The Exchanges

Q So we'll come back to you alluded to, I think, or you, um, prefaced some of the applications that might be applicable in a climate context. We'll come back to those in a minute, but first let's just quickly talk about where we're at in the trajectory of quantum computing. What is the, what is the current state of the technology today and sort of who are the big players?

A So we are in an era called the NISC era for quantum computing, the noisy intermediate scale quantum computers. Um, that is, we have quantum computers that work, but they're very, very small in the number of bits that they have, and so we are no longer in a realm of, of theory, whether we can make a quantum computer, or whether or not a quantum computer can do something that a classical computer Cannot do, or, or they could do it faster than what a classical computer can do. So we have crossed that threshold. Um, actually that, that's only been crossed in the last couple of, of years. So we are in a really, really interesting time where we have, uh, working quantum computers, but they're, they're, they're, they're early. They're, they're small and they, they, they lack a lot of power and usability and functionality. And so people are just starting to be able to play with these and test these. These types of, of computers, and the way we measure the power of a quantum computer today, there, there's several different types of, of methods, but we really talk about the number of quantum bits or qubits in, in a computer, and the most recent announcement actually came from IBM, uh, announcing the first ever three-digit or north of three-digit, uh, qubits, um, in their quantum computer, so they are Um, uh, have the, the largest, uh, functioning, uh, quantum computer that's been announc…

AI assessment note: “we are in an era called the NISC era for quantum computing”

Q good at everything, and so it's not like we end up some number of years from now with exclusively quantum computers. There are things that they will be particularly good at, potentially, and things that they may not be as good at. Can you kind of outline at the high level, what are the things, what defines something that we think a quantum computer might be particularly well-suited to solving?

A There's a couple of different classifications of things that quantum computer is really good at solving, but, um, most problems that it's going to solve are, uh, problems you can boil down to an, an algorithm or an equation that has, uh, massive numbers of permutations, um, that if you try to go through that brute force method, you won't ever get there within any meaningful, you know, human time scales. And so these problems are typically problems of, The physical world of chemistry, of biology, of pharmacy. They are problems of optimization. Um, they are problems of, of, of solving, um, really complicated mathematics around factorization, other things that we only have algorithms to solve that through a brute force method of trial and error. And so if you can take something that historically has been a trial and error type brute force method and do things in parallel, that's a really great problem for a quantum computer. Do I need it to, um, uh, use on my cell phone to play whatever game I'm going to do or, or, or in a, um, an Xbox or a PlayStation? No, I don't need it for that type of thing, and, and you, you may not use it for, for that type of application. It, uh, and if you look at the structures of quantum computers, they may not fit into a personal device or into a home, at least in the, in the near term, but we said that about Computers back in the day, famously, about …

AI assessment note: “problems you can boil down to an, an algorithm or an equation that has, uh, massive numbers of permutations”

Q a commercial fusion reactor. In the case of nuclear fusion, it is Q equals one, which is energy breakeven. There's like this seminal moment that we're gearing toward that like may or may not get hit by One or a number of the players in the next few years, and that's, that's one milestone on the path. Is there anything like that with quantum computing? Is there a clear milestone?

A That's a great question. Um, we, we theoretically have hit that Q equals one with quantum computing already, uh, with a Google experiment that was done, um, uh, a couple years back, which proved that they did a calculation, um, On a quantum computer faster than what you could do classically. That's sort of that Q equals one moment. Now, um, the detractors of that experiment said, well, this is a made-up mathematical problem that has no use whatsoever, that was designed to be successful within this space. Uh, in my opinion, I'm very excited about something like that, because who cares? It did something that no one, nothing else could do on Earth at that time, and, and they proved That was, it was a really interesting, um, sort of moment for the quantum computing space. And to take your fusion analogy one step further, Q equals one is incredibly important for a, for a fusion reactor, but if that Q equals one comes at a cost of 20 dollars per kilowatt hour, it is still useless for us, uh, as a society, because it's too expensive. And so that's the analogy of where we are right now with quantum computing. We have that moment, but that thing that it's done, Not all that helpful, right? Um, so, uh, but I, I will say there are, um, use cases today, and, and this is another, a way of how this, uh, this industry is going to evolve. It is not going to be, um, on or off. It's, it's not bl…

AI assessment note: “we theoretically have hit that Q equals one with quantum computing already”

Q the sort of known cathode material, uh, characteristics and known properties of other potential materials, and I can use that to predict, and that'll make me more precise in my testing. Or at some theoretical point in the future, I can plug all of the necessary parameters into a quantum computer, and it can tell me what cathode materials I should use. Is that... About right for one potential application?

