bad fliesen ideen grau
yeah at d-wave we have charted a course ofaggressively pushing the frontier of quantum annealing processors andperformance and we continue to do this with our new quantum computer the d-wave2x the most advanced quantum computer in the world back in 2010 you release thed-wave one which included a 128k with quantum annealing processor which wasrealized in a truly scalable architecture in 2013 we ship the firstd-wave to which included a 512 quantum annealing processor which outperformsgeneral-purpose optimization solvers and achieve performance on car and highlyoptimized terrific sellers now in 2015
we have announced the d-wave 2xincluding quantum annealing processor with over a thousand cubits as well asmany technology improvements aimed at boosting quantum annealing performancethe new system runs at a colder operating temperature below 15millikelvin which helps the quantum annealing algorithm staying low energystates and improved solution accuracy other improvements include increasedcontrol circuitry precision and fifty percent reduction in cuba noise whichalso combined to improve performance with over 128,000 josephus and tunneljunctions the new processors are believed to be the most complexsuperconductor integrated circuits ever
successfully used in production system our latest benchmarks on the d-wave 2xshow competitive performance against the best-known highly-tuned softwaresoftware running on cp on benchmarking problems that you have 2x times nearoptimal solutions to six hundred times faster than classical solvers if youcompare the quantum processing time to the cp the d-wave 2x was recentlyinstalled at the quantum artificial intelligence lab at the nasa amesresearch center in california this lab supports research into machine learningand optimization by google nasa and the us are a memory interest
i know everything intervals google andits partners to keep the system at the state of the aircraft seven years withnew generations of belief systems being installed the nasa ames as they becomeability to be with 2x is the culmination of years of hard work and we can't waitto see how the changes the landscape of kitchen i'm jeremy melton i'm the vicepresident of process development of the wave systems i'm responsible for the processordevelopment group which involves the design test and development and releaseof the quantum processor that the wave makes which is the underpinning of thequantum computers that we built each of
these black systems is a quantumcomputer inside that system is a thumbnail size quantum processor inorder to operate and develop quantum processors we require a fairly extremeoperating environment which includes ultra-low temperatures and an ultra-lowmagnetic environment and to achieve those we need a fairly large andsophisticated system the main room behind us houses the quantum processorand the data racks in the front fold the quantum server where users can accessand program the quantum processor and the other data tracks include thecooling system which includes pumps and and is responsible for the trippingsound you can hear in the background
inside the room before we get to thequantum processor there are many layers of fielding which allow us to create thelow-noise magnetic environment the chip system if you look inside the data racksyou'll see that they don't contain what you would normally expect so here wehave some scroll pumps which are mechanical parts associated with pullingvacuum on the fridge inside the room behind so this is obviously highlyunusual for for a date iraq but has been integrated so that it it wouldn't lookout of place in a data center and in fact the the requirements for runningthe system are consistent with all the services that would be available in adata center which would include chilled
water and our predominant in in thisrack is the quantum server which allows users from anywhere in the world tointeract with the processor itself this server is receiving information fromusers converting that into the machine language of the processor and sendingthat machine language into the room behind where high-precision analogelectronics will convert those signals into electrical pulses which are sentthrough cables into the refrigerator and down through the cooling system to thethe ultra-low temperature of processor it executes this problem and then andthen data from the processor is returned to the server and return to the user andthe system is very flexible and that
users can be programming in whateverlanguage they are comfortable with and interacting with it either with thissystem in their data center or even here in d wavelab the quantum bits in ourprocessor store information in the form of little magnetic fields to 0 and 1corresponds to a little magnetic spin the controls over the qubits are all inthe form of magnetic fields so the quantum processor is very sensitive tomagnetic noise and the magnetic environment that it sits in in order to keep that magneticenvironment really low as well as support to alter low temperature that isoperating at this system behind me has
about 16 layers of fielding betweenwhere i'm standing and where the chipsets at the center these these blackpanels are really for aesthetic purposes and they contain they integrate thesedata racks with the shielded room which is the first of those layers ofshielding and is filtering out rf signals so if you were to stand insidethe room and close the door your cell phone would stop receiving asignal the fridge itself has a combination of radiation shields as wellas magnetic fields which helps support the ultra-low temperature operatingenvironment and achieve the ultra low magnetic field inside the room we haveto keep electromagnetic noise very low
so even things like power outlets areallowed because power can be quite noisy without extensive filtering so insidethe room we have the cooling system and these are radiation fields so this isthe next layer of shielding inside the room the chip itself the quantumprocessor is sitting at the bottom of this array of field and i this is againthe first of several more layers inside these shields vacuum which is theinstallation of the cooling system when we want to swap the processor or conductmaintenance on the fridge we warm the system up we remove these shields andthen the processor is accessible as well as all of the internals of therefrigerator
