00:00hi everyone welcome to the a 6mz podcast
00:02I am sonal today we're talking about one
00:04of the most exciting advances in the
00:06history of computing and next platforms
00:07quantum computing we start by talking
00:10about the almost impossible march of
00:12Moore's law going beyond debates around
00:14whether its reaches limits finally or
00:15not to what that means not just for the
00:17potential of quantum computing but also
00:19advances in parallel computing for
00:21machine learning and GPS to FPGAs and
00:23more broadly how the economics of all
00:26this continually change who gets to and
00:28how we innovate in the computing
00:30industry and then we cover what is
00:32quantum computing and where are we right
00:34now in the practical reality of what
00:36people can actually do it that including
00:38what the first applications will be
00:39especially given what classical
00:41computers can't do and who are the
00:43players in this global race our guest on
00:46this podcast is the CEO and founder of
00:48quantum computing company Righetti
00:50computing Chad Righetti in conversation
00:53with a 6 + z partner Chris Dixon so let
00:56me let's start with what's kind of where
00:57are we in the in the history of
00:59computing modern computers started and
01:01probably in for real world war 2 ish
01:05you know the PC revolution began the 70s
01:08and 80s Internet in the sort of 80s 90s
01:12mobile phones now right sort of in the
01:14heart of that revolution and you know
01:16you're working on this one new thing
01:17quantum computing yeah there's been
01:19several many revolutions and computing
01:22capabilities at the hardware level and
01:24not the software level I think there's
01:26been a few kind of inflection points in
01:28that and one was back in the late 50s
01:30when we figured out for the first time
01:31how to wire together many transistors on
01:33a single on a single chip and this was
01:36an invention of the planar integrated
01:38circuit over the past 50 or 60 years
01:40since then we've had we've had Moore's
01:42law scaling of those of the silicon
01:43based devices that have led to really an
01:47almost impossible scaling of the power
01:50that these microchips have and they have
01:52completely changed the world but that
01:54really just kind of the substrate layer
01:56of computing technology really the chip
01:58level you know over that time we've gone
02:00from from tiny chips with a few thousand
02:03transistors to chips with billions of
02:04transistors on them and the size of
02:06those transistors have shrunk by many
02:09many orders of magnitude over that
02:11and we're now at the point where
02:12individual transistors are about 10 or
02:1520 nanometers in size to put that in
02:17context a human hair is about 10 or 20
02:20microns I think so a thousand times
02:23larger and 10 nanometers is about a
02:26hundred atoms wide and a transistor by
02:28the way corresponds to a two inch vacuum
02:31tube if you see those those computers
02:33from like the nineteen forties that
02:35two-inch thing now fits in the scale of
02:38thousands in a human hair yeah and if we
02:39zoom out in terms of layers of
02:41abstraction what what is this transistor
02:42and what are we using it for in any case
02:44the transistor is the core logic element
02:47on on on a chip and traditional
02:49computing works by encoding information
02:52in zeros and ones in digital binary and
02:54those transistors represent that
02:56information so you can represent a
02:58massive amount of tree of information on
03:01these chips and we've also learned how
03:03to wire together millions of these chips
03:05and millions of processors in parallel
03:07to build large-scale supercomputers
03:09we're also at a point though where the
03:11ways in which we've been making those
03:13transistors more energy efficient so you
03:15can pack more of them more densely on a
03:16chip making them smaller and are
03:18starting to run into fundamental limits
03:20particularly the shrinking the
03:22fabrication technologies are hitting
03:24physical limits correct there are there
03:27are many there's a constellation of
03:29challenges that it's more than just a
03:31physical size one is the power density
03:32on a chip so when you when you switch
03:36the memory state of a transistor you
03:37generate some heat and that heat has to
03:39be extracted from the device to keep it
03:40from melting and that problem gets
03:42really hard as you pack them pack them
03:44more and more densely and another
03:46challenge is if you want to build a
03:47supercomputer out of these things you
03:48need to use many millions of processors
03:50can parallel well that was one of the
03:52responses to the diminishing Moore's law
03:54right with multi-core processors yeah
03:56parallel system of your typical you know
