Subscribe
Share
Share
Embed
How do we begin creating the future we want, today? Design This Day takes you on a journey to our future world. Futurist Devin Liddell sits down with visionary leaders from some of the biggest names in tech and innovation. Each episode features a brilliant mind who is building the opportunities of the future before most people even know they exist. What will living in microgravity in space look like in the future? Can driverless vehicles go off-roading? What unexpected roles will robots play in our future workplaces and homes? We explore the role that design plays in shaping our future – with the big thinkers and doers who are creating tomorrow, today.
Design This Day is an original podcast brought to you by Teague.
About the Host: Futurist Devin Liddell
Devin Liddell is the Principal Futurist at Teague. With over two decades of experience in innovation and design strategy, Devin has worked with industry giants like Boeing, Intel, and Nike, helping organizations anticipate changes across both near and far-term horizons to create their preferred futures. Devin is a frequent contributor to Fast Company.
Get in Touch:
Have a complex problem that needs solving? Have a great guest or topic idea? We want to hear from you. Visit us at teague.com or send us an email at hello@teague.com
Michael Brett:
It's a really exciting technology, it deserves a little bit of hype, and there's this constant tension within the community of the dangers of too much hype like what if we overpromise and under-deliver and then there's no longer any funding for it and we enter these dark times of not being able to progress because we overpromised too much? And then we look at what's happening in AI right now and go, "Okay, we're not really in any danger of over-hyping something if that's the benchmark."
Devin Liddell:
Welcome to Design This Day, a podcast about the futures we want and the people working right now to make those futures real. I'm your host, Devin Liddell. I'm a futurist at Teague.
This year we're celebrating Teague's 100th anniversary, so I'm thinking a lot about the challenges of today and how we're going to solve them over the next 100 years.
In today's episode, we're talking about quantum computing, a technology with a promise to change how we solve some of the world's hardest problems.
To help us understand what that future actually looks like, I'm joined by someone who has spent his career at the application layer of this technology, Michael Brett.
Today, Michael is the worldwide go-to-market leader for quantum technologies at Amazon Web Services, but he got there by way of aerospace engineering, a defense computing career, and co-founding one of the earliest quantum software startups.
That background in software, not just research, means Michael thinks about quantum computing in terms of what it does, not just how it works. He's the perfect person to help explain what quantum computing makes possible that wasn't possible before.
Michael Brett:
Yeah, so there's a couple of different classes of problem that we look for. The holy grail ones are the ones that we just cannot solve today. So these are problems that our customers do not even approach, because on a classical computer it might take 10,000 years to solve or something like otherwise intractable that we can't even contemplate. Or just so expensive and so resource-consuming that they don't even solve it or they sort of avoid the problem and maybe use heuristics or some other technique to solve it.
So there are those unlocks, like the brand-new applications, and those are the really big ones that we're aiming for. Then there's other applications where we would get some performance improvement over the alternative. So they are doing it today on classical, but can we get a faster, better, cheaper outcome by adding some quantum compute into that workload?
And so it's really trying to find that sort of price performance advantage where one part of the workload, if we can allocate that to a quantum computer instead, can we accelerate the entire problem set and get a benefit from that in some way?
Devin Liddell:
And so to sort of ground our listeners in quantum computing, can you give as a conversational as possible overview of what quantum computing is and where it's likely headed? And I'll just put the asterisk there that quantum computing, which extends from quantum physics and quantum mechanics, famously sort of hurt the brain of Einstein, right? This sort of vexed Einstein. So there's a lot of complexities here, so would just be curious how you describe that.
Michael Brett:
Yeah, yeah. We've got a lot of advantages over the time that Einstein had and that we've had about 100 years now of trying to figure out how to explain this thing. So we've had more time to come up with better explanations that can help us get through this.
So quantum computing is an emerging technology. It's a very early stage technology that's in development. In some ways it's just another computer. It's a computer that we might use to do some computational work that can help us solve some algorithms and solve some problems, which is what classical computers do. They do it in their way. Quantum computers do it in a different way.
And the unique properties of quantum computing are that we're using quantum physics, a different kind of physics to classical computers, to do that work, to do that computational work. So we're encoding information within the quantum physics. We're doing some operations on that information and getting some result from that.
