Quantum Computing—The What, Why, and When | SXSW LIVE

[Applause] all right hello um so if you uh you're steeped in science you're a PhD and nanotechnology but I imagine some people in the audience are like me and Quantum Computing sort of makes their brain hurt so I want to start with the very Basics great and ask you a simple question so what is a cubit tell me what a cubit is what is a cubit all right so for a lot of people in this room I think the best place to start is what is a bit right so let's think about the way that our computers today work right your laptop your cell phone data centers they work um using this fundamental unit of computing called a bit right and the easiest way to understand this is think about a light switch everyone can understand that you turn the light on you turn the light off there's two states that's exactly how bits work in classical Computing you can have bits in a state of one or zero on or off um and that is used to execute a lot of different logical op operations right things like and or if then you know what happens um in all of these cases and humans like are ingenious right over the last many decades they've been able to use bits to create all sorts of computation for adding subtracting calculus prediction models optimization even today a lot of the AI is all based on classical Computing so that's how a bit Works which is which is great it's transformed a lot a Quantum bit a cuit operates according to the principles of quantum mechanics so rather than needing to be in just zero or one Quantum bit can be in elements of zero and one at the same time so sometimes like this is where people are like what like how does that work how do how do you think about that how can I get my mind around and it's hard because it's not something that we experience it you know every day as something that that makes sense I think one picture that I used to think about it which can be helpful is thinking about a a globe almost right so think about the North Pole and the South Pole that can you're on you're off you're one you're zero Quantum bit can be pointed at the North Pole or the South Pole that's like your standard bit but it can also be pointed in any of these different directions and that gives it this zero and one character at the same time so you could have a Quantum bit pointed at Austin or California or Japan Australia um it has elements of zero and one at the same time and that's what can lead to some really fundamentally new properties why why does being in in two places at once in two states or multiple States at once why is that better for calculation yeah so one punch line that I would love for people to take home from this conversation right if you just took home one message it would be that quantum computers are capable of solving problems that are impossible for AI or supercomputers and so I'll say that one more time quantum computers are capable of solving problems that are impossible for AI or supercomputers even in the best case and you know we should dive into into why and how and and what kind of problems one answer to think about is this fundamental fact that nature itself is quantum mechanical so what does that mean right think about if you go outside right look at the leaves trees flowers grass um cells in your body chemical reactions happening underground those are all quantum mechanical in nature right they are fundamentally at their heart electrons atoms molecules all interacting that gives rise to all of the properties that we see in the world and that is quantum mechanics it is not zero it is not one it is states of zero and one at the same time and so right now what we've run into with classical computers is really a challenge right because people are trying to use zeros and ones to describe this very very complex language of electrons and atoms and and and we've made some good progress as as a society right we've been able to make approximations estimations but we cannot calculate exactly the properties of most molecules in the universe and the ability to use a quantum computer that fundamentally harnesses quantum mechanics is going to be a GameChanger yeah it's I wonder um if you could talk about so when something is a Quantum State and I think we'll get into how you get something into a Quantum state in the first place but when it's in a Quantum State as you described one of the things about it is that you can't observe it without it going out of that state so how do you how do you know what's happening there how do you actually understand how that calculation is happening when you you ask it to do something so to speak good question Reed so um I think this is important for the group to understand maybe um actually we can put a visual up on um the slide that can um help people understand so oh yes perfect thanks Chris um so you can see a comparison here on the left classical computer on the right quantum computer um so for classical computer right and again this is what most people are familiar with it's easier to understand um you could have n bit so in this case you know you've got three bits there each of those bits can take the value of zero or one and you can imagine setting up all different combinations if you've got three bits you actually end up with eight different combinations right they can all be zero they can all be one or they can you know take turns having one or or two um zeros and one so you get eight different combinations with a classical computer you are able to access all of those combinations either in series so you could set up three bits to run these eight um different combinations in a row or you could use parallel Computing right you can set up eight versions of these three bits and have them take all the different combinations with a quantum computer really getting to Reed's question um you're able to use uh two different properties which are really important the first is superposition which I talked about briefly each of these Q bits each of these Quantum bits can be in a zero and one state at the same time it's in a super position of zero and one at the same time it can take those flavors now where you really get get the power of quantum Computing is when you start to have multiple Quantum bits together acting in concert um this is called entanglement um it's the idea that you can get multiple cubits talking to each other really related to each other as you do that you can get um elements of all of the different states accessible in one state so what do I mean by that in the same way that you had eight you know opportunities to combine with a classical computer three bits there's eight possible combinations with a quantum computer but but if you entangle all three cubits you can get flavors of all states um in just one state at that time now you said read you know okay when you when you measure it then you know this is what's known as wave function collapse the wave function collapses and you can only get one measurement out you don't get all eight measurements out of the system you could only get one um and this is really