKER_02 (19:27):
That's I think that's very interesting because
one of the themes that that uhoccurs quite a lot, and we
discuss about it quite a lot, isthe difficulty in essentially
embedding real-world problemsinto the space of quantum,
right?
Defining a problem in a way thatcan actually be addressed,
(19:50):
right, with a quantum device ora quantum computer or however we
want to name it.
And I think besides the specifictask of building these devices,
the theoretical problem ofdefining the problem in a way
that's suitable, I think is iskind of the other big big goal
(20:10):
here for everyone in the in thefield.
SPEAKER_01 (20:13):
And uh do you know the fundamental reason for such
difficulty for cus uh for peoplein in needs uh to formulate the
problem to run on quantumcomputers?
What is what is it?
This is because in many cases,say quantum computers uh have
(20:40):
been designed based on theunderstanding of scientists
criticists.
Right.
So so it is that okay, so forexample, just to use this
gate-based.
So so in fact, uh so there is norule set that uh so we have to
(21:01):
use the gate-based.
It was that okay, so when weopen a classical computer,
right?
So you will have the memory,right?
You have the processor, and uhthere that's a lot of
transistors, right?
So uh the easiest way to picturewhat uh a future quantum
(21:28):
computer would look like is thatokay, how about just uh we
replace all the parts of aclassical computer with a quanta
equivalent and to it's what weknow.
Yes.
If you wanted to build a newtechnology or build a new type
of hardware uh by harnessing atotally different uh physics,
(21:58):
right?
So you're starting.
point is should not be try tobox the new physics into the old
architecture.
You should uh go deep into thenew physics itself okay and then
try to find the most logical wayto build the hardware according
(22:23):
to the physics itself.
SPEAKER_04 (22:25):
This is why in every sci-fi movie all the aliens look
like humans.
They just have different likenose ridges and things like
that.
It's it's the lack ofimagination is what you're
getting at.
SPEAKER_01 (22:36):
Yes and uh and actually so I but that um um
actually so I have been I mean Ihad uh uh been struggled uh in
in the same puzzle and but thenum um since uh about 10 years
(22:58):
ago I started to tell myselfthat okay let's uh really try to
uh let quantum be quantum and uhtry to make friends with quantum
instead of trying to box thequantum in a classical
architecture so let's build uh aquantum machine that uh um
(23:25):
observe the unique property ofquantum physics itself and it's
and then see how we can leverageit.
How can we can leverage it.
So this is why we are buildingsome quantum machines at QCI
that is very different from whatanybody has done.
But hey so it works and uh uh wehave find that uh so they can
(23:50):
indeed solve um problems veryefficiently and uh give us
better solutions and uh I infact so when we first announced
this uh machine many people saidoh this is not a quantum
computer because in people'smind it is that for a quantum
(24:11):
computer you have to have thegate base you have to have the
quantum memory and we are doingsomething very different but my
definition of a quantum computeris that a machine that utilized
the quantum effects to solveproblems with advantage that is
(24:34):
unmatched by any classical andthat's working for you because
you guys just announced um uh anAI edge product I believe didn't
you yes uh this is a another uhpretty exciting development uh
in fact uh so uh we know that uhAI is here to change the world
(24:58):
uh for good or of or for bad Idon't want it to it's like the
internet it's gonna changethings but it's up to us the
jury's still out on that oneright and uh now so but uh one
thing that we know for sure isthat uh AI is going to consume
(25:20):
more and more electric power anduh it's not a sustainable mode
uh to uh further uh develop uhthe AI architecture by by
building uh nuclear power uhplants next to data centers I
(25:43):
don't like that idea and I don'tthink uh that idea can carry us
much further.
So instead so at QCI we'reasking ourselves can we offload
some complex and uh powerconsumption intense calculations
(26:06):
from the digital to the analogoptical domain because the
moment that we can do that wecan for example reduce the um
power consumption percalculation by many orders of
(26:28):
magnitude.
SPEAKER_02 (26:28):
In fact we have uh we have uh done some research in
the lab and we find that uh uhwe can use a single photon for
machine learning task and uh asingle photon has energy of 10
to the negative uh 19 joule soto just uh to uh put this uh in
(26:55):
the context uh project so uh thelight bulb uh in your room now
is probably emitting if you haveit on is probably emitting like
10 to the uh twenty two uh yes10 to the 21 photons per second
(27:18):
so yeah we just need uh almostno energy to to perform that
project particular ai so nowthis is why at qci we are
looking at uh photonics approachand by incorporating uh
photonics neural netlock uhneural network layer with the
(27:45):
digital layer so that we canhave this hybrid uh machine
architecture that uh can reducethe energy consumption by i mean
10 times i mean that's 100 timesthat's amazing i mean that's so
much bigger than what we've seenin other things so um i don't
(28:06):
know if you saw this in the newsa while back but google came out
with a way to compress memoryusage by uh the the leading
models and training by sixfoldthis is a much bigger savings
and so that kind of thing Ithink it was called turboquant
or something on that name uhfrom Google that came out memory
(28:27):
management and and it it itchanged the metrics of uh the
cost the the data centers theywere gonna build so this this
kind of thing could also have abig impact is it's it's
something you just announced uhin the last last month I believe
right yeah so we our first uhcommercial product was uh
released and uh and uh lastmonth and it's called neuron
(28:51):
wave and uh what it does is thatuh so inside the box uh so it is
a hybrid uh uh neural networkthat uh processfully does what I
said so a part of thecalculation is done still done
uh in digital so uh FPGA and apart of that is done in the
(29:15):
optics and uh by combining thetwo actually so we have already
seen significant uh um savingsin the power consumption but
this is is just a start yes welland and machine learning
underlies everything I I learnedthat from Cyprian right if it's
weren't for the machine learningwe wouldn't be where we are in
(29:37):
the in the LLMs right no but Ithink this this comes at the
perfect moment because whatwe're seeing there is the end of
the free tokens for everybodyparty um I think people are
starting to realize that thesehuge models that are being used
out there right are are likehuge energy consumers and moving
(30:02):
even kind of forward with theseit's just going to be not
sustainable.