my name is mary tom i'm the director ofprofessional services range i talked with our customers and partners aboutwhat quantum computers are what kind of problems that they solved and work withthem to develop software that uses computers we're standing inside thequantum computer and you know that the quantum world is delicate it's easily disturbed so there are manylayers of shielding and assistant which we've removed in order to look at theinternals here a traditional supercomputer may have many thousands ofprocessors and but in this quantum computer is only one chip located rightdown here when you send an instruction
of a quantum computer a delicate problemcalculation is taking place on that ship to produce solutions the system we'relooking at here is an ultra-low temperature of crack and encouragedbecause it works successful temperature use the kelvin scale to measurehousehold is getting in the kelvin scale a warm room is like 300 kelvin and zerokelvin is the lowest temperature that can be reached businesses percolate isabsolute zero there's two levels of cooling in thesystem the top one is a close to refrigeratorwhich takes us from 300 kelvin 250 kelvin and then down to three kelvinsince roughly the temperature of
interstellar space and then a secondrefrigerator at least refrigerator takes us in this plate from points7 kelvin . 1kelvin and then down 2.01 kelton when the system is operating this plate andopening mounted below it is ten thousands of a degree above absolutezero which is more than a hundred times colder than interstellar space here andthis whole system is designed to sustain those low temperatures in a continuouscycle when you send an instruction to a quantum computer the signals cast downthese wires along the sides of the refrigerator and at this point there's abreaking the wires were all the signals become superconducting superproductivity is this amazing effect
where you can send electrical signalswithout any resistance whatsoever so the electrons are really interacting withthe atoms in the wire down that's great because it means thatthis portion of the system has no energy dissipation of this is no key but italso means that we can't rely on the wires to cool the electrical signals andeliminate their noise so we have a custom filter bank accounts of herewhich cools cools those electrons an important aspect of the superconductivity is that these signals and the common processor unit and soonvirtually no power whereas the supporting systems around it seemedabout fifteen and a half camelot and
this remains the same regardless ofwhich processors and install the bottom if you were to take a fully populatedserver rack out of a traditional supercomputer in might consume 10 timesthat much power so for our first customer lockheed martin when theyupgrade from their first trip to the d-wave to they saw a relativeperformance improvement on the order of 10,000 times but the power consumptionfor the system was the same i'm mark johnson i'm a scientist d-wavesystems and i work with a team of other scientists and engineers developing thesuperconducting integrated circuit chip that's at the core of our quantumannealing processor we have our chips
fabricated on aidan silicon wafers andwhen we receive them back from the foundry they look like this here if youlook closely you'll be able to see in our repeated array of squares that aremultiple copies of our integrated circuit chips that integrated circuitchip is really just a series of metal layers predominantly which you can seeand cutaway here other the process that goes into fabricating this integratedcircuit chip is actually not that different from that which goes intomaking the chip that's the drives your cell phone or your laptop one of theprincipal differences though is that these metal layers are actually made uppredominantly of niobium niobium was
chosen because it's a superconductor sothe thing about a superconductor is that when you pull it down below a certaintemperature all of the electrons really collapse into the same quantity state itsince they cooperate with each other in the way that electrons in a normal leveldon't ordinarily quantum mechanics only exhibits on the scale of atoms electronsand photons superconductors are unique in this way and that you can see quantumbehavior on a much larger scale and it's these properties quantum mechanicalproperties that enable us to use these materials to make our quantum processorso this is a blueprint of one of our processors the design and constructionone of these is not really that
different from that of a conventionalprocessor that are really important difference is that the conventionalprocessor the ones and zeros are encoded as voltages and currents that on that chip in this processor thatinformation is actually coded in small magnetic fields called single fluxquarter for qantas magnetic slugs when we're programming the processor we loadhundreds of thousands of these blocks quanta into the into the chip intovarious circuits then the process or undergoes algorithm quantum annealingwhich is how it solves the problem that we opposed to it at the end of that allthe answers in the form also magnetic
flux quantity which are then pulled outof the processor to the four corners and then up into the room temperatureelectronics and outdoor conventional computer so the basic computationalelement in a quantum processor is a cubit qubit is a lot like a bit in thesense that it can encode the state 0 or 1 but it's a quantum bit you for quantumbit and it has the additional properties that can be in a superposition 0 and 1at the same time so this trip is made up of repeating tiles or unit cells andeach one of them is made up of eight cubits now if you take one of these unitcells well we can blow it up and get a bettera better view of it here