And so a different kind of physics is described by a different kind of math. And a different kind of math means we can have different algorithms, and different algorithms mean we can solve problems in different ways.
So in some ways it's just another computer. It does some interesting computational work. In other ways it's really, really special because it's using a different type of physics, and that can unlock some really new and interesting, unique ways to solve problems that we're excited about, our customers are excited about, and I think is worthy of investigation and further investment as well.
Devin Liddell:
Basically, rocket science was just not complicated enough for you. You just needed something more complicated.
Michael Brett:
It had been done. It was time for a new adventure. Yeah.
Devin Liddell:
Obviously we're living in a moment where certainly AI is, if anything, over-hyped. It's not under-hyped. But from your vantage point, is quantum computing under-hyped at the moment?
Michael Brett:
So yeah, some hype is appropriate. I think it's an exciting technology. I think what's going to be interesting about quantum computing as it becomes capable of running these big commercial workloads and create value is will the average consumer ever really notice that impact?
Because the customers that we think we're going to have for quantum computing are running these big industrial-strength workloads. They're pharmaceutical companies, they're aerospace companies, they're doing material science at a really deep level.
And the consumer will see the products of that, like the result of it in that their 401(k) might perform slightly better, they will get different technologies in their phones that they're not sure how that material's created, but it somehow works better than it did before, or their aircraft might fly slightly more efficient than it did before.
And these are all profound things that will have a big impact on the economy and products of the future, but will the average consumer ever really know that a quantum computer was involved in that process? We're not sure. We'll see how that plays out.
Devin Liddell:
And that's actually maybe a little bit of a comparison to artificial intelligence as well, and in the sense that it seems likely that even in five years' time, AI might in some ways disappear into a broad computational power type of domain where you don't regard an experience as sort of AI-enabled or not AI-enabled. It's just sort of baked into all of these things.
Michael Brett:
Yeah. And there'll certainly be transformative things and some hugely impactful technologies that'll come from this. I think we can sort of, when we think about what quantum computing might do in the future 10, 20 years away, there are possibilities that this could get involved in really specific drugs that can help all sorts of maladies that we suffer as humans.
That could be hugely, hugely impactful and unlocked by quantum computing, but it's one ingredient in a broad portfolio of innovation that needs to happen that enables that, so yeah, we'll see.
Devin Liddell:
At AWS, Michael's work centers on Amazon Braket, a service that has made quantum computers radically accessible through the cloud.
Since 2020, any developer, researcher, or company with an AWS account has been able to run workloads on a quantum computer. Most of Braket's users are researchers probing the technology's limits and figuring out what it can do, but interest is starting to spread into the private sector as well.
Michael Brett:
Yeah, so we've got a whole range of startups that are building software on top of AWS and on top of Braket. These are companies that see an opportunity that a new technology can provide to differentiate themselves in the market, provide a new capability to software products. We do a lot of work with startups in this space that see the potential.
And then on the enterprise side, we've got customers like JPMorgan Chase that see an opportunity with quantum computing. They want to be first movers into this space as the technology improves, and so they look at a wide range of different algorithmic techniques that are relevant to the bank. It could be in options pricing or portfolio optimization or in fraud detection, looking for these kind of edges that they can get that a quantum computer could provide through algorithmic performance.
Devin Liddell:
You mentioned earlier that this is accessible to anyone. So anyone with a credit card can access a quantum computer through AWS.
It raises the question of course of, for example, if I were to do so, would I know what I'm doing? Right? And you mentioned your background in software engineering. My question around this is less about my incompetence in terms of using the computer and more about is there a software layer that's required from a design standpoint that's actually different than the software layer in terms of how we would experience traditional computers?
Michael Brett:
Yeah. So one of the principles behind what we build at AWS is to make it consistent with the rest of AWS. So our quantum computing service bracket has a standard AWS ABI, there's a software development kit that's written in Python that customers use. So it is very accessible to the average software engineer, developer, your coder, your hobbyist. These are tools that they're very familiar with already and can also get a lot of support with through the various AI coding assistants that are out there.
So from that sense, very accessible. This is kind of a normal kind of environment to work in. There are also a huge variety of open source third-party software toolkits that we integrate with, we help support. And we're really lucky that there's this huge developer community out there of open source enthusiasts that have built tooling for quantum computing over the last decade or so.
