Quantum Everything Explained - EP 81 Prineha Narang

Okay, we are heading off into the quantum realm. We're going to talk quantum computing. We're going to talk quantum encryption, sensing, quantumverse AI, all the quantum stuff. It's fascinating area, science, technology to explore. It's always a bit daunting to understand, but we have a great guide with us today. Pranay Narang is a professor at UCLA. She is an incredibly accomplished scientist and academic. She has done, um, her studies at Caltech. She's worked at MIT and Harvard, and now obviously UCLA. She advises the government on quantum policy. She's also a venture capitalist at DCVC, where again, she's helping out on their strategy around quantum technology in this episode. I'm always curious what's real about quantum, even though it's come so far. It still feels like as time goes on, it's unclear if there are tons of things where it's useful or if we're getting, we're getting quite far on our own. And so we get into all that. We obviously get into where quantum and AI competes or complements with each other. And on this quantum sensing stuff, which is an area of extreme expertise for Pranayaa. It's just fascinating, the idea of being able to use these quantum properties to detect things out in the world. Beyond being an incredibly impressive overachiever in her academic and professional career, Pranayaa is also a— she's a competitive marathoner, ultra runner, mountaineer. Um, everything she does, she does with maximum effort, and she's a wonderful guest. I am Ashlee Vance. This is Core Memory. Off into the quantum realm we go. Pri, thank you for coming.

Thank you for having me here.

How are you doing? You're, you're always traveling all around.

Well, it's just been a very, very exciting few weeks for Quantum.

It has been. Yeah. So we got, we've got some news items that I was going to, I was going to jump into and then obviously we're going to do all sorts of Quantum things. I was going to start us off a little more on you because just some of your background. I tend to cover fairly exceptional people with rather extraordinary backgrounds. What I noticed with yours is a— I mean, you're fast-forwarding life. You've accomplished rather large, remarkable things in a short period of time, both professionally. And then I noticed last time you were here, we were just hanging out, having lunch. You had this bag that I think you take on these ultra hikes or camping.

Jaap lines.

And then, you know, if anyone digs in to your background, there's so much about you were going to be perhaps a professional athlete. You're deeply into running. And I noticed you strike me as somebody who does just about everything you're into quite intensely. And this made me curious about It's always, when I see this, it always makes me curious about somebody's background. It just, you know, sometimes these things just happen and then sometimes there's forces at play.

Well, not a professional athlete. I wanted to be one, realized you have to be really, really fast as a runner to even sort of make it.

Yeah. And you wanted to be a marathon runner?

You know, I ran track and I didn't know that marathons are where most track athletes end up., but I thought I'd be a, you know, fast track athlete and it's just all going to work out. And then I realized, wait, could I had this physics thing too, right?

Yeah. Well, and I saw there's an anecdote about all this, but before we get to that, so where did you, where did you grow up?

I grew up on the East Coast. Okay. I don't normally talk about my early life too, too much outside of athletics and the fact that, I think I've always been obsessed with physics, always been obsessed with science. And for a very short amount of time, I thought I'd be a mathematician. Then I realized proof-based math is its own thing.

Yeah.

So anyways, that was my, uh, you know, how did I get to this part of science?

Yeah.

Story.

I'll ask this delicately then. Is there, is there a reason you don't talk too much about your, your background?

Uh, it's just a little complicated and You know, just, I don't have family anymore, so I try not to talk about that publicly.

Okay, okay.

It's nothing, nothing nefarious. Yeah. Nothing too exciting. It's just, yeah. Yeah.

And so you were, you, okay, so clearly you were into athletics as a kid and then obviously math and science. I mean, were you a, like this picture I was portraying, were you quite a driven, teenager and you were feeling compelled that you were going to go into math or science or—

I think competitive.

Yeah.

You know, the typical athletic competitive kid. I wanted to be first and I cared about that for some reason. And, you know, with science, it was always, okay, there's this thing called the Physics Olympiad, which is basically track and field for physicists. And so I was like, well, I wanna be good at that. I wanna win that. What does that look like? And I think in all of these things, I started to develop what we now call problem-solving skills. You know, how do you look at something that's framed as this like awkward physics problem? And I'm gonna use my hands to describe it because at least all the Physics Olympiad problems from my era were classical physics. And that meant these weird pulleys and weird masses in different spots. I used to have to go from this picture to equations to then solving something. And I didn't realize, but my desire to be winning this, whatever that meant, also meant that I was learning how to actually problem solve.

Okay. And I mean, I interviewed lots of people who've done the Math Olympiad. I don't run into too many who've done the physics. So that sounds daunting. Horrible to me.

It's same idea. You know, people come up with more and more complex problems to solve every year. And the Physics Olympiad is actually interesting because you can't do quantum mechanics. It's not fair game. Okay, so, so I do all of this and I now do zero classical mechanics. But at the time I thought all of physics is classical mechanics and you have these textbooks, they have complicated problems, you go to the next one and it's really the analogy with athletic endeavors is very, very clear. If you look at these and you're marked on these problems and then you can, I'd say graduate from one to the other if you actually can complete all the problems in some amount of time. That's at least when I realized I enjoyed physics a lot.