so r $operand cubits are actuallysuperconducting wire loops that are interrupted with a circuit elementcalled josephson junction and we make our cubits long and thin you can seethat there are four of them running vertically here and another for runninghorizontally here wherever these qubits cross or intersect with each other andwe put a tunable coupling element that could be used to allow the qubits toinfluence each other in a programmable way a cubit is engineered so that thereare two stable states corresponding to circulating current going clockwise orcounterclockwise and associated with those two stable states are magneticfield either coming out of the surface
or into the surface and they correspondto the zero and the one-state respectively for the equivalent ofcourse the cube can be in a superposition 0 and 1 and so in thiscase that means that these qubits can be in a superposition of of those two circulating current state'sgoing clockwise and counterclockwise at the same time so the idea that amacroscopic object like a wire loop could really be in a superposition stateof circulating going clockwise and counterclockwise is is common to it it'squite on but this is the sort of thing that happens in common visits on themicroscopic scale and this is exactly
the sort of phenomena that we'reharnessing to quantum annealing processor i see quantum computation is being atool that can solve problems that you could never otherwise solved doinggenomic analysis to help find better drugs debugging large complex systems tolooking for bad guys and large amounts of data space exploration try tooptimize water works building artificial intelligences that can help us all verydifficult complex problems building a quantum computer is a historic thing andits really like venturing into uncharted territory
there's no precedent quantum computingliterally taps into our most fundamental understanding of the laws of nature innature there's is somewhat spooky phenomenon called quantum mechanicswhere i think can be in two states at the same time and so the d-wave quantumcomputer uses that strange property what the quantum computer actually does itsable to compare many possible solutions at once so if you imagine a regularcomputer if it was searching for a solution it would have to go throughevery possible solution one-by-one see if that's a good fit what the quantum computer does is it itkind of life all the solutions down and
compare them all at the same time so itcan pick out the best one the d-wave system is built to solve discreteoptimization problem these are types of problems where you'vegot a load of constraints and you're trying to find the optimal answer toyour set of constraints so you've had some kind of natural disaster and you'vegot a lot of people who need help and you've only got a limited amount ofresources to go help so you've got a few lifeboats and a fewhancock what you need to do is work air order to send things and where to sendthings to save the most amount of people as possible
this is a hugely complex problem becausethere's so many possible different combinations of sending them indifferent so this is the kind of thing that finding a fast answer is is reallyimportant the d-wave computer has to what i would say really sort of exoticoperating conditions we put our quantum processor into something called thedilution refrigerator which cools it down to 10 millikelvin e which is abouta hundred times colder than interstellar space and essentially is the coldestplace in the universe we have to also create an extremely low magneticenvironment for the processor to city
that's about 50,000 times less than theambient first magnetic feeling and so there are these two extreme conditionsthat the processor has to be running in this company you know a few dozen peoplein 10 years i've managed to make a technology that can match and beat theexisting normal computer technology that's been developed you know allaround the world 60 years we have a incredible future ahead of us first andforemost we're just going to increase the capability of this processor willhave at least two more generations of this processor over the next five yearso from raw performance point of view this is gonna be incredible capability
will make a huge impact on the way things are done in this spaceof high-performance computing and solving these really tough problems foryou from my perspective the really the really crucial aspect here is that we'vebuilt a truly scalable technology and this is what we've what we've shown nowwith the 120 cubic processor the integration of this programmablemagnetic memory the fact that that that integration hasn't caused an environmentthat is damaging to the quantum mechanics that's going on in theprocessor we've been able to emerge on all of these different reallychallenging components to make this
technology real and to me that's why wecan say it's truly a commercial quantum quantum computer and in a lot of theseother other proposals for quantum computing there's no they are trulyscalable designs there's no there's there's there's always a limit to howmuch it can be scaled up to before yeah you know some other technology hasto be invented or created in order to in order to help that with the design thatwe have now it is it is a truly scalable design in addition it's it's incrediblyperformance i mean the way that we've designed this system you you can havethe flexibility to run whatever types of problems you want it's entirelyprogrammable and that it gives a very
important element we've tried very hardto make it has as closed and consistent with our our normal experience withcomputers as possible as opposed to being some exotic quantum computer and ihave to be a you know a phd physicist to to know how to even interact with thisthing it's actually something that's very simple that you know high schoolstudents can understand good program and actually have we brought high schoolstudents through here to take a look at it so i mean that's that's also a keything you could argue it's not it's a crucial thing too because this is aquantum computer technology that's not a crucial thing but i i actually think itis that that connection with the user is