So there's a particular toolkit that is specialized for quantum machine learning. So if you are a machine learning enthusiast and you know machine learning on classic compute, well, great, you can pick up this toolkit, it's called PennyLane, and it's got all of the machine learning algorithms in there.
Devin Liddell:
I love the name, by the way.
Michael Brett:
Yeah, it's great. And there's a variety of different tools out there too to help people do that. So it really is accessible.
And then the tricky part is working with the hardware that we've got today, there's only certain capabilities that can deliver to a certain capacity, and so often you have to fit the problem, you have to make trade-offs and cut things down and kind of fit a square peg into a very round hole and make it work. And it's those trade-offs that are really tricky.
So we often joke that the hardest part about programming a quantum computer is the math on the whiteboard. It is doing that algorithmic kind of innovation. And once you've got that, once you've got the hard math, actually programming it, actually writing the code out for it is relatively easy with all the tooling that's available.
Devin Liddell:
In artificial intelligence, I think there even to this day there still persists some debate as to, for example, will we ever even get to an artificial general intelligence, for example? There's some people that say, "Oh, absolutely, we're headed to an artificial general intelligence." And there's some people say, "Actually, I'm not even convinced that that's possible."
Using that as an example, are there any similar debates within quantum computing in terms of like, "Oh, this is where we're going to arrive in five years' time," or, "No, this is not where we're going to be, we may never get there"?
Michael Brett:
Yeah, I think things have changed extremely quickly just over the last couple of years. The primary debate two, three years ago was can this scale? Can we build quantum computers that are big enough to run problems that are of value? And can we operate these computers long enough so that we can get to the result out of it?
And the key technology that needed to come to fruition there was in quantum error correction, so identifying and correcting for the errors that build up on these computers as they perform their computational work.
And that was a really tricky problem. This is something that's been studied in-depth over the years and is a real conundrum. And we weren't sure if we could build systems that could do that work.
Over the last two years probably, there's just been a series after series of announcements around experiments showing the error correction pathway is working. We're able to implement this error correction regime, it's scaling better than we thought, it's more efficient than we thought it was going to be, it is unlocking that large-scale computer that we need to be able to run these commercially-valuable workloads.
And we're so confident in fact that AWS just announced expanded commercial partnership with one of our hardware companies to deliver a fault-tolerant quantum computer onto Braket in 2028.
So this is now a, it's a commercial contract, it's between us and one of our hardware companies, one of our suppliers into the hardware space, QuEra, who are now going to deliver to us one of these computers in a not-too-distant timeframe, just about 18 months or so.
And this is really exciting. This is a threshold moment, building up these large-scale, computationally-useful quantum computers that can run very large scientific kind of applications, things that, like what I was describing before, that we can't even contemplate running on a classical system today. And so that is now sort of on an almost tomorrow kind of timeframe in terms of the event development that we need to do.
Devin Liddell:
We had a conversation on the show with Rob Meyerson at Interlune, and they're working on moon harvesting. And one of the needs that they're looking to satisfy, of course, is helium-3 is important to quantum computers in terms of achieving the super cold temperatures that they require.
Just to use that as a backdrop, when QuEra delivers this computer, what does that actually look like? This is not a FedEx showing up on your porch, right? This is a different proposition?
Michael Brett:
So at this point, all of the computers we provide access to are at the manufacturer's facility because they are best place to operate it. They built it, they know how it runs, there's calibrations required, there's constant sort of tinkering and improvements required to operate these machines.
Eventually we will of course be integrating them into AWS data centers. That will be part of the future roadmap of how we operate them, but at least for now, the machine's at their facility.
For a QuEra device specifically, they're based on a modality of quantum computing. There are different types of quantum computers, we call them different modalities. So their modality is something called neutral atoms. And neutral atoms use lots of lasers to interact with atoms that are captured in a vacuum chamber together.
So what this looks like is kind of a laser bench, which you'll see at almost any physics lab at universities around the world. So it's a stabilized bench with lots of little holes in it and then lasers and lots of mirrors and lenses to direct beams around. And so it's a pretty crazy looking device, but within that, the core, the beating heart of it is this vacuum chamber with a sort of floating cloud of atoms that get stacked and locked inside these laser beams that interact with it.