A A hundred percent. If, if you know what, um, what properties that you want out of that cathode, you, you've got, it's not going to be, um, a, a magical, um, you know, genie in a bottle to, you plug in a thing and it, it pops out an answer, uh, but if you're able to define, um, the, the, the qualities and, and, and boil down that problem into what is effectively still mathematics. Mathematics is the core of all the different physics that we do. You should be able to do that with a quantum computer to varying degrees of accuracy as the power of that quantum computer increases, and the quality of the question that you're asking it, which is very important. If you don't know all of the knowns going in, you're not going to get the quality answer out. So it's still limited to the capabilities of a human to define what that problem is, but a quantum computer will be able to give you that answer, which is really, really exciting, and that That works for things that are incremental to say a better and better cathode. How do I remove Cobalt, or are we looking at wrong combinations of elements for what that cathode would be to improve something by five, 10%, which can be important for lowering the cost or increasing energy density or charge times of, of batteries? Or it could be a fundamental new alloy that is stronger and lighter that allows us to have better energy efficiency for cars. …

AI assessment note: “A hundred percent. If, if you know what, um, what properties that you want”

Q All right, so material science and biology, we talked about the third category that, that you mentioned to me is optimization, which, um, as I've been reading about and talking to folks about quantum computing, in general, people talk about the traveling salesman problem. A lot. Can you describe the traveling salesman problem and how it applies?

A So, uh, the traveling salesman problem is a problem that is currently unsolvable by classical compute in a time scale that is, um, uh, on the order of the universe. Uh, so it's, it's something very challenging to, uh, to solve, but effectively, uh, imagine I have a salesperson who, uh, needs to visit a certain number of clients and customers that are, um, On a map. And, uh, as I want to optimize the amount of steps that my salesperson takes, I want them to take the most optimal route through, um, through an ecosystem or through a landscape. Or if I have multiple salespeople, you know, how do I, how do I organize that? As that problem scales, it is something that gets so incredibly challenging that, again, because of exponentiation, we are unable to solve those problems. We can get close through Um, uh, through our classical compute, but we can't prove whether we are at a, a local minimum or a global minimum for, for something, and that means we, we get an answer that we think is right, but is there a better solution for that? There might be, and mathematically we can't actually prove that. Um, but because a classical computer, uh, is, is doing things through brute force, and a quantum computer does things in parallel, it is a perfect problem for a quantum computer to potentially solve, um, uh, through being able to Use less resources to do the same things that we do today, whic…

AI assessment note: “imagine I have a salesperson who, uh, needs to visit a certain number of clients”

Q lower zero carbon, but then you also need to meet these sort of output characteristics of, like, this is the nature of the carbon fiber I'm trying to build, or whatever it's going to be. So, theoretically, with a quantum computer, you can state both of those things. These are my input constraints, these are my output constraints, and then it will tell you your ideal formulation. Is that right?

A That is the future vision for what you could do with a You know, very powerful quantum computer, um, and what's exciting about that is if I think back to all the major, like, true decarbonization efforts that we've had and the true successes we've had, which are solar, batteries, uh, you can put wind and composites and magnetics in that. They're all based off of fundamental material science innovations, and unfortunately, you know, it typically takes Decades, 20 to 30 years to go from a university idea of something into a full-scale commercialization because it's very, very challenging to be able to do material science research with, with PhD horsepower. It takes a very long time of, of doing that brute force parallel track. You've got some theories, but it is, it is going through a, a scale up that takes a long time. And what if you could short circuit decades of that work By, by doing things and simulating what I call in silico, which is how do I figure out and, and do those experiments in a computer, uh, to be able to know what they might be, to be able to shrink down the millions and billions and trillions of possibilities of combinations of things down to something that's tractable and, um, and really supercharge material science to be a tool for, for climate change.

AI assessment note: “That is the future vision for what you could do with a”

Q Alright, well that leads me to my final question. Gun to your head, what is the first significant way in which a quantum computer is used to solve a problem that impacts climate change?

A I feel like I, I risk repeating myself, I, I do think it is going to be those optimization problems, honestly. I, I really think it's going to be in transportation, I think it's going to be in optimization, um, because that problem is far more definable, it is easier, it takes a lower power quantum computer, To, to solve some of those things, and you can do it along a continuum, which I think is really, really important, where if I'm doing material science, I either have that material or I don't, right? Either works or doesn't. Where optimization, there are still levels of, of, of performance you can do, and so I'm really excited about, uh, the ability to, um, to, to continue to, to show tangible emissions reductions from, uh, from transportation within, within a, a quantum computer. Um, but I will hold out hope that somebody discovers a, a, a, a catalyst or something that solves a really, really interesting challenge in, in, in, in fertilizers, in, um, in making other different chemistries that, that lowers significantly the energy footprint or the output of that, but I, I come to my head, I'm, I'm gonna stick with the, uh, uh, the transportation optimization examples that, that we see, because there's already red crumbs for that today.

AI assessment note: “I really think it's going to be in transportation, I think it's going to be in optimization”