And this is in your high school years. So you're super competitive, you're running track, doing physics. And very unpopular. Were you a nerd?

Yes. That's putting it mildly.

Yeah. Yeah. I mean, you know, meaning you're just off studying all the time and not socializing a lot or.

Yeah, I think the social skills Maybe, you know, of all the things that could make the 24 hours, that was not making it.

Okay.

And so that required some development later in life.

And then, I mean, there's a story you tell of, and I couldn't tell, I think this was probably in high school where one of your track coaches—

In early high school.

Okay. So you are thinking, I want to possibly be an athlete. And this track coach, coach, I don't know if it's a man or a woman, is like—

He's a guy, yeah.

He's a guy and he's like, maybe you should focus on the academics or—

You know, I think somebody put him up to it.

Yeah.

I think somebody probably said, you know, you gotta redirect this person and they're not gonna take advice from anyone that isn't a coach. I actually have, you know, this immense respect for coaches. I still have a run coach and I can't, even if I try, call him by his first name without saying coach. His name is Blue, so I call him Coach Blue. And folks are like, well, you know, you can, like, once you're not in high school, call coaches by their first name. Nope.

This is your coach. Like, you still run all the time.

I still run. I still think I can be, you know, reasonably fast.

What do you do? Do you do the long distance stuff?

I am training for a few different marathons and trying to hit some, ambitious time goals. And Coach Blue's whole philosophy is, you know, leave everything out there, but for adults.

Okay.

So that's, yeah, trying to see what I can make happen.

Is Coach Blue at UCLA or?

No, LA-based and used to coach, I think the Nike team and now is with a different shoe brand and doesn't really coach anymore, but, took me on as this hopeless but will not give up athlete. So that's at least—

What's a good marathon time for you?

What's a good marathon time for me or what is a time that I really want?

What are you aspiring to?

Well, the OTQ standard for women is 2:37. So that is my goal. I'm a little ways from that. I have a sub-3, but not, you know, So Delta at that level is pretty significant.

OTQ is Olympic, is that?

For the trials. Okay. Lots of people run the trials. Okay. And it's like, I wanna say hundreds, maybe even a, you know, 1,000 women will run the trials. Okay. And the time standard is just a time. So it's not, you know, there'll be some folks who are elite, elite, and then a lot of folks who are like me who are just, kind of obsessed.

Okay. But yeah, I mean, it sounds like you're no, no joke.

Well, see, you know, and it was one of those things and actually saying it out now publicly, or at least on air. Yeah. Feels like it's real, but I had it in my mind. I wanted this goal and I didn't really know if I could even tell anyone this. I was like, they'll probably think I'm not gonna focus as hard and work if I'm doing this other thing. So anyways, that's when I brought on Coach Blue, who specializes in people who also have work.

Okay. Okay. Well, and I'll stop badgering you about all this in a second, but I was just struck the last time. You seem very passionate about all this. And so what's the, as far as the outdoor stuff goes, what's kind of like your best adventure you've been on?

Okay. My most recent, which tends to be how I remember things. Rank these because it's hard to compare which mountain is more, you know, beautiful. Or I was just in Ecuador this past December and I got a chance to have this gorgeous sunrise summit of Cotopaxi. And that was another one of these goals that was on my list. And that— it's not a very, very technical mountain, but it was erupting for a while. Turns out volcano active, erupting, can't climb. So it worked out for various reasons and I was really happy with like this sunrise summit. I'll send you a video of that.

Okay, okay, okay. So you do the, you do climbing and camping along the way?

Yeah, yeah. It's like, you know, I'm not so much into what is the most technical route I could find. It's more, you know, what's up there?

Yeah.

Can I get up there? Oh, could it be a little more challenging? Yeah, my, my thing.

Okay. Have you ever done like the ultra marathon things?

Everyone keeps asking me this. It's, it's on my list to do one of these like insane long runs. Actually, this past weekend was the Speed Project where folks ran from LA to Vegas. I was at the start line and, you know, folks are like, are you running? It's kind of, you know, there are many ways to get from LA to Vegas. Running seems like one of the harder ways.

Yes.

Could bike there maybe, but you know, it's 300+ miles.

Yeah.

So I'm contemplating what, how I feel about these long distances, but after I hit one of these faster time goals.

Yeah, okay. I remember I read that, what was that, Born to Run, that book about the barefoot runners?

Yeah, yeah.

Yeah, I mean, it sounded like such a, it's not something I would do, but it sounded like a very interesting world.

A very, very intense set of people out there. And I think as you, track folks and their times, you realize they're not really running slow. They're still running, you know, a 6:30, 7-minute pace. They can just do that.

Yeah.

For days.

Yeah.