I think uh a few months ago oreven perhaps a year there were a
few very interesting papers thatwere published that were using
the laws of thermodynamics toprove that there is not enough
energy on this planet to achieveuh like the Skynet level of
(30:24):
artificial intelligence becausethe curve is just exponentially
in in increasing so I think thetiming of your announcement is
is just great because people arestarting to realize that the the
party's over and especially theenergy party is is over and
there is like no unlimitedpossibility of moving forward
(30:45):
with what we have unlesscompanies like you come with
these amazing innovations wherewe literally scale it back down
orders of magnitude the energyconsumption.
SPEAKER_01 (31:01):
Yeah thank you so much I I think uh yeah so and uh
uh we are um we are uh strivingto contribute uh uh but uh uh
say I'm glad to say that uh umthere are um uh other players so
they are making the effort alongthe same lines so I'm actually
(31:24):
pretty optimistic that um um uhfor photonics uh uh will uh uh s
play more and more importantrole in ai so you know even a
couple of years ago Nvidia isreally looking at photonics to
(31:47):
move the data from GPO to GPOright but now we we are we
should be looking at photonicsnot just for transferring
information but also but alsofor computing processing the
information and uh I believethat uh uh there is a a very
(32:10):
bright and um exciting futurethere and one thing that uh I
believe everybody will benefitfrom is that uh the tokens could
uh become cheaper and cheaper aswe adopt more and more photonics
because the energy consumptionwill be will uh will drop down
(32:36):
significantly as we integratemore and more photonics into the
AI architecture.
SPEAKER_02 (32:42):
And we have to say something before we we stop
right which is criticalphotonics and if we talk about
in the context of energyphotonics has one like huge
advantage which is it operatesat room temperature.
SPEAKER_01 (32:57):
Yes that's yeah well so actually uh this is a uh not
a very um very uh interestingtopic that uh um uh if I may I
want to share some uh thoughtsthat I have had for a long time.
(33:17):
Please uh as a physics reresearcher actually so I feel
that uh um there's a uh majormisconcept that quantum can only
happen at room temperature so infact if you open any quantum
(33:40):
mechanics book there's no suchthing called a cryogenic or or
uh or cold or or closed systemin fact so we have seen so many
signs that uh quantum effects doexist at room temperature in
fact so quantum effects arealready with us in the in many
(34:05):
biological systems so and I I'mnot sure if if you guys heard
that uh actually so thephotosynthesis uh processes that
happen uh now in any green plantso there are quantum effects
(34:27):
there right so in fact quantumeffects help uh uh transport in
the the the the energy and tostore at the bottom of the uh
green leaves okay so quantumeffects already exist uh in
(34:47):
nature at a room temperature incomplex uh biological systems we
don't need to put the leaves ina cryogenic fridge in fact we'd
better not do that's right sothen the big question question
is that uh why do we need to goto cryogenic to build quantum
(35:12):
computers and and so nature ishas already told us clearly that
no you don't need to do that andthis is uh also uh a guiding
philosophy at qc i i've beentelling everybody that okay guys
I know that uh it's easier forus to observe some quantum
(35:35):
effects at cryogenic temperaturebut we don't do it at QCI
everything we do is roomtemperature why because we want
every technology that we areinnovating engineering and
turning that to manufacturingwill end in people's hands right
(35:59):
we don't need uh we don't wantany of our users to buy that big
cryogenic fridge in order to useQSI product instead amen they
should be able to buy a productand put in the backpack and just
use it as another say computingdevice communication device I I
(36:22):
really love the philosophy youhave of like letting the
technology letting the physicsguide how we use it and and uh
not try trying not to resort tocryogenics.
SPEAKER_04 (36:31):
I think we're out of time but I really hope you'll
come back and talk to us again.
SPEAKER_01 (36:35):
It's been a great conversation is there any
anything uh at that you stillwant to mention before we wrap
up uh I I just want to say thatI really appreciate the for the
this opportunity um uh toconnect uh with uh with uh you
and uh I hope that as uh myhonest uh uh opinion uh actually
(37:02):
so maybe as a physics uhprofessor and so so so uh uh can
be perceived uh as a scientificand and uh scientific insights
instead of uh um uh uh uh mycomments as the CEO because I
(37:26):
really shared uh what I havebeen thinking and uh what my
true belief is uh yeah I thinkwe'd be all better off if all
the CEOs were also physicsprofessors.
SPEAKER_00 (37:39):
I'm not sure about that thanks again for joining us
and uh and hopefully we'll seeyou again soon thanks again bye
thank you very much it's been apleasure all right bye everybody
see you next time thank youthank you bye bye cybercrime is
one of the biggest threats tobusinesses of all sizes and
industries with almost half amillion open cyber positions the
(38:01):
problem is compounded by thelack of available talent in the
marketplace at Pulsar Securityour elite team of highly
credentialed experts collaboratewith you to assess your current
defenses and develop solutionstailored to your specific needs.
With services ranging fromcybersecurity education to
advanced penetration testing andred teaming you can start
(38:22):
reducing your risks today visitpulsarsecurity.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.