03:59MacBook today I think as you know it's
04:00multiple cores your iPhone does but
04:02these are essentially multiple computers
04:03running in parallel and then that at a
04:07larger scale is a data center which
04:08might have you know 10,000 of these or
04:10something but those also have limits
04:12because these things have to communicate
04:13with each other and there's just
04:14diminishing returns right as you as you
04:17connect more of these together in the
04:19same way that if 100 humans working
04:21together don't are 100 times more
04:22efficient than one human
04:24this is actually described by something
04:26called arm dolls law which is the lesser
04:28known but perhaps even more important
04:30today cousin of Moore's law and what it
04:32means is as you start to paralyze a
04:34computation across many many processors
04:36in parallel you get a diminishing
04:38returns because not every step in the
04:40computation can be effectively paralyzed
04:42some of them just have to happen
04:43serially and so the basic approach of
04:45building more more powerful computers is
04:47starting to hit hit some real limits and
04:49the physical size of the transistor is
04:51one but what that leads to from an
04:52economic perspective is that the cost of
04:55building the manufacturing
04:57infrastructure of the pot the cost of
04:58putting up a fab to build technology at
05:01the ten nanometer 20 nanometer scale is
05:03is extremely large we're talking tens of
05:06billions of dollars now to get to the
05:08latest generation of technology and
05:09there's very few organizations in the
05:11world that kind of that can afford to do
05:13that so the competitive dynamics have
05:15been shaped by that by this economics
05:16one counter-argument to this is from the
05:18outside Moore's law looked like a law of
05:20nature from the inside if you talk to
05:22people that worked at companies like
05:23Intel they would say every time they
05:26were you know they felt like they'd hit
05:28the limit and then somebody came with a
05:29breakthrough and it felt maybe they'll
05:31continue to come with breaks here's
05:32number one number two people say Asics
05:35more specialized chips if you look at
05:37your smartphone you have a video
05:38processor comms processor all these are
05:40the things maybe you will get more
05:41specialized chips and keeping Moore's
05:43law going for another few decades yeah
05:46and I think it's possible that that it's
05:48going to continue I believe there's a
05:50there's a seven nanometer node target
05:52from Intel the cost is monumental and
05:54right a point a very significant
05:57diminishing returns and so what's
05:58happened in the industry over the past
05:59five years or so is just relying on
06:02brute force acceleration through
06:04improvements at the hardware level at
06:06the a at the at the integrated circuit
06:08level have kind of slowed down and
06:09people have looked for other ways to
06:11accelerate data processing and one
06:13engineering right here is that people
06:14confuse Moore's law with Dennard scaling
06:16you know some Dennard scaling is a law
06:18of the sort of physical law around
06:19transistors getting some physical law
06:21but physical like the pattern of packing
06:24more transistors into a smaller space
06:25whereas moore's law really the spirit of
06:28Moore's law is an economic principle yes
06:30which is when the computing industry
06:31really cares about something and the
06:34economic engine gets going
06:37things tend to get better very quickly
06:38and so you see this with networking you
06:40see this with storage you see this sort
06:42of kind of across the board right one
06:44thing happening now is the computing
06:45industry is getting very excited about
06:46things like machine learning and then
06:48there's people like you working on kind
06:50of the next generation of things like
06:51quantum computers and so you know if you
06:54think of it Moore's law kind of writ
06:56largely it's this broader principle that
06:58like this whole broad system of
07:00capitalism plus research plus you know
07:04lots of smart people plus critical mass
07:06of ideas plus a whole bunch of other
07:09things has led to just this very very
07:11steady kind of rate of improvement yeah
07:14and what we're seeing now is that
07:16quantum computing is getting pulled into
07:17that into that ecosystem is beginning to
07:20be driven by the same the same economic
07:22forces that have been driving other
07:24other forms of Technology thus far and
07:26so what's happened right now as a result
07:28of all this economic pressure on the
07:30semiconductor industry there's
07:31effectively a sort of Cambrian explosion
07:33happening because companies can you know
07:36invest ten million dollars in
07:37manufacturing infrastructure and build