really important Point uh which is often fundamentally UND uh understood incorrectly right so you make one measurement you have to set up your problem in a super creative way so um algorithms are designed to be extremely clever such that you increase the probability of getting the right answer at the end you decrease the probability of getting the wrong answer such that when you measure one time you hopefully have exactly the answer you want so let's start with you you um let's go back to like just how you create this thing you you Google Quantum AI just came out with this Willow chip as as we just heard um inside this chip is you know a number of quantum of of cubits essentially what are those things and how do you get them into the quantum State I mean I've been to your facility in Santa Barbara I've seen the giant chandelier that just freezes this thing tell me what's go what's going on there yes absolutely so um a couple things to think about right and there's a lot of different approaches to getting to a quantum computer and I think it's important for this room to understand also right there's different components so often times what gets the most um attention is the hardware right so that's what we're talking about we could dive deep into the hardware but you you can imagine that there's lots of elements that are required to make that Hardware work right so there's the chip itself there's all of the surrounding Hardware that supports in Quantum state of that chip um there is the software that controls the chip and tells it you know what computations to run and then there are the algorithms that really identify what are useful applications so when you look at kind of different approaches different um companies different um groups are really trying to play in different parts of that space what Google's approach really has been and and we're confident in this approach is to try to take what's called a full stack approach so Google has tried to work on not just the cubits themselves again which we can talk about but the cubits the full computer um all the hardware surrounding that the software and the hardware because we think that's going to be kind of the most seamless approach and so you have to you have to freeze it you have to freeze these things get them to to I think what is the temperature of like there's only certain it's colder than space essentially right that's really hard to do I imagine yeah it really is so of often times you know people say like it's cold it's colder than outer space you know well you came to our lab sometimes people say this is the coldest place in the universe we have to get down to specifically a 10 Melvin so that's um 10 Melvin just just above absolute zero um and the reason that we have to do that is because of the way that we're making our cubits so um maybe we can put um this yes exactly thanks Chris so you can see this slide here there's a lot of different approaches to building a quantum computer um and uh you you know there's it's it's important because this is a super hard problem how do you get to different types of cubits you can see several examples up here on the screen um Google is pursuing superc conducting cubits so we can you know dive into that a bit more deeply but um one way to think about this is you can essentially create a cubit out of anything that can be in a Quantum state right so you could use electrons you could use photons you could use atoms um you could use anything and and the real question question is there are trade-offs to what you decide to pick right every every option has its advantages every option has its disadvantages so you know you got to think about what's easy to make you know H how can you actually can you make this thing can you measure it can you control it is it uh reliable can you start to program it um how do you manufacture this on a large scale these are all types of questions and you can see examples um different companies are taking different approaches but um we're quite confident at Google in the superc conducting approach let's go into the the impacts of this a little bit now just to I think give us a sense we've gotten I think deep into the into the tech and what's happening but what can these things actually do that a classical computer can't how is it going to change the the world and I I love the stat what was it how many the calculations did the willow chip do you know compared to uh 10 septian which would so it would have taken 10 septian years on a supercomputer what took five minutes on the willow chip that was a benchmark problem it's older than the than Universe essentially older than the universe possible so you have this this power to do these these massive calculations what what are we going to do with it yeah so I'll give an example that is um really kind of personal to me and then we can talk about you know many other examples so 21 years ago my husband then friend was uh facing a very tough cancer diagnosis and for those of you in the audience who you know have gone through this yourselves or um have a friend or loved one that have gone through it's it's horrible it's very rough and he went through chemo he went through radiation nothing um was working uh and ultimately we were very lucky we were at Stamford at the time there was a clinical trial he got onto the clinical trial and um and it worked um fast forward to today he is a oncologist himself he's a cancer researcher and um he's helping to treat other patients right so that's amazing but the the problem is that um it's not a it's not a happy solution for everybody today right there are lots of untreatable cancers there are lots of untreatable diseases across the board right and the idea is if you could use um a tool if you had such a tool that could really get to the heart of what is happening right not just estimating what's going to happen or learning from past data but calculating fundamentally what is happening in these molecules in your cells what are potential drug targets that could actually make a big difference um that would be a game Cher it would you know increase our problems uh increase our ability to solve these problems and and help more people and that's just one of of many um reasons why people are so excited you know I think we've all taken you know biology in school we know what a cell what makes up a cell right I think there's I think for me it was a big eye opener when when people talked about modeling a cell that we don't actually have a mathematical model of a cell right explain why that is because you you hear all this stuff about Ai and and biotech right and we sort of have this sense that we've mathematically mapped out you know human biology but that's not really the case right what how how could you use a quantum computer to say create a mathematical model of a cell yeah um this a good question so let's think about um you know one case of a simple case of proteins right this is even you know simpler than a salad just you know a single protein um a lot of people have been very excited about a tool that came out from