something that really makes thistechnology adoptable and applicable to the really hard problems that would beable to solve we've essentially removed to reduce those barriers as much aspossible and it's all those things to me that that makes it a commercial quantumcomputer imagine being able to predict patternsin weather so that you could prepare for really dangerous storms or earthquakesor maybe drought imagine being able to predict patterns in the stock market sothat you could detect if if something was going wrong like if there was animpending market crash imagine discovering patterns in genetic datathat could lead to the advancement of
new treatments for cancer or differentdifferent treatments for diseases all these tasks at the moment require animmense a merit of computing the processing power and what we have is wehave computers that are just really inefficient doing this and we just throwmore and more resources at these problems and we're still not managingsolve them very well so we're designing is a computing system that's much moreefficient at solving these types of problems it's really just a case of using theright tool for the job i was at ucla studying physics
i've always been interested in thefoundations of physics the deepest aspects of nature and obviously quantummechanics is at the most foundational level and i got recruited while at uclaby a place called trw a professor some people were interested in me i get somework at ucla as an undergraduate published papers as a junior college andthat got the attention of some people at trw and notably a kind of famous guyarnold silver he had he was the co-inventor of the squid which is anacronym standing for a superconducting quantum interference device the squidsand interesting object because it's actually the basis of the quantumcomputer working on and and others in
the superconducting community and it'san object that was predicted historically to be the first object todemonstrate weird quantum phenomenon at macroscopic scales something you can seewith the naked eye my first job out of college was with theco-inventor of the device which would eventually be the device to show amacroscopic quantum effects and so when i got hired into this group the the whatthe group was working on was a quantum electronics superconducting electronicsso that had the promise of operating much much faster a thousand times fasterwith a thousand times less energy consumption right and this would be ableto extend even moore's law well into the
future and and solve problems that wouldotherwise be beyond the pale of what cmos could do right so that's what thisgroup did and so when i got hired my initialattraction to that whole field was one they offered me a job too i was working with some very well-knownpeople who had done some really at work and three it was quantum mechanics on atabletop so we were building this technology and we have the most advancedprocessing capability in the world at the time and i learned a ton of stuffand learn how to build superconducting electronics from the ground up and learnall kinds of cool physics and what i
used to do for fun i always you know wastrying to educate myself about new things going to feel caltech had aseries of free lectures the beckman lectures and in 1997 i think it wasthere was a lecture entitled quantum computing by a guy named john press killa very well-known and respected theoretical physicist and i knew whatquantum mechanics was i knew what computing was but i didn't know whatthey had to do with each other it was it wasn't the buzzword that it istoday so i went to that lecture and i took arnold silver you know adventurethe squid with me and a couple other people from trw and we went to thislecture and blew my mind
yes sir this alice in wonderland stuff imean one interpretation of quantum mechanics is that there's paralleluniverses right and that when you look at the quantum-mechanical equations andthey tell you that all these disparities physical phenomenon happening at thesame time you know like the same physical object being in many places atonce or living out many possibilities simultaneously that kind of thing the idea is the basic idea behindquantum computing is if i could have a if i can have single object without manypossibilities simultaneously what if i could have computing elements the samephysical hardware i behaved as if it was
many many pieces of hardware for many isone processor operating like it's an incredible number of processorsoperating in parallel doing different parts of a very difficult problem butthe thing is the way parallel processes done today you actually have to build physicallyyou processors right what if one processor could behave like10 to the five hundred processors operating in parallel you know itstaggers the mind right so he was giving this talk and david deutsch choose sortof the progenitor of quantum computing easy big fan of the many-worldsinterpretation of quantum mechanics he
came up with this idea by saying youknow what if what if you could access all these parallel realities intodifferent parts of a complex computation and them simultaneously and and somehowextract information from the multiverse right this is an amazing thing you'reharnessing a whole new resource in nature your technologically accessingthese parallel universes and this is something that human beings have notbeen able to do therefore so i went to this lecture waskinda had kinda had my mind blown and at that moment decided that's what i'mgoing to do and beyond that i said i'm already working with materialssuperconductors and building
superconducting circuits and objectsthat behave quantum mechanically a macroscopic scale so when i walked outthat lecture i thought he's talking about doing quantum computing he talkedabout you know doing it with ions you know nuclear magnetic resonance photonsand when i walked out i thought well yeah those are all microscopic systemsand it's very difficult to access and control and manipulate and engineermicroscopic systems but superconducting circuits are macroscopic if you couldget and they already demonstrate a lot of the requisite behavior if you couldbuild a quantum computing elements out of superconducting circuits you couldhave a macroscopic technology where it's