07:39individual chips that are close to
07:42rivaling the capacity of an entire
07:44supercomputer and that supercomputer the
07:46chips and that were built on a four
07:48billion dollar fact and it's not just
07:49quantum computing is really going to
07:50drive another acceleration of the pace
07:52of advance in computing capability it's
07:54also neuromorphic and we're talking
07:56about neuromorphic chips in individual
07:57handsets being available soon for
07:59machine learning NVIDIA has built an
08:01incredible business around GPUs which
08:03which really kind of owns a
08:05parallelization of tasks across a count
08:08a small number of processors hundreds or
08:10thousands of processors and a single die
08:12and then obviously FPGA based computing
08:15is a very Singh has been a significant
08:17advance as well with you know Microsoft
08:19embedding embedding FPGA in in Azure
08:22cloud servers as well yeah so modern
08:25computing has evolved to such an
08:26incredible level of performance it's
08:28really shaped the world but ultimately
08:29there is a conversation that is not
08:32happening today around all the things
08:34that computers do not do and well our
08:37you know the our laptops and our super
08:39computers and and Amazon Web Services is
08:42our amazing computing resources there
08:44are a bunch of things that they simply
08:46cannot solve and the reason that's the
08:49case is because they compute
08:50they operate in a manner that is very
08:52very almost very very dumb in a sense
08:54they map information into digital binary
08:56and there's nothing in the universe that
08:58computes in a similar manner except for
08:59our own computing technology today and
09:01ultimately the universe itself in nature
09:04at the lowest level operates on quantum
09:06mechanics and that's kind of the Machine
09:07language that nature uses so tell us
09:10what is quantum computing and maybe if
09:12you could go back a little bit or 100
09:14years or so when I like just briefly
09:16I'll talk about if you could the the
09:17kind of history of quantum physics and
09:19how that leads to quantum computing so
09:21quantum mechanics is a theory that's now
09:22over a hundred years old it was
09:24developed in the first two decades of
09:26the of the 20th century and for a long
09:29time it was really instrumental in
09:30understanding nature for very long time
09:32but is now at the point where we are
09:34able to build machines that explicitly
09:36behave according to the laws of quantum
09:38mechanics rather than classical
09:40Newtonian physics and we're able to
09:42control those systems in the laboratory
09:45and we're able to build artificial
09:46quantum systems on a chip and to control
09:49the quantum mechanical states of those
09:51devices so what quantum computing really
09:53comes down to is encoding information in
09:57quantum mechanical states of nature that
09:59we can control and deterministically
10:01steer to represent data in a computation
10:04and why is representing data in a
10:07quantum particle the quantum state why
10:10is that advantageous to the traditional
10:14method there's really two core reasons
10:16that it comes down to the first is that
10:18quantum mechanics quantum mechanics is a
10:20is a continuous theory and quantum
10:23variables are continuous variables so
10:25what quantum computing allows you to do
10:26is to compute with continuous variables
10:28rather than digital binary so any any
10:31any fraction from 0 to 1 as opposed to
10:37as opposed to just two digits yeah yeah
10:41exactly and the second is that the
10:42number of such variables that we have
10:44access to to encode information in a
10:47quantum system grows as an exponential
10:49function of the number of quantum bits
10:52on the chip this is completely different
10:53than how traditional computing works so
10:56if you have a if you have a chip with a
10:57million transistors on it and you
10:59add one more transistor you go to a
11:01million-in-one then you have a part per
11:02million performance increase in that
11:04chip roughly speaking in the best-case
11:06scenario and with a quantum computer if
11:08you have if you have a hundred qubits
11:09and you add one more you don't have a 1%
11:12performance increase you double the
11:14performance and that persists
11:16independent of the memory size so every
11:18quantum bit you add to to the system
11:20doubles doubles the number of continuous
11:22variables which we have access and what
11:24it means is what appear to be
11:25rudimentary quantum mechanical devices
11:28can encode a tremendous amount of
11:30information and can be used to compute
11:32things that are physically impossible to
11:35compute not only with today's super
11:36computers but with any foreseeable
11:38supercomputer that we're going to be
11:39able to build in our lifetimes or anyone