Google deep mine called Alpha fold so um when the Nobel Prize last year is really an amazing tool what does it do okay predicts how protein is going to fold and that 3D structure matters a lot for how that protein operates in the cell how it interacts with other proteins how it interacts with drugs it's awesome to have ai that can do that however Alpha fold wasn't able to do that just on its own Alpha fold had to train from roughly 100,000 proteins that were in a protein Data Bank that were painstaking data you know really collected over the past many decades by protein biologists you know crystallographers people that are working to actually get real data for AI to learn from and that's been amazing you know it's going to have enormous impact but there are many many many situations including in the cell as you said where we do not have data right we don't have any data for AI to learn from and the best classical methods are very um crude right they're making crude estimations because we cannot you know fundamentally again it gets exponentially complex as you increase the size of the molecular system to try to use binary um modes right zeros and ones to complicate uh and and compute exactly what's happening so the idea is if you can use quantum mechanics itself not be like overwhelmed like oh my God this is way too complex there's too many electrons we can't do it end of story let's just try to learn from data if you can harness Quantum Computing itself it operates according to Quantum Mechanics you could ideally really get these answers understand from the ground up Abino what is happening inside of a protein what is happening inside of this all right and and it's also not replacing classical computers right I think that's something that is is important to understand as well this is not like a you're not we're not just all going to have quantum computers in our pocket at some point yeah I'm glad you brought this up I really want to make sure people understand this point too you know the two types of computing are going to be complimentary to each other and sometimes we'll hear okay quantum computers it's it's going to be our new form of computing so in 10 years 20 years it's going to replace everything we have today um the answer is no you think of quantum computer as yet another tool in your toolbook yeah so it it really will change the world in fundamental ways um but but we we can't do it yet so what is the challenge why why is it that you know you've been able to demonstrate such amazing things with Quantum error correction uh which which I'd love for you to explain as well but like what do we need to do to actually build you know an ATS scale quantum computer as that as they say yes when you think about that question it's important to think about where are we trying to get to like what do we need in a quantum computer in order to solve all of these problems we care about in chemistry and biology and and others right and um the best estimates today um really H can capture this in a couple of metrics one you can think about the number of physical cubits that's required so as I said you know it's just great to have one cubit but one cubit on its own isn't going to do much right 10 cubits on their own they're not going to do much they important stepping stones but in order to run problems that we care about like in chemistry and Material Science it's expected we're going to need on the order of at least a 100,000 a million or more cubits so one thing that's important as you ask like how do we get there is you've got to think about how to actually scale to build um reproducible reliable sets of cubits that are not just on the order of 10 or 100 which I you know I don't want to undersell that it's been a very very hard road to get here but you've got to get to a larger scale quantity and that includes not just the chip itself it includes all of the surrounding Machinery you've got to take a 100,000 a million cubits down to 10 millin scale how do you get refrigerators that big that powerful um these are really important problems maybe the second thing I'll comment on and you can talk more um is it's not only the quantity of cubits that matters right sometimes people will say yeah this is the quantity everybody knows you know it's not just quantity that matters it's quality you could have a bajillion cubits that size and you're not going to get anything useful out of it so what is a measure of quality one thing that is very important to look at is what's known as the error rate um so how many errors are these cubits making um A good rule of thumb again from the kind of current estimates looking at what we need for chemistry materials is at least on the order of 10 to Theus 6 logical error rate so what that means is you're making one error one mistake in every 1 million or more algorithmic steps uh now with Willow as you mentioned we had you know an enormous breakthrough and we can talk about that in more detail um this is the very best error rate reported to date in a Quantum Computing system our error rate uh is uh 10 to the minus three logical error rate of 10 the minus 3 that means we have one error in every 1,000 algorithmic steps which again has been enormous amount of work to get there but that is not one in a million so we've still got three orders of magnitude or better um to go and that is a lot of really hard work and I think it's fascinating what actually causes these things to make errors right why do they make why do they make errors there's a lot of reasons that there are errors in Quantum systems right one you can imagine is just it's it's it's fragile it's a fragile state um it can be um disturbed by energy by Noise by vibrations in the environment and so you know you mentioned these giant dilution refrigerators actually we can put the slide um back up on the screen um so people can get a picture if you look at the far left on superconducting cubits the image um that we're showing on the screen is of one of these sometimes people call them chandeliers they're called dilution refrigerators they really are the tool that gets um our Quantum chips down to 10 Melvin so each of these flat stages that you're seeing it goes all the way from room temperature down to 10 Kelvin 3 Kelvin then all the way down to 10 Melvin and the chip is screwed on in a package um to that coldest stage um now why does it have to be so cold well that is we're superc conducting cubits start to get their Quantum nature so at that temperature they take on Quantum character and we're able to actually get the cubits that we care about now for those who may have seen we end up shielding these um refrigerators right so you can see it unshielded now but we put these um different types of cylindrical Shields around that's really to protect Quantum chip from noise protect it from the environment you know unwanted microwaves uh cosmic rays things like that so that we really get um the computation yeah Shield it from like so basically tiny tiny particles that we can't even see are passing through the Earth constantly and they they