already known how to build large-scaleversions of that so i got very excited about that yes i i proposed it to arnoldand at the time he said you seem to really be into this i like that youridea i'll pay one day a week to think about it so one day out of the week just at andgive me a lecture every couple of weeks about what you're thinking so that's howthat started and in the course of that over the next two years i read everypaper i could find on the internet spending late nights at the ucla libraryand is interesting because what i discovered was this amazing story thatwas going on this idea why is it if the
microscopic world behaves in this alicein wonderland like way and were made out of those atomic constituents why don'twe be that way this was a big quandary why does quantummechanics not work when you get to a certain scale and yet we saw signaturesof quantum mechanics and things like superconductors at large scales is veryintriguing so when i started looking into thisfield i found a paper ucla by a nobel-prize-winning physicist tonylegged and what he predicted in this paper was if you wanted to see aschrã¶dinger's cat now not actually a cat but if you wanted to see amacroscopic object very very large
compared to scale of atoms that would beyou could put into two states simultaneously and predicted it would besquid and that was and i was working for the guy who invented it and i was in agroup where we built them routinely i went to ucla library i read that paper iwent wow this was nineteen ninety eight and nine i was reading all these papersfor two years i've read papers i reported arnold silver about my ideasabout how to maybe pursue this feel about how we might leverage what wasgoing on in superconducting electronic material you and then i went in twothousand i went to a conference it's called the asc the appliedsuperconductivity conference and at that
conference there was a group at in newyork suny so state university of new york in stony brook and a group of a guyjonathan friedman and jim lukens succeeded in making the world's firstrunners cat they put a squid in a macroscopic superposition of twomutually exclusive states and it was in new york times and i had been readingfor two years anticipating the day when someone wouldfinally get to superposition so they had tunneling their energy and i went tothat conference and i saw his poster and i haven't seen the new york timesyet and it was like my god there it is and what he had done a simple way ofexplaining it is if you take take a ring
of metal it's just a ring of metal youcan look at the periodic table niobium you know it's it's a pretty commonmaterial but when you put that material at low temperature in liquid helium kindof thing it becomes a superconductor has these macroscopic quantum propertieswhat they have done is taking it to very low temperatures millie kelvin andessentially put this in a very strange state of affairs they had this ring ofmetal there's a detail something called thejosephson junction in there but they ran a current clockwise right all thecurrent was going clockwise it's important all the currents goingclockwise and all the current was going
counterclockwise simultaneously so theydid an experiment that demonstrate this so that's rather like you being at thevelodrome on your bike and going clockwise and counterclockwisesimultaneously one individual rights that that kind of crazy and i said thereis and so that new york times article was true owners cat lives so quantummechanics does apply on the macroscopic scale its universal and beyond thereason we don't normally see it is because of all these interactions withthe external environment with this object they had isolated the objectsefficiently and had the right for what records of properties to demonstrate thefull-blown quantum mechanical effects at
macroscopic scales so this was to me youknow a revolutionary thing and i said there is that's my quantum bit becausethere's lots of ways that you can represent zeros and ones you could bethe light switch you know up and down to be a transistoron and off and in this case current going one way could be a zero currentgoing the other way could be a one but you can put in a state of being a 01simultaneously and that quantum bit that's your cue bit and isaid that's the basis of the quantum computer i want to build and i justpatch this idea i thought what if you could get some of the best researchersand feel like the group that
demonstrated this get them well-fundedcollaborate with a group like trw that had this industrial might behind it tobuild devices do experiments microwave engineering all that stuff materialscience because the materials there's there's fluctuations and materials thatmake the quantum effects go away but we had a whole material science division isaid what if we were to take that industrial might kind of like when theyramped up the manhattan project with all the isotope separation and and thencouple it with some of the best physicists and feel and then beyond thati said you also have to look at applications you have to look at themathematics that underlies the quantum
information processing and those twoyears of reading i had surveyed the entire field i said here these thesephysicists are doing some great stuff here is here w that has all thisindustrial might hear some guys who are doing really great algorithmic work so ihatched this idea for a quantum computer manhattan project two weeks later to theday a researcher at the colorado branch of trw randomly called our group darpahas just issued a baa which stands for broad area announcement for where theyare putting aside 100 million dollars for projects in quantum computing i saidi have this little mini manhattan project i'm going to figure out i'mgonna say how many people what resources