11:41else's lifetime what are some examples
11:42of computational problems that you could
11:45solve with a quantum computer and you
11:47couldn't with a classical computer I
11:48think there's really two categories that
11:50we're seeing today where this is going
11:53to be taken up first for practical
11:55computing applications in the first is
11:57you know very naturally derives from
12:00what computers are ultimately that's in
12:01computational chemistry so in that in
12:04that world you're using a quantum
12:06computer to to simulate and understand
12:08another another system that is itself
12:10intrinsically quantum mechanical things
12:12like small molecules or materials and
12:15ultimately what that's going to allow us
12:17to do is just get a much deeper
12:18understanding of how how different
12:21molecular species are generated what
12:22properties they have in ways that are
12:25physically impossible to explore today
12:27because the combinatoric
12:28of molecular spaces is is extremely
12:31large it turns out that those quantum at
12:33those those equations basically the
12:35Schrodinger equation which describe
12:36systems at the quantum mechanical level
12:38is extremely hard to solve on even on a
12:40large scale classical supercomputer we
12:42can write down the equations we know
12:44what the equations you know we know how
12:46how they behave we simply cannot solve
12:48them for systems of meaningful sizes a
12:50small molecule was something like 50
12:52atoms it is almost impossible to compute
12:54the exact molecular structure or the
12:56exact electronics just takes the it's
12:58just if you sort of graph the computing
13:00required with the number of molecules it
13:03just gets big it's unfeasible very
13:04quickly is the reason is because that
13:06small system is to some extent a small
13:08quantum computer and it behaves
13:10according to the same laws that give a
13:12quantum computer it's
13:13power so there's another area I talked a
13:14little bit about earlier about corner
13:16computing allows you to encode
13:17information and continuous variables and
13:19we're starting to discover ways in which
13:21we can take this compute power and map
13:23it on to optimization problems that
13:24underpin a lot of a lot of machine
13:26learning and over the next few years the
13:29quantum hardware that we're building is
13:30getting better at such a fast rate that
13:32we're reaching this point where the
13:34bottleneck is going to be understanding
13:36the best algorithms to run on those
13:38machines to get the most value out of
13:40that given compute resource that the
13:42quantum chip provides and part of what
13:44that implies is that you need to build a
13:46very sophisticated classical computer
13:49around the corner' computer to to both
13:52leverage its resources and to offload
13:53anything from that corner computer that
13:55can be offloaded so that the computer is
13:57doing the things that only it can do so
13:59there are two classes of applications
14:00one is our systems in nature which
14:02themselves have quantum properties and
14:05the second are kind of more you know
14:09classical computing problems that are
14:11just are so difficult so complex that
14:13they are unfeasible for current systems
14:16and so that would include things like
14:17you know machine learning and op there
14:22are kinds of optimization problems
14:23yeah and large-scale optimization
14:24problems now I'm always very careful to
14:26predict with the applications of a
14:27fundamentally new and very profound
14:29technology are going to be there's
14:31always stories you know retroactive
14:32stories you're gonna be able to tell
14:33about about the lack of vision that
14:35people show when you go back to
14:36technology look at like the early 1980s
14:40everyone talked about how the only use
14:41for computers was like recipes and like
14:43keeping your recipes and they had a
14:45whole bunch of predictions if you look
14:46at the old like ads or hidden very few
14:48of them predicted Facebook and Wikipedia
14:51so we use all the time so I would argue
14:53that the past 30 years have shown us
14:55that humans are amazing for learning how
14:57to use use computers to the to the large
15:00entire life's and when we as we build
15:04these systems and as the industry itself
15:06develops I think one of the things that
15:09I'm most excited about is watching the
15:11unforeseen applications start to
15:13materialize well there's kind of a
15:14yin-yang here right where so much work
15:17if you go to compute typical for your
15:18science department they're people
15:19working on better chips and things like
15:21this they're also some people working on
15:23they're all working on algorithms for
15:24classical computer has never been a few