often slam into these these Quantum these cubits essentially right exactly and and it's important because this is the um kind of the trick it's one of the hardest parts of quantum Computing is you want the cubits to maintain their fragile state but you also want to be able to manipulate them so you do want them to be sensitive to the environment so you know we send these you saw in the image we have all these control wires that go in you want to be able to talk to the cubits you want to be able to read out what they um processing you want to be able to adjust their state but only in the way that you have full control that's a trick uh I'm going talk about policy a little bit just to to change gears I mean a lot of this came from basic research right funded by the US government other governments around the world um and that's sort of what what got us to a point where a company like Google can can take on a a challenge like this um how important is that how important is to to continue that do you think it's critical basic research is critical so you look at you know even if we talk about classical computers right the transistor you know that was all at its heart funded by um US Government fundamental research investments in science in in in really understanding at its basic level how do electrons behave how can you set up materials that have such properties um that is what has given rise even to in a lot of the we have in Quantum Computing today it's all based on basic science um even you know in what we're building at Google we learn every single day from the work that's happening in academic research Labs we're always reading the literature we're looking at papers there's new advances you know no one company is going to do this on their own no one team can do it on their own you've got to have the whole community of researchers and government investment in this is critical part so and we have China which I think is outs spending the US 2 to one on Quantum research what does that mean I mean we're we're in this technological race with with China yeah there is really good research coming out of China really good advances right so if you look I think even um at you know recent announcements we seeing a lot of progress in Quantum Communications in Quantum Computing all coming um from China and and look you know you want to make this you know case you talk about government investment let's talk about you know the Human Genome Project or um the race to the moon there's numerous other projects these would not have been possible by scientists alone absolutely not these would not not have been possible by one company alone absolutely not they were required really government investment and also International diplomacy in many cases to advance Technologies um and that's exactly what we expect to be required here and you worked in the White House Office of Science and Technology policy under Obama right how do you think we've seen a lot about you know basic science funding right now in the in the current Administration how do you think they're doing with regard to Quantum um you know I think it remains to be seen um a couple points I will raise you know the national Quantum initiative in the United States and this was um announced in 2018 it was actually under the first Trump Administration so we saw a lot of support for Quantum Computing and Quantum technology broadly even in the first Trump Administration which is great to see um and a lot of progress that's been made since then continued under President Biden um so you know that was wonderful I think with the um Trump Administration coming back we've heard Quantum Computing is a priority right I think the government recognizes both from an economic as well as a national security standpoint again we could dive into some of these in more detail this is a priority um but it remains to be seen if um the government is going to be able to play the role we need to what do you think should happen what do you think the US should do to get ahead to stay ahead yes uh a couple of things that um would be uh you know our recommendation from Google I think um the first thing is for of course the US government to really invest and you talk about investment what does that mean okay invest in basic research invest in engineering invest in you know collaborations okay what is required well we've got to make cubits again that are um higher quality we've got to scale the number of cubits what kind of research does this look like what looks into research in in new materials you know are there better performance that we can get across the board this is basic research how do we think about engineering there's all these different components beyond the chip itself from the wires the electronics the software how does the US government invest in that type of research and invest in that type of talent really you know we're going to need an entire Workforce here um a lot of work and coming from uh the National Science Foundation the dod the doe all of that work is going to be really important to raise up um the workforce that's needed maybe the last thing I'll mention is supply chain um this you know comes up often in you know today's semiconductors it's absolutely also the case for quantum computers so um again all of these different components they're not made in one single country right so you need the government to really step been to play a role of diplomacy to play a role of bringing different people together across Industries across countries to build um what's going to be a very complex and ambitious project and and why why does the US you know want to why should it win this race what would happen if it if it doesn't yes uh so two things I could talk about here I think number one and we talked about you know the ability to really understand nature fundamentally right and and that solves I talked about you know cancer research and opportunities for new Pharmaceuticals of course that's one area but if you understand nature fundamentally that has implications for lots of areas right for better batteries better solar cells maybe Fusion Energy um you know better chemical reactions more efficient energy usage across the board you know some people will say climate change is is the climate crisis that we're facing is really the reason to pursue Quantum Computing so if the US had the ability to do that that would be an absolute game Cher the second reason which many people in this room have probably heard about and we could talk in more detail too um is that it is expected um thanks to an algorithm from Peter Shore in the 1990s that a powerful enough quantum computer um could solve factorization of large numbers that means it could break the encryption of a lot of what protects um secure Communications today and we should talk about how that works um it's important to note that that's expected to take a computer that is one or two orders of magnitude more powerful than that would be required for chemistry materials so it is farther down the road um but you can imagine that has massive National Security