But to the user, all of that is totally abstracted away from them. They don't need to worry about the physics of it and lasers and light beams and all that kind of thing. What they will interact with is a software library and an API to write algorithms and to send that workload off to the computer and get a result back from that.
And so that's the kind of stack, the design thinking that needs to be built out so that all of that magic complexity of quantum physics, they don't need to worry about that, they can focus on building the best possible algorithm for their workload.
Devin Liddell:
So quantum computing is here. The infrastructure is being built, the early adopters are in position, and a fault-tolerant quantum computer is now less than two years away. The next question is the big one. When access grows, what does quantum computing actually unlock for all of us?
Michael Brett:
Yeah, so I think as we think through the different phases of what's available and what the capabilities of these machines are, likely in the first phase of value, it's going to be quite esoteric, scientific kind of workloads that are really only relevant in the deep material science or deep nuclear physics or chemistry analysis kind of domains.
So these will be problems where using a quantum computer helps you create a model of a quantum system in the real world. So the way that molecules come together in the real world, the way that atoms interact, that is quantum mechanics in the real world. And so can we use a quantum computer to create a better simulator of that?
So likely the first applications will be in this physics modeling, chemistry modeling kind of space, which has got a lot of application in pharmaceuticals, in industrial chemistry, in material science to build sealants and adhesives and just different products that go into industrial manufacturing.
After that, it'll be sort of the larger scale kind of optimization problems. So very large scale like routing and logistics optimization problems, finance optimization problems like portfolio optimization, risk allocation, that kind of work, scheduling, these kind of big multi-variable, tough nuts to crack sort of problems that require a lot of computational horsepower today. That will be later with these systems, but there's huge opportunity there too.
Devin Liddell:
Is there a category that you think would surprise the average listener in terms of like, "Oh, I didn't think that that category would have an interest in quantum computing"? They actually seem quite pedestrian in their approach to problem-solving, but they're already using it. For example, will dating websites use quantum computing?
Michael Brett:
There is a startup out there that promotes themselves as a quantum computing-powered dating website. I-
Devin Liddell:
I knew it.
Michael Brett:
I don't know how many successful relationships have been formed here as a result of their technology, but we wish them well.
And I think what's really interesting is just how pervasive this sort of chemistry simulation problem is. And this is all across industry. The world is made of stuff, it's made of atoms, and the way that those atoms come together is incredibly important to everything that we build.
So the companies that are looking at this include pharmaceuticals, as you'd expect, but then agricultural chemicals, the future of fertilizers, the future of pesticides and herbicides and that kind of thing, and then the future of food chemicals.
So when you look on the box of cereal and the ingredients that are on there and there's red dye number 13 or whatever, that's a food chemical. It was produced by a company that invests deeply in understanding all of the properties of that particular chemical compound to make sure it's safe and it's manufacturable and it's transportable and all that kind of thing.
And so there's investigations into understanding better how to design safer alternatives to that or more easily transportable or longer storage life. The little improvements that can be made to chemistry and the building blocks of the stuff around us, those little changes that could improve something marginally can have huge economic impact for those companies.
So just getting longer storage life out of some input chemical to some product by years or months can change their business. And so that kind of chemistry aspect applies across the economy. It's going to be really, really interesting to see how that has a follow-on effect.
Devin Liddell:
Well, and you touched on this example earlier, I mean, even just small seemingly minute changes in, for example, how aircraft are routed around the world ends up actually saving gajillions of gallons of Jet A fuel, which also has a corresponding effect in terms of emissions. But that's a small change but that requires a big problem-solving engine to get to though.
Michael Brett:
Yeah. And I'm reminded of a problem. This isn't anything to do with quantum, quantum is not involved with this, but I visited a milk technology lab in Wisconsin a couple years ago. This is University of Wisconsin's dairy industry think tank that does innovations in dairy technology.
And one of the innovations that they came up with was a way to take or to kill the bacteria and make safe mozzarella cheese and to extend the shelf life of mozzarella. And what that did was it meant that the shelf life was now long enough that they could ship mozzarella cheese to Asia.
Devin Liddell:
Oh.
Michael Brett:
So it opened up another 4 billion people market available to them that otherwise just wasn't available because of the transportation time and distance around that.
And so profound effect for the dairy farmers of Wisconsin and the cheese production industry there to have an entirely new market available to them because of these tiny, tiny little changes in chemistry. And so opportunities like that abound in the quantum side as well.