15:25people working on quantum algorithm for
15:27the most part they haven't been focused
15:30on AI because they don't have those
15:31computers to work on and tests on yeah
15:33and so therefore you don't know even
15:35what the algorithms what even that layer
15:37is gonna look like like the programming
15:38languages and the algorithms and
15:40everything else let alone the end user
15:42applications right so we know enough to
15:45know how much we don't know and people
15:47have developed some early applications
15:50that will be able to run in very early
15:51in near term quantum hardware and these
15:54are predominantly quantum classical
15:56hybrid algorithms where you use a
15:57quantum computer to provide
15:59directionality into an optimization loop
16:01that you're running in conjunction on
16:03classical hardware and that's really
16:05exciting application because it it
16:07really puts the quantum processor in a
16:09position where it's doing the thing that
16:10it's exceptionally good at and in insane
16:13for the classical computing hardware you
16:15have a little bit like a CPU GPU and
16:17then you'll have your exact quantum
16:20exactly think about quantum computing as
16:23providing a new kind of computing
16:25capability that will be deployed in a
16:27heterogeneous computing environment and
16:29we've worked really hard to develop
16:30software that allows us to integrate our
16:33quantum punic capabilities seamlessly
16:35into existing classical cloud
16:37infrastructure we've developed an
16:38instruction language it allows you to
16:40write some simple programs that target
16:41both classical and quantum computers in
16:43the same in the same instruction so
16:45where are we on this you've been talk
16:46about quantum Peters for a long time and
16:48there's been various debates as to you
16:50know how it's progressing what's the
16:52state of the world in quantum computing
16:53right now on computing is arguably the
16:55most sophisticated technology that
16:57humans have ever developed we're able to
16:59leverage a physical theory that we as
17:01individuals that never see on a
17:03day-to-day basis because the world
17:05averages over all of that kind of
17:07quantum mechanical behavior and we just
17:08get the Newtonian universe and so it's
17:10extremely hard to build these these
17:13chips in to have the quantum mechanical
17:15effects that you utilize in a
17:16computation to have them persist for for
17:18a meaningful amount of time and that was
17:21the real bottleneck in the field for a
17:23long time this is coherence this is
17:25quantum coherence when I started my PhD
17:26in 2002 I think there's one or two
17:28groups in the world that had ever built
17:29and demonstrated a superconducting qubit
17:31with a measurable coherence time so
17:34so you the fundamental elements of a
17:35quantum computer and it's used both for
17:37the mathematical abstraction that
17:39algorithms theorists can use to develop
17:41an idealized two-level quantum system
17:43that has two physical two available
17:45states at the same time a qubit is also
17:47used to represent the physical
17:49instantiation of that of that two-level
17:51quantum system and this is very
17:53different than how we talk about
17:54classical computing in classical
17:55computing we talk about bits and
17:57transistors it is the logical element
18:00the transistors are physical I'm an
18:01executive it is represents both the
18:03nomenclature that we have today it uses
18:06qubit in phimosis in this kind of double
18:08mat double meaning and so we we use
18:11superconducting qubits to represent to
18:14manifest these two-level quantum systems
18:15that we use to encode information when I
18:17started my PhD at Yale the field was at
18:19a state where quantum computing was the
18:21excuse to do this fascinating physics
18:23research but there are very few people
18:25who were thinking seriously at that
18:26stage about building a real quantum
18:28computer we spent about 10 years as a
18:30community as a whole really
18:32demonstrating that we could increase the
18:34quantum coherent lifetime of the devices
18:36to the point where they could be used
18:38for a computation and then learning how
18:41to solve the fundamental problems about
18:43putting more and more quantum bits on a
18:44chip and so where the field is today is
18:46that we're really working on packing
18:49enough quantum bits onto a single chip
18:51where you can run a useful computation
18:53and to simultaneously increase the
18:56quality of the quantum boolean
18:58operations that you do during the
18:59computation so the error rates are
19:01sufficiently low that the computations
19:03are reliable so how big is the quantum
19:05computing industry / for research world
19:07right now like how many people are
19:08working on these kinds of problems as a