implications and you would want the us to have that capability yeah and actually let's talk about the timelines here too I mean uh Sundar Pai Google CEO has said he expects this to be affecting it affecting Google's business within five years is that tell tell us like does that mean we're there in five years what's the what's the trajectory here yes so we expect within the next five years to have a useful application using quantum computers right so you talked about this Benchmark problem that we ran um again for the audience it's called RCS RCS Benchmark um we were able to run this RCS algorithm uh on a willow chip in about five minutes it took um an estimated using best supercomputer at Frontier uh it was calculated to take an estimated 10 to the 25 years 10 septian years which is basically impossible you can't run that on classical computers um but that's a benchmark problem right it's not designed to be useful um the real thing that people are looking for is what are the useful applications as Sundar said we expect within the next five years to demonstrate at least um a if not a few useful applications and we expect it continue even beyond that so it would be sort of like like give us an idea of what that might be like it won't be solving the the mysteries of the universe kind of applications it'll be more what like helping optimize you know ads or you know recommendation algorithms what are we what are we talking about that would be good for Google yes um you know look I think our our best guess right now based on where the technology is going and um what we can see with our own systems is within the next few years an example would be um being able to solve uh the structure of a molecule that is impossible to do by any other means so that you know would have enormous implications you can think about molecules that play a role in biology in chemistry in um you know batteries these are really important types of things that you could imagine many different Industries would care about that's something that we expect quantum computers to be able to do yeah that makes sense one of the one of the things that got a lot of attention in your in your uh your paper tied to the willow chip coming out um in nature was that you talked about this proving essentially that parallel universes exist and we've all you know probably all seen some of these Avengers movies with the quantum Universe um multiple you know alternate universes what does that actually mean what what are what should we take away from that yeah so for people who aren't familiar there is a concept called the Multiverse uh not to be confused with the metaverse the Multiverse um which is really this idea that there are multiple parallel universes all happening at the same time now this has not been proved um the willow chip doesn't prove it by any means um but it is one explanation right you look at um a lot of physics today and you actually see over the course of history physics and philosophy have always been very closely tied together because physics can tell us what it doesn't necessarily tell us the why um and you know I think a lot of people can see you can see the way that electrons behave you can see the way that atoms behave um it doesn't necessarily make sense in our minds why does it behave that way some have hypothesized that um Multiverse is the reason why it behaves because you know you're you're seeing combination of zero and one while it's actually testing multiple combinations in all these parallel universes at the same time now in our Willow announcement and this was in the blog post it said like look you know the fact that Willow can run something that would take otherwise you know an estimated 10 septian years um that is you know potentially some evidence that what's happening well all these computations are happening you know in other universes possibly maybe that is um it it doesn't contradict that idea but it doesn't prove it and and maybe to to get a little philosophical here I mean one of the things that's sort of interesting about this is like we don't actually have a unified theory of physics right that explains all of this so we are sort of we understand that this is useful but we don't necessarily understand it the thing so how do you reconcile that it's like a you know you're you're putting information into this computer and you're getting out these amazing results but you really have no idea it's it's might be going into parallel universes to get the answer I honestly think that's one of the most exciting things about being a scientist you really realize like the world is so amazing like this entire there's so many Mysteries we do not understand and that's what makes it exciting to come into work every day like what new thing can I try to even you know wrap my mind around a little piece of it you know to your question you know maybe I make an analogy like most people here probably know how to drive a car right do we understand all of us the inner workings of how that car Works no we don't don't but we absolutely use it to get from point A to B it's useful and there are ways to test it ensure the reliability ensure that it's secure there's a lot of tests we can run I think it's very similar you know with the mysteries of the universe you think about a quantum computer does anybody in this room fully understand how it works no there's a lot of mysteries we don't we find that it can be useful and if we can set up the right tests to make sure that it is safe that it is secure that it's reliable giving us what we need then absolutely let's use that as a tool yeah I think about it is like if you if you thought about scientists in in in history right who didn't have the instruments that we have today who didn't even understand you know that we are a round thing in the middle of the Universe um and didn't know what stars were I mean can you think like is is there analogy like an historical analogy for what we're sort of you know on the cusp of with Quantum yeah I think there's like many such analogies right even think about let's take one simple one in in recent history the Human Genome Project right that's an enormous undertaking um led by the us but with collaboration countries around the world to really try to understand can we solve the sequence of the human genome at least to a certain degree um and use that to to become you know actually in information that impacts our everyday life and that's had enormous implications does that mean that scientists today understand every single bit of the last you know every single DNA based pair in the human body no do we understand how that interacts with all the different cells and how you know chemicals can come up and and modify the DNA no we don't understand all of it but it's increased our understanding and we're already using it for useful applications so I I was at when I visited Santa Barbara I was talking with harmat who runs who runs the whole thing and and he said you know I was asking him similar questions like how do you do this how do you scale this and he said you know