Devin Liddell:
That's fascinating. Yeah, and that's where it gets into pretty transformational stuff. I mean, it's almost so big that it's actually hard to grasp how all of those types of changes could actually change the world pretty fundamentally.
The industrial use cases for quantum computing are compelling, but Michael explained that some of the most interesting potential users are hiding in plain sight, running cities, managing airports, and maximizing the efficiency of infrastructure, organizations that deal in massive optimization problems every single day.
Michael Brett:
You'd be surprised that big cities around the world do a lot of computational work. They are heavy users of high-performance compute today for solving some of these problems in routing optimization, traffic optimization-type problems, airports and ports and that kind of stuff.
And so already that is going on in the classical compute world, and we do see some natural kind of curiosity and adoption and investigation for the role that quantum computing would play as part of that portfolio of compute that they look at.
So I think the key takeaway from that is having that agility to adapt that computational strategy, that is a huge factor in running any large-scale operation today.
Devin Liddell:
Are there any other design challenges in particular that keep you awake at night in terms of, "Look, we need to solve this problem if we're going to get to a point where quantum computing can live its best life"?
Michael Brett:
Yeah. So I think right now the things that keep me up at night are often relatively short-term kind of problems, but big problems.
So in quantum computing, on Braket today I have literally seven computers, seven devices. My colleagues in AWS HPC, they will sell 10 million cores of compute before breakfast when they wake up, but I've got seven to work with.
So I've got a scarcity problem. I've got a very small number of computers to work with and a lot of demand. And then as we bring on these fault-tolerant quantum computers, like this system from QuEra I was talking about in 2028, that is one system. I'm going to have one computer in the world capable of running these workloads and the workloads will take some days to run.
And so it's a scarcity problem to be solved through design thinking and how we structure the work that we're going to do, what runs first, what runs second, et cetera. And then how do we integrate that with the rest of the compute that we need at AWS? We need to be able to provision CPUs and GPUs to run alongside the quantum computers, to be able to have elegant handoff between those at low latency.
The one computer is sitting in Boston at that QuEra's facility that's connected to a data center in AWS in Virginia. We need to be able to think our way through how do we create a overall holistic experience of compute that includes both the quantum and classical for something that's incredibly scarce so that customers feel like they're getting a really, really good experience out of this thing that they're excited to get some outcomes from, but it's very high consequence whether it works or not.
Devin Liddell:
I'm showing my age by saying this, but when I was in college, personal computer ownership was actually quite rare, and you would go to the lab, you would go to the computer lab. And to your point about scarcity, you would sometimes just wait there until someone finished doing whatever they're doing and then got on with the rest of their day and you could go sit down.
How do you actually, and I'm especially curious to dive a little deeper into this notion of the elegant handoff, which I love that phrase, how do you actually manage scarcity at the moment? How does it actually work for a user to go about using one of these seven computers?
Michael Brett:
Yeah, so it's relatively simple today. The way that we handle it, we have a queue. So you submit your task in a queue. Some days the queue is busy, some days the queue is not so busy and you get-
Devin Liddell:
So it's like the lab. [inaudible 00:25:45].
Michael Brett:
Yeah. So you get processed in the order that it submits and you'll get a little notification when it's completed. And for the most part, that works reasonably well. And it's a matter of communication and expectation-setting.
One of the things that we need to sort of educate our customers on and kind of nudge them along with is that this isn't the elastic cloud of infinite compute. This is seven computers and there is a queue and here's your expectation around that.
We also have an option where if you've got a particular workload you want to run and it's super valuable to you and you want exclusive access to that device, we can set up a reservation for you for Tuesday next week at 3:00 PM. You can have the device configured to your liking and kind of set up to do that and get that in advance.
But yeah, that's something that we're constantly thinking about is how do we set the right expectation with customers? They're working on a frontier technology that isn't quite as scalable yet as they would like it to be. They're still getting a quality experience out of that.
Devin Liddell:
This actually comes up quite a bit just with any kind of emerging technology. When a technology emerges, it poses questions around equitability like does this technology exacerbate the have, have-not problem, or does it actually liberate people from the have, have-not problem in some regards? Meaning that the technology makes existing patterns worse in terms of who has access to it or it subverts those old status quos and changes who has access to those things and makes the unpowerful powerful, for example.