you know how many experiments i figureall that stuff out i wrote milestones and all that mrs. what would it take todo something substantive in five years minimum and and even that i was underestimating and i came up with about 10 million bucks five years and they saidyou'll never win that i said that's what i'm gonna go or so i wrote a very ambitious proposaland in the meantime i went visited all these scientists who would be on theteam and got another team's the research they were doing and put it all you knowput all the stuff and follow about what we do organize all the efforts thatwould interleave in the right way all
the right stuff and send it in and ithought well if nothing else i got to meet all these really great smartinteresting people and been exposed to this field and you know i know its highmaybe improbable two months later a letter arrived you've been selected foryour project you want to ten-million-dollar dr. project inquantum computing and i was principal investigator and that's how it started so that was my darpa project now thedream project very quickly i i came to realize that when working on governmentprojects you know tremendous bureaucracy slow working with academic groups thataren't used to working on a manhattan
project they were disparity groups indifferent locations they had you know they had teaching responsibilities theyhad you know that kind of thing the culture wasn't the right kind ofculture you know i started realizing i need to co-locate these people they haveto be dedicated to this and this only they have to be used to like hittingmilestones and one of the first one of the first things that i knew wouldn'twork is i said well okay you need it here's 10 million dollars put in thebank account and then based on the milestones that we've agreed to i willpass health money but they did instead was they said oh here's how much moneyyou had allocated to you know cindy
group and this group nectar and theywere they were giving it to them individually which i'm and i said youjust ruined the team because while they'll be nice and collaborative withme unless there's somebody who's handing out the money and their deliverables andyou know all of that i can see the writing on the wall i saidthis is not going to happen well i was giving one of these talks tothe broader community some people i i piss them off they thought i was beingcritical i thought that was great physics goingon but i said this activity the way it's organized will never build a quantumcomputer and shortly after that i you
know i was reading the book on craigventer and how he map the human genome so there's this big internationalproject rights gonna take a decade billions of dollars and he said i'll doit in three years for 300 million and he did right and there's caveats there butnevertheless i looked at that same thing and i said i want that's another modelof a more recent variety and i was also directly involved with a friend of mineleft here w built a company and build a a micro fluidics technology so thismicroscopic you know chemistry lab on a chip kind of thing problems that existedfor decades he solved in months and i saw how he did it you know rapidprototyping lots of experiments . of
time you know kind of a similarphilosophy what i had so i was at a conference and i was giving a talk tothis effect like here my impressions about what we need to change to makethis happen as a technology one of the people in the audience was colinwilliams he was a pretty prominent guy used to work with stephen hawking youknow in cosmology in cambridge got very interested in artificial intelligence iended up being the user professor at stanford wrote the first text book onquantum computing the textbook that george rose-red that got him interestedin the field now at the time geordi rose who was doing thisintellectual property company you know
we had a few people the idea was ifsomehow i could shuttle venture capital two groups that had promising sort ofmodels for computing hardware implementations because i'm a theoretical physicist andthis group i'm working with understands that stuff we can we can be this kind ofthe people between the resources and the people who are doing their research theexperimental research so they said we would provide direction you knowtechnical direction theoretical support money equipment whatever to assess thevarious models of quantum computing shuttle money and resources to the bestgroups doing what we think are the best
ideas in exchange for which we would getintellectual property and then we build this big patent portfolio and whenquantum computing comes of age and and we have also select property we clean upthat was the business model that was the 1999 d-wave it started up and he wasdoing that for a number of years and came to the conclusion the sameconclusion i did for different reasons that this disparity amalgam ofresearchers distributor researchers will never build a quantum computer there'snot enough cohesiveness culture drive ambition whatever so he was gettingfrustrated this is bad i think it was about fiveyears of the project yeah it was about
two thousand four gave this lecture iwant my diaper contracted 2001 worked on it for three years cool science but i was just made it thathow that the pace of things where we go i gave this talk so colin williams whosebook i also read approaches me after this lecture and says we should start acompany now i said to colin no offense for theoretical physicist don't know howto start a business i don't either i think i know how to puttogether a good technical project and run it and inspire people and all thatbut raising money and venture capital and anything about that now in we weresitting at stanford late at night in a
little cubbyhole you know what we'relooking some nice garden and he said to me well you know there's this guy georgezeros canada and he said this guy raise twentymillion dollars and that was the moment where i said haha that was my next ahamoment i said doesn't it make more sense for us to team with a guy who's alreadyraised the money because you think that my manhattan project the right way toproceed and he has the money and how about that so eventually thought you know it'd benice to know more the company but maybe that's a good idea