if the aliens were invading we could we could scale this today and I think what he meant by that is if aliens were going to invade we would need some pretty advanced technology to defend ourselves and this is how we would get there but he I I what does he mean by that I mean could we actually just do this to sort of Brute Force this this right now and should we yeah so I it's funny harat's a very colorful person so um he has lots of good kind of analogies and I think that you know gets at this sentiment of a lot of the science has been proven out right 30 years ago it was this crazy idea could you build a computer based on quantum mechanics and use cubits and and do some calculations it's okay it's great in theory today over the last many decades I think a lot of work has proven out that this idea actually is possible it is probable and we can make it happen how do we make it happen well there's a lot of hard you know fundamental engineering is what Harman was you know getting that but if all of us really you know focused and we kind of prioritized this problem and tried to work on it um we believe this is a solvable problem so the we're going to get to some audience questions I think the last thing I'll ask you though is I think an important important thing for for everyone here is you there's a lot of hype out there around Quantum and there are companies today that say they are are actually using Quantum Computing right now um to solve various algorithms um how do we think about that how do we separate hype from reality and how do we you know should should we be cynical skeptical you know help us think about it yeah very important question uh I will give this uh audience a couple of tools that I would use um really to assess claims right as you say Reed there is a lot of hype out there right there's a lot of um announcements there's a lot of frankly some misinformation out there also which makes sense because it's a very difficult um you know field to understand if you're not in it day in day out um even if you are it can be difficult to understand so you know what are some of the tools that people can think about um I would list a few so number one um if you're looking at you know some news or some claim that is out there how many cubits are there like let's look at the quantity of cubits again because as I said to get to practical applications best estimates you're going to need something like 100,000 million or more cubits so does this claim have even one cubit um are there 10 cubits 100 um that gives you some measure of how far along we are on the way to getting you know practical useful computer um I talked about also quality of the cubits right you can look at the error rate at what um stage is this claim you know to what degree does it get to the one in a million or better error rates that we need how far along is it that's an important thing to think about third thing you can look at is actual performance like what can be run on this chip and we've talked a couple times now about this RCS Benchmark that's almost table sticks can you run this Benchmark algorithm and how does it compare to the best classical supercomputer okay awesome but Beyond The Benchmark what's possible on this chip what can be performed how quickly does it do the you know computation that's an important thing to think about maybe the last thing I'll really encourage this group to think about um and I do this myself especially in areas where I'm not an expert is what are the experts saying right there's a whole community of experts that have decades really looking at this problem understanding it thinking deeply about this what do experts say about this and that can help inform your own view thank you so let's do where where do we oh they're right here so uh let me read I'll just start from the top so how can we ensure Quantum is developed safely and transparently are there open- Source initiatives or public forums to track progress and mitigate threats yeah thanks for the question so um I think it's very important to ensure that it is developed safely and transparently and one of the ways that this is done today is through peer-reviewed research so you'll notice even over the last year a lot of papers published in journals like nature to really ensure that there are um experts looking at this that there is scrutiny um you can see a lot of transparency in that reporting facts and really being able to test itself um are there open sources initiatives there are many such initiatives um for example um we've got a couple open source programs at Google ourselves but this is really where the field is trying to go um to be able to track progress as a community and see how the field is doing this is a good one given that both quantum computers and AI work in a probabilistic manner should we ultimately look forward to Quantum AI yes so the name of uh the team at Google trying to build the quantum computer is is actually Google Quantum AI is for um you know very um similar sentiment as raised in this question at the end of the day we really believe that quantum computers again are going to be able to do things that AI cannot if you can combine Quantum and AI together you envision that you can get a lot more kind of a lot more benefit out of that so here's somebody who's wondering whether they should sell their Bitcoin how do we best prepare uh oh sorry did that go away that went away um sorry there was a Bitcoin question maybe it's the next one um oh yeah so based on the timelines will Quantum get us cures for cancer before it can break Bitcoin encryption sorry was that y so um yeah I talked about this briefly and that is really according to the best estimates uh in the field today um it's estimated that it'll take on the order of about a 100,000 or a million physical cubits and logical air rates as I said 10 Theus 6 or better to get to chemistry and materials um you know being able to break encryption it's expected to require a quantum computer that has has 10 or um 100 times right an order of magnitude two orders of magnitude more powerful than that for chemistry materials so answer to this is ideally yes so how do we best prepare the kids in school right now for a Quantum world that seems just 10 years away I love this question uh I have a 5-year-old daughter I know you've got kids to read and um I love it you want to think about what's happening in the future right it's not the same as what we learned when we were in school um it is how do you prepare them I think there's a number of things to do obvious viously you know math science um Humanities all very important I think it's important to think about a world of possibilities right how do you have the skills to really um try to understand how you could use a quantum computer what are the applications that are out there really understanding um different opportunities and being able to dive deep into that would be awesome uh so what are the main concerns regarding the um eth ethical implications of it how can we ensure its benefits are not limited to a few major companies uh I guess