19:10field of physics research the the
19:12community has grown substantially over
19:13the past 10 years or so there's probably
19:15thousands of people around the world
19:17that would identify as researchers in
19:19quantum computing in terms of real
19:21effort to build practical quantum
19:23computers it's a much smaller universe
19:25and of course iBM has a very significant
19:27effort in this Google has a significant
19:29effort and there's an there's amazing
19:31scientists and researchers in both of
19:33these the these places Microsoft more
19:36recently has gotten involved and has
19:38started to make significant investments
19:39in quantum computing and then around the
19:41periphery there are smaller
19:42organizations exist in the ecosystem
19:44that are doing some combination of
19:47of research and in some cases building
19:49software tools or working on developing
19:51applet potential applications for long
19:53term corn computing quantum computing is
19:55very much a global effort there are
19:58significant efforts in Australia and in
19:59Western Europe extraordinary people at
20:01ETH Zurich at Technical University of
20:03Delft and all over the place obviously
20:06can't name them all there there's also
20:08signs of significant progress in China
20:11who recently saw a paper with a multi
20:13qubit experiment that was successfully
20:15run by a Chinese group this is a global
20:18race in many in many ways and quantum
20:20computing is going to reshape the world
20:21in a significant way I think because the
20:23impact of this technology will be
20:24profound and be felt across industries
20:26and around the world there's going to be
20:28another Silicon Valley where the quantum
20:30ecosystem is it kind of comes up we
20:33often use that term everyday and it
20:34doesn't sink in that hey it's called
20:36Silicon Valley because of silicon
20:37microchips so I picture a quantum
20:39computing company I imagine a bunch of
20:41physicists is that you know tell me
20:42about who works at Verde computing so we
20:45are a full-stack quantum computing
20:47company we design and manufacture
20:48quantum integrated circuits we integrate
20:51these quantum integrated circuits into a
20:54complex system that cools them and then
20:56operates them using a microwave an RF
20:59control system to to run computations on
21:03those on the on those chips and and then
21:06we have a software platform that
21:07connects up that corner computer to
21:10cloud infrastructure and allows you to
21:11run - run run quantum algorithms on that
21:14on that machine there's a lot of
21:15physicists physicists at various stages
21:17in their career we have what we call
21:19junior quantum engineers who are just
21:21coming out of college and really really
21:23great and brilliant young physics majors
21:25we have theoretical physicists
21:27experimental physicists we also have
21:28computational chemists we have a lot of
21:30technicians Electronics Technicians
21:32we've hired systems engineers from from
21:34Jet Propulsion lab and NASA we've hired
21:36FPGA developers from the aerospace
21:38industry who were building autonomous
21:40drones before it turns out that the kind
21:43of core technology problems at once
21:44needs to solve in order to build quantum
21:46computing are being solved in other
21:48places what doesn't exist all of those
21:50skills under one roof in one
21:51organization with all those people
21:53pulling on the same rope we have
21:54incredible business and people
21:55operations folks we have a lot of
21:57software engineers and there
21:59this is one of the most impactful things
22:00that a software engineer can work on
22:02today you have the opportunity to to
22:04make foundational contributions to an
22:06entirely new computing paradigm that
22:08will lead to fundamental advances in
22:10many different fields when do you think
22:12regular companies people will have
22:14access to quantum computers the idea is
22:16around neural networks and deep learning
22:18have been around for twenty thirty years
22:20people even trace the ideas back much
22:21earlier than that and ultimately from
22:23one perspective it was the availability
22:25of phalanx --is of GPUs over AWS that
22:28allowed this to really to really take
22:30hold because the number of folks who can
22:32contribute to improvements in lead from
22:34an algorithmic perspective was
22:36significantly increased quantum
22:37computing is at the early days where
22:38there's maybe a few hundred folks who
22:41are working on quantum algorithms around
22:42the world and I would argue that every
22:43software developer to some extent is
22:45working on better classical algorithms
22:46and over the next five years or so I
22:49think the number of folks who identify
22:50as quantum engineers or quantum software