so he he wrote a very nice letter ofintroduction to georgie saying if you if you want to build a computer should talkto this eric i think he has the plan to do it and vice versus any nice letterabout jordie introduced us and this was the difference i noticed with jordy youoff times you'll meet someone have a conversation live action items so let'smeet the month and have lunch and discuss so jordie got that letter i gota call the next day he was in france with this group into clay there'sactually doing some good work and he said i hear you have a plan to build aquantum computer i said yep and i said i'm in france and i thought he would saymaybe we'll meet in a couple of weeks or
could you come to vancouver and he saidi'm going to buy plane tickets tomorrow i was at uw and georgie showed up iremember picking him up at the he was a little hotel by the airport and you knowwe didn't know what each other look like but somehow i think colin told him i wasa rock climber he said something about jordie being you know a wrestler in theolympics or something so i showed up in this little hotel and i see a guy whodidn't look like a physicist this big bulky guy and i said the wrestler hesaid the climber i took my mini manhattan project proposal and set thisis what i would do is what i wanted to do in this government contract butbecause of all this you know despair
labs and conflicting cultures the wrongall that and he said that here's my business model what do you think of it as it'll neverwork and he asked me why and i get the whole set of several reasons and heagreed with them always said that's exactly been my experience the last fiveyears what would you do so for three days i told what i would do who i wouldhire you know how you hire a team it has to be applications driven likeeven those at an early stage like what are you gonna do with it if you have ityou have to start thinking about it because the the applications define thearchitecture define the device
requirements right so i said let's startfrom a processor not from individual elements you know what is this thinggoing to do and then let's ask yourself is there a model that's easier toachieve so i had this idea like there's this continuum of quantum effects someare easier to achieve and others could we build some device that didn't havethe full pallet of quantum mechanical effects but still did something usefulon path to the dream machine you know we have the this revolutionary idea forquantum computing and then you know after three days he said let's make yourvision into the new vision for d-wave this is phase two
let's let's go from an ip company to atechnology development company let's do it you just told me so i said and said about 20 millionbucks higher your dream team let's go never had anybody say that to me beforeso it's like wow so i decided to take him up on it i tooka little time at first just doing kind of part time and then pull the triggerand said okay and our first you know dilution refrigerator was something ibought on our quest project and talk to darpa and got them to transfer it to usand we paid for it but so instead of 18 months we were online a month you knowrecommended a group of people so they're
a bunch of world experts superconductivity circuit designers and physicist from t really worked with mebefore and i hired them are all some of the best people working tmw i you knowwe hired a guy that he both he and i like maryland has shown first quantumentanglement and superconducting devices and then it was i started hiring likethe decor technical team and then he had you know that the theoretical physicistyou had before and we put the team together the next big hurdle was how arewe going to build things we need a superconducting fat this brought upanother issue this is about doing things in a different way the conventionalwisdom for a long time in the
superconducting community was that wellyou can't get like the semiconductor community guys to build superconductingyou know circuits all the superconducting groups in the world thatdid that built superconducting circuits were usually physicists you know withsome technicians or something they would build their own cleanroom they would buya few pieces of equipment the problem with that is it was neverdone in a production facility and it was like you know little wafers one way forthe time and i looked at that when i was at trw because i did that too i learnedhow to build superconducting circuits from the ground up at trw because i hadto for my research and i said you know
doesn't it make more sense to teachpeople who make a billion devices on a chip how to do superconducting stuffit's not exotic to build the physics is exotic that load temperature but it's not that exotic tobuild so very long story short myself and jeremy hilton who you've met the vpand we went we searched the world for semiconductor world-class kind ofenvironment like an intel environment that was willing to do rnd you knowsmaller volumes and and different materials and landed here at thisfacility and i taught semiconductor guys all about superconducting stuff i didn'thire superconducting people i heard
semiconducting people and some of thesome best-in-class who were responsible for 65 nanometer technology you know andthat became the d-wave team i put a deal together with this place in siliconvalley and leverage the capability here and in two-and-a-half to three years webypassed the world capability and superconducting circuits that have beenbuilding for 40 years and i said this is the craig venter manhattan project it'slike because we had to because you can exist in in these other groups they getgovernment funded over the years they never get past a certain place but it'sokay you keep getting this government money in our case if we couldn't build aprocess that exceed everything in the