they're talking about Google um and or government yeah um I think this is really important question you know like any um new technology Right comes with opportunities it absolutely comes with risks so what are the ethical implications you know what could you imagine happening if um a bad actor had power over such a technology how do you think about the security implications there's a lot of questions like that I think one of the best ways to do this and you know we've seen this you know starting to happen with AI I think you know with Quantum Computing we could get ahead of it actually move faster let's start to have more public discussion about this let's have more people involved really understanding what the implications of this technology are how it can be used and getting involved in how it is used um and thinking about applications um so um I heard there are people using Quantum Computing to make market stocks market stocks predictions optimizing logistic routes exploring new materials th are those cases is real that's what we were kind of alluding to earlier yeah so I think it is real that people are using quantum computers for these problems I am not sure it is real that they are getting useful answers for those problems today's quantum computers so um really want to make that case right to date there is only one known algorithm that has outperformed a classical supercomput and that is the RCs benchmark stands for random circuit sampling again awesome it's a benchmark problem um this again you know is is really important measure but that is not something that is used for useful applications today all of the applications that you've seen that are here in this question and also you know I I I want to give credit there's really interesting science experiments that are being done on today's Quantum chips that is awesome but nothing to date has been out of the reach of what you can do on classical computers can actually just those science experiments those are can you just quick tell us what those are cuz those are really interesting yes yeah so um you know I think even if we look at what's possible on Google's Willow chip our previous generation was called Sycamore um we have collaborations with a lot of scientists um you know outside of Google and we've been able to have really interesting new discoveries and things like magnetism um discoveries of um systems like time crystals and a lot of different understanding of how are you know electrons even situated in this in this um you know 1D material do we think about fundamental physics I there's a lot of interesting answers you can get from Quantum chip today now again underscoring you can also get these answers with classical computers today this is just another tool to get is another instrument right to get the answers um we're continuing to do that we're excited about it you know our collaborators are very excited about it but we really look forward to the day where it's possible to run things that are impossible otherwise yeah um uh so how can we think Quantum under understand what is C what it is capable of doing how can we evolve our binary mindset yeah this is hard I you know you think about your you know however many years you've lived on Earth and your everyday experience is very binary even our you know our phones our computers everything that we think about it's a very um binary way of thinking and you know there's a couple of different ways you can do this I would say like you know even you go on a hike I would you take a walk I love you know going going up in the mountains in California and you just you know breathe the air and you're look around there's everything around you is made up of quantum Computing if you could really look at that appreciate it and start to kind of dig deeper whether it's through books whether it's through videos or anything else right just doing your own understanding of like how does this work what is here fundamentally it'll kind of Open Your Eyes in a very different way so recently Microsoft announced majorana one as a breakr in um uh in development uh in developing quantum computers how does compare with Google's uh superconductor process yes so Microsoft announced this a couple of weeks ago and uh it's been a longstanding problem for them I think you know they've been working on this problem 17 plus years um so it was really good that they were able to release something now um you know the claim here is that there is one cubit I think that's a myana cubit um but that has not been proven yet by public data so that is something the whole field is very eager to see from Microsoft I think we heard um you know these experiments are consistent with potentially um one myana cubit um and they said that they've got data to come in the coming weeks so we're all really looking out for that um you know in contrast to Google and this was the question um Google's Willow chip has 105 cubits um we've demonstrated a logical error rate of 10 to the minus 3 that is one error in every 1,000 computational steps so these types of things are really important to be measuring as you're looking at progress thanks for helping me with the pronunciation there um can Quantum Quantum Computing potentially make it possible for us to travel to other galaxies and even test the theory of the Multiverse to bring that up again maybe uh I I think that remains to be seen um okay you've you've talked about us and China is there a specific role you want Europe to take on in the quantum scene yes so Europe has a very important role to play I think we've been really excited to see kind of the EU coming together on a lot of these kind of quantum initiatives and specific countries um including in the Netherlands and Denmark and Finland really taking a role here um look I I want to underscore this right quantum computers is is not just a you know one computer there's tens of thousands of parts that go into this right again no one country is going to be the expert of all of those parts even at Google ourselves we collaborate um with over a 100 institutions in a lot of different countries including in Europe as well as Asia and Canada really to bring this quantum computer to fruition right so Europe has a lot of expertise for example in um cryogenics getting super low temperatures Europe has expertise in wiring um Europe has a lot of expertise in algorithms and and other so yeah there's a huge role to play um this one I I thought cubits have three states one zero and both but it seems there is a scale can you elaborate more on the on the error rate elimination and observation bias I guess those are kind of two questions yes um so let's start with the um let's start with error rate so three okay maybe we'll start with the first part of the question so three states so the Cubit actually has if you think about a Quantum system and the way that you want to set up a cubit it has two different states that it can take character of both right so it can be in zero or one or some combination of zero and one and I I wouldn't call that a third state I wouldn't call that