Is there anything along those lines when it comes to quantum computing that you're especially concerned about in terms of what happens if too few people have access to scarce resources?
Michael Brett:
With quantum computing, it's never been more accessible. This is an incredible moment to get engaged in this technology. If you're a young person listening to the podcast for the first time and you want to get involved in quantum computing, you can do it, it's available.
When I started in quantum computing 13 years ago, so not even that long, the only quantum computers were in physics labs at fancy universities and you had to be a postdoc with a professor and have access to that particular system, and it wasn't easy to access any of these computers at the time, and now it is. You can use a AI coding assistant and get some time on the credit card and run a quantum compute workload. And it's incredibly accessible. And in some cases, some of these devices are made available at no cost as well.
So anyone with an internet connection and a little bit of time on their hands can contribute and be active in these open source communities around quantum, which is great.
We've got customers all over the world in all sorts of places that are able to access these remarkable machines and do something that even 10 years ago they would not have even had the opportunity to program any of these devices.
So I think that's on one side of the coin. On the other side of the coin, we are going to have the scarcity problem that the most advanced computers, the frontier computers are going to be in really high demand and it will be difficult to provide everyone access to that. We're not going to be able to fulfill everybody's desires for the access level out of that.
We'll do the best that we can, but the reality is that it will be a scarce resource. And at least in the early stages, only a few are going to, relative few, are going to have access to these remarkable machines and potentially change their businesses with, so develop your strategy accordingly.
Devin Liddell:
Absolutely. I feel like there's a trope within origin stories, sometimes within startups, but even sometimes even with filmmakers you'll hear these stories of like, "Well, we maxed out our credit card to pay for this production or we maxed out our credit card to buy computers for our startup." So this is just the latest entry into that trope. People can just max out their credit card and get access to quantum computing.
Michael Brett:
Don't max it out. Let's not get into too much trouble. But drop me a note first at AWS and we'll see what we can do to help you out.
Devin Liddell:
Be responsible about it. All right, I appreciate that. Yeah.
And now as always, it's time for the lightning round. We're doing the lightning round a little bit differently this season only because Teague is celebrating its 100th anniversary.
So I'm curious, what do you think, and we've talked a little bit about hype, both over-hype and under-hype, what about under-hyped? Are there any technologies that you think like, "Wow," you look back at the last century and we completely take this one thing totally for granted?
Michael Brett:
One of the examples I always give of the profound impact on society is refrigeration. It completely changed our world. And we just take it for granted as a natural part of our living environment that we have refrigeration and air conditioning and transport, supply chains, food security, the ability to support a population in the way that we do and just enjoy a cold beer on a hot day. Imagine a world without a refrigerator and it is a grim, grim world.
Devin Liddell:
Absolutely.
Michael Brett:
Yeah, so that's one of them.
Devin Liddell:
Almost cave-like, really. Actually caves probably actually served as-
Michael Brett:
That was the coolest place to keep...
Devin Liddell:
... early refrigeration systems. Yeah.
I'm also curious, are there any books or shows or other podcasts that you're listening to that you would love to recommend or that you have sort of regarded as especially formative in terms of how you approach your work?
Michael Brett:
One of the books that I really just love and read kind of once every couple of years is the Skunk Works book by Ben Rich. That was a book that as a young aerospace engineer, these were hero kind of stories and incredible stories about innovation, but I always get something different out of it each time I read it.
As my career's progressed and I've kind of learned new lessons along the way and then can kind of recognize those lessons being reflected back in those books around organizational structure and thinking and innovation and how you create culture and impact out of that, as a concise, quick read about an exciting technology, I think there's a lot of richness in the leadership lessons out of that Skunk Works book by Ben Rich and some other books alongside about that time at Lockheed Martin.
Devin Liddell:
Interesting. Love that.
Michael, thank you so much for being here. This has been super, super fun.
Michael Brett:
Oh, pleasure. Thank you, Teague. Thank you to you, Devin, for a great discussion.
Devin Liddell:
That's it for today. Thank you for listening to Design This Day, a podcast by Teague.
Subscribe on your favorite podcast app so you don't miss the next episode. And if you have a complex problem that needs solving, visit us at teague.com or send us an email at hello@teague.com.