world in a couple years we're out ofbusiness so we did so we had to get creative so i came in silicon valleytaught semiconductor people how to do superconducting stuff and we bypass allthe capability in the world in a couple years so and that kind of bring us towhere we are today that we have now we have and in the interim you know we have70 people in vancouver you know a team at jpl still does rd and materialscience for us you know world-class superconducting appear and uh and you know in the meantime raise 80 or90 million dollars right which was bigger than the entire us budget for theso it's like it worked out some of the
technical things that you have toovercome as a builder of one of the machines requires that you build partsof your computer in ways that are entirely novel no one's ever done beforeso what if you're going to build a quantum computer has to be really coldand they the temperatures that you have to achieve are some of the tempostemperatures that we know of the entire universe and in fact the temperatureswill be achieved in our computers when operating are several hundred timeslower than the temperature eventual interstellar space and they wereachieving temperatures that they have no there's no nothing in the natural worldthat gets you down to temperatures like
this if there are no other intelligentlife-forms in the universe these are the coldest places in theuniverse when we operate the d-wave one computer the temperature of the chips islikely among for the coldest . in the entire universe there are other very difficult technicalproblems that we've had to solve the way one of them is the problem which is howyou program a quantum computer without disrupting its behavior so quantumcomputers have this strange seeing a paradoxical set of requirements that onthe one hand you as you were large hot blob of a lot of mass highly classicalin greece have to somehow tell this
thing that lives down in the very coldand very quiet what you wanted to do you have to beable to change what you wanted to do from run to run you don't always saythat solves a problem has to be programmable so how do you convert theseinto that and that is a an unsolved problem for every quantum computerarchitecture except he waves we solved that problem we solved it in a way whichis scalable to as larger computers you can imagine building of this search thatis was one of the biggest marvels of engineering brilliance and creativitythat i've ever witnessed the team that did this comprises the best circuitdesigners combined with the best
physicists available for this type ofwork and form they worked on this problem for many years before coming up with thisbrilliant solution its creative its exceedingly simple as all goodengineering designs are but it's a it solve the problem and what's even morespectacular is that the very first time we tried this idea that scale it workedand that given the complexity of this type of system isn't as an astoundingfact what ended up happening and how the course of the project went through andwe actually took the entire first month just establishing what the requirementsof the bumper and really understanding
that we were building but we needed tobuild that we understood what requirements were and that we were goingthrough with the designer was going to meet those it took a tremendous amountof work and coordination on everyone's part the group grew from about fivepeople to about 12 people by the end of the project new people coming on boardwe were talking about the constraints the design everyone was trying to helpout in whatever way that they could in order to accelerate that progress andfinally at the end of the project the whole system came together we installedit on the refrigerator and i watched two of my friends and colleagues basicallyput the system in and begin its
operation that was part of a project youknow that it gets handed off to the testing analysis that's when they canbegin testing chips and we had some goals and deadlines for when we need toget operating ships completed by and they then carry the torch and work lotsof all time and then ran through problems and they were able to achievethe goals that they needed by the subscribe time with six hours to spareso when you think about you know project as a whole if you actually build it andthen you begin to test in it it's not working you have to go through anotherdesign iteration your son so every every generation along that line has basicallyhad to work first time complete
that wasn't sort of magical magicalprocess we basically know but the approach we took at a time was to justdo a lot of rigorous testing during the build basically jack and recheck and anddip you do several cycles and dip things and problems and receive the great asyou're going along because once the system is built together you can take itapart you basically start from scratch again so we did a lot of rigoroustesting during the build phase and then we came together we cross your fingersand held our breaths and the system work the opportunity to do something that wasthat was unique that was new and and also and it wasn't just quantumcomputing it was actually quantum
computing the d-wave that reallyattracted me because it was it was an opportunity to do something with it wasa group of people that were really obviously serious about trying to buildsomething and trying to sell something and so i think that's that's what youknow that aspect of it fascinated me as much as the fact that it was quantumcomputing quantum computing is an amazing idea and it's not one idea it'sactually sort of a broad collection of ideas but it's basically is trying toharness aspects of quantum mechanics to do autumn to do computation in a way thatthat hasn't that hasn't been done before
at least that this is sort of a verybroad way of defining it the the people that that this problem is attracted youknow that our artists you know they're they're some of the most interesting andintelligent people that i've interacted with and understand it's just a pleasureto work with all these all these people so i mean that's the other thing thatmakes it really every angry sort of enterprise to be involvement