both because it can really be you know mostly one or mostly zero or halfway zero and there's really like a huge set of possibilities in between those two so you think about a two-state system um now interestingly and I I put this as an aside but there are people that are looking at three state and four state and higher Cubit systems I guess that's called CIT or you know Q quadrant so essentially like a zero one two three and then can you think about you know uh states that are kind of in between those it's very exploratory research but um yeah for the types of cubits that Google is working on we have zero one and then a kind of linear combination of of the two um can I elaborate more on error rate elimination so this is um actually good for this group to understand right every single Cubit inherently has error in it we talked about this a little bit right so um every Cubit can have you know whether it's due to some imperfection in the material itself whether it's due to um you know outside interaction or interference you know there's error inherent with it and when you think about scaling up a quantum computer right this has been a long-standing problem for decades people say all right if you just keep adding more and more cubits you're compounding the error if your one cubit has error and you're adding more and more you get to 100 thousand they all have that same error you're actually introducing huge amount of error into your system and um you're not going to be able to do useful computation so one of the things that the field has really been working towards and this was part of Google's Willow announcement in December is called Quantum error correction so how do you you know the errors are going to be there is that's just a fact of nature we can keep trying to push it but there's always it's never going to have a perfect Cupid so how do you correct for the errors that exist um this is really borrowing an idea that also exists in classical Computing rather than having a standalone Cubit on its own um what we've done and others in the field have been working on this is actually can you have multiple cubits working as one logical single Cubit together to protect for the individual errors that are arising so if you can have one protected error um one protected Cubit that can actually get decreased amounts of Errors you add more and more cubits but you protect against it kind of have some redundancy built in there um you could actually have a much higher quality cubid um and that's what we've been able to demonstrate and it's really important direction for the field do you want just on that topic do you want to address the Amazon recent error correction um or I guess it's almost like an error reduction method where they have this Cubit that kind of oscillates I guess is that is that interesting uh yes it is and and I really want to give credit right to all of these um advances because you know I underscoring this is a very difficult field this very very difficult topic it's not easy you've got a lot of hard work from multiple people in the lab you know over decades so every Advance really um should be celebrated and and recognized um Amazon and others really looking at this idea of error and you know how can you start to mitigate some of the error that's there if you can't correct it completely are there ways to reduce it even just a little bit um and these methods are going to be important too that's interesting um what could be the consequences if one foreign Nation like China gets a high functioning quantum computer before anyone else we kind of touched on this earlier but if you have anything to add to that yeah we did I think nothing you know more really to add essentially just think about the opportunity that is here it's enormous we really want to think about how that can be um used in exactly the right way what would be your um first priority sluse case if quantum was was ready was ready to use secure reliable and safe yeah okay so speaking to me personally I would love to work on cancer drugs I I that's is the thing I'm very very motivated by I think that would be amazing to be able to really understand different kinds of systems that cause cancer in the body and um new types of materials and molecules that can actually help solve that I think that would be amazing but there's a lot of other very important things to work on um for example the energy um opportunities and solutions better batteries better solar cells nuclear fusion um these are all huge things that could change the world so would Quant to me anything for the creative industry um audio visual content um any of that stuff yeah actually this is a great question it's interesting so um recently actually there's an art show that started in Berlin is a collaboration between our team and um uh some artists and it is really this idea around like okay what can Quantum do so most people in this room may be familiar with generative Ai and the use of AI to you know create imagery and you've seen amazing collaborations between Ai and artists and using it to create all sorts of new things um so one of the questions that we've been asking Rec recently is about this idea of quantum noise all right so Quantum noise you can get from you today's Quantum processors you know previous generation Sycamore generate Quantum noise you think about um Can this Quantum noise be actually used Fusion networks to generate more natural images more interesting images things that somehow appeal to the human eye because you know we're accustomed to seeing things in nature that that come from Quantum noise um and in this art show actually um there's a dial that um participants and and viewers have been able to see like kind of turn up the quantum noise versus turn up the classical noise get the image generation and qualitatively we've seen that a lot of people prefer the large Quantum noise giving rise to the image um so that's something that we're really pushing on and kind of excited to see where that goes wow that's fascinating um is this technology going to be accessible to the public to the public like a like a server on gcp where I can just like use a quantum computer remotely yeah we certainly hope so this is really the ultimate goal right now is very much kind of initial experimental prototype phase almost right and as I said we've got collaborations with a number of academics but this is very kind of selected use cases things that we think are going to be you know the most chance of having something meaningful happen on the computer but longer term absolutely this is something we want to open up and make accessible to people so what would be what would be the impacts on market dynamics and pricing yeah this is a big question I think we hear this a lot especially from people in financial services you know how do you think about this now we talked a lot in this conversation about Quantum simulation about the ability to use quantum computers for chemistry for materials that seems like a you know clearcut case a huge area of exploration

2025-03-16 00:24

Other news