22:52engineers is going to go from
22:53approximately zero today to a meaningful
22:55number and that the progress on that
22:57front will really accelerate and we're
22:59really focused now on on developing
23:02applications and working with early
23:03customers in these core application
23:05areas that we discussed and in really
23:07engaging with folks to to kind of kick
23:10off the flywheel of application
23:11development and discovery so like all
23:13computing platforms they'll be sort of
23:14this mutually reinforcing interaction
23:16between the computing platform and the
23:19software developers side and that hasn't
23:21begun yet until you get these things in
23:23people's hands and you see what they can
23:24do with them and all the inventive
23:25things they come up with exactly that
23:27flywheel won't start so it's quanta
23:28computing sounds like a very hard
23:29research problem and and it's not
23:32surprising that IBM and Google and
23:34universities are working on it how can a
23:35start-up possibly you know compete
23:38against these giant companies that's a
23:40great question and it's something that I
23:41I've thought about a lot and you know
23:44before I started the company I looked
23:45around at the world and ultimately it's
23:46a kind of mission that is best served by
23:48building an organization from scratch
23:50where you can kind of hand select or
23:53curate the the DNA of the different the
23:55different organizations within that
23:57larger company you got to build to
23:59uniquely position it to solve that set
24:01of technology problems at this point in
24:02history and that opportunity to build a
24:05company from scratch is very hard and
24:07there's a chasm you have to cross to get
24:09there but if you can do it it gives you
24:11a compelling competitive
24:12against a larger existing incumbent
24:15organization whose quantum computing
24:16effort is not going to move the needle
24:18in their culture think of this in
24:19analogy to electric cars you know
24:20General Motors would build an electric
24:22car and still there's got to be a Tesla
24:24and eventually electric cars and hybrids
24:26are going to kind of be a technology as
24:29adopted across the industry but there's
24:31one electric car company there's one
24:33that matters there's an economic angle
24:35to this too and the economic angle is
24:38that quantum computing sounds hard but
24:40it is very much a you know we're
24:42knowledge workers and ultimately it's
24:44the knowledge of how to build this
24:46technology that sets you apart and that
24:49is not something that can be reproduced
24:51at this stage in the industry with mere
24:53scale an army of fabrication process
24:56engineers is useful only if you have the
24:58foundational knowledge about what it is
25:00you're trying to accomplish and how to
25:01diagnose whether you've done it or not
25:03okay so it's a persistent rumor I on the
25:06internet forums is that quantum
25:08computers will destroy you know all of
25:10our cryptographic systems what what do
25:12you think's gonna happen there what
25:13you're referring to is Shor's algorithm
25:15ultimately and in Iran in 1995 a
25:18mathematician at Bell Labs named Peter
25:19shor discovered an algorithm that if one
25:22could build a large-scale quantum
25:24computer one could run an algorithm that
25:25would that would be able to factor large
25:28numbers in polynomial time what that
25:29means is that you'd be able to threaten
25:31the standard encryption protocols that
25:33are used around the world from Wall
25:34Street to the battlefield
25:35we're probably 20 to 30 years away from
25:37having a machine that would really be
25:39able to run Shor's algorithm on on
25:41practically relevant problem sizes at
25:44some point in the future corner
25:45computers will be able to crack RSA
25:47encryption yeah and so the question
25:49really becomes what is a shelf life of
25:50your secrets is the blessings for the
25:52field because that discovery led to
25:54research investment from the government
25:56that got the field started it's it's a
25:58curse to some extent because that
26:00application from my perspective is one
26:03of the least interesting it's not as
26:05interesting in relation to the other
26:06things that quantum computers are going
26:08to help us do ultimately the things that
26:10we get really excited about are using
26:11these machines to to build fundamentally
26:13more powerful artificial intelligence
26:15using these machines to to disrupt wet
26:18chemistry and to do simulation driven
26:20design in silico of new
26:22or new drugs this is going to
26:24significantly affect healthcare it's
26:25going to affect how we treat disease
26:27it's going to affect how we how we
26:29generate energy and how we how we feed
26:31ourselves as humans ok great thank you