I'm not quite there yet.

Okay. Okay. Okay. Well, I will drag us now towards, towards quantum land. You, from high school, you went to Drexel University. Yeah. Did you, I mean, obviously you're doing physics in high school, so Was that what was always calling to you? When do you start to, to make this push towards quantum? I mean, did you know that that's what you were going to do once you—

So I didn't. I, you know, got to university. I actually went to Drexel. They gave me a full ride and kind of actually paid for me to go to school, which is really important to me. And I just thought, I'm going to do something that is practical. So it's going to be more, you know, engineering physics, maybe something that's hands-on.. And I really wanted some experience in a lab. So I wrote completely out of the blue, an email to Frances Ross at IBM TJ Watson. And she's now on the faculty at MIT, but she'd run a group there for some time. And she's doing in situ electron microscopy. So effectively ways of looking at how things change under the microscope, electron beam. But doing that in a way that you could also do some liquid cell electron microscopy. And I knew very little about all of these things, but I said, I'd really like to have some experience in the lab. T.J. Watson, everyone's heard of it. And would you take a chance on me? And I have no idea why she responded. I've asked her that question. She's like, it just felt like something that you could work on and I had a project for you. So I did an internship with her and then that actually became something that I did my senior year as a part of my senior thesis as well. Ashlee Vance as well, and looking at how materials behave really, really at the atomic scale, trying to figure out the kinetics of these different processes. So you have electron beam and you're putting in a lot of energy, but it's in this closed liquid cell, watching how things change. And I realized, okay, to model this system, you actually need to do this quantum mechanically. A simple kinetics model or something that you can do that's semi-classical is not sufficient. I'd taken, like, quantum mechanics at that point, but I hadn't done enough to actually be able to address this problem.

Okay.

So I, again, at lunch would see all these really senior theorists and they would talk about, well, here's a textbook, here's a paper you should go read. And I started looking at these papers. It's like, man, there's a lot of jargon here, a lot of things that I've never seen in a classroom. And that's actually what got me excited about going to grad school.. And I didn't actually know how, you know, folks think about, you know, you have to do all these things in undergrad to get to grad school. I didn't have folks who walked me through that. So I just, you know, applied to a few places, all in California. I had at this point decided East Coast, not for me. Yeah. So 3 schools in California. I'll let you guess which ones admit me. And two were in the Bay Area and one is in SoCal.

Yeah.

And I visited Caltech, and Caltech at the time actually didn't have just a visit weekend. They would call it an interview weekend. It's since become the same as a regular visit. And I had never been to Pasadena. Didn't really know the fact that, you know, Pasadena is this like tiny town. It's not really LA. It's about an hour away from what people think at the city of LA. So I get there, absolutely fell in love with it and decided I'm gonna go to grad school and do something that involves something quantum, but still something practical. And it's gotta be a little bit mathy, so probably theory, but I didn't know yet.

All right, what do we do at Core Memory? We cover innovative, fast-moving, forward-thinking companies, which is why Core Memory is sponsored by Brex, because Brex is the intelligent finance platform for many of these companies. 30,000 companies from startups to the world's largest corporations rely on Brex's technology for their finances. They've got smart corporate cards, high-yield business banking, and expense automation tools that are fantastic. I hate doing my expenses, and Brex's AIs and software run right through those expenses, figure out where we're spending money, and take care of so much stuff for you so you don't have to waste your time on it yourself. Go to brex.com/corememory to learn more and just, you know, get with the program. Let's get going. Let's get out of this archaic finance software and move toward the future. CoreMemory and Brex. In what year did you graduate from Drexel?

2011.

2011. And so it's like a— it's a pretty interesting time for the quantum field. You already had, I guess, D-Wave had just gone, whatever, definitely de-stealthed and come out in the world and said, we think we've built a quantum computer. And you started having— I mean, for people who don't follow this stuff, I always hate doing this because I feel like You have to define some of these terms and then it still doesn't, doesn't get us very far. But, but in traditional computing, we have bits, um, ones and zeros, and obviously all modern semiconductors are based on, on this principle and have been since the 1960s. And then in the quantum world, we have qubits, which are again hard to define, but, but start to get you into these quantum properties that are different than the traditional bits. And I remember back in that period around 2010, 2011, I mean, you did have a company like D-Wave, which I remember as being the first quantum startup to get a bunch of attention. But it was still like the quest was really just, can you build qubits? Can you build, you know, it's like, can you build one? Can you build 4? Can you build 8? Can you build 16? And you have all these labs trying to, trying to do that. I mean, that, that's how I remember that time. But I'm—

so the experience on the grad school side was totally disconnected from what was happening in the startup ecosystem. And at Caltech, there was what became the core set of faculty that would be the Institute for Quantum Information. And then eventually an M was added. So it became from IQI, IQIM, Information and Matter.

Okay.

But there was a core set of faculty that were kind of starting to meet, stuff was happening. As a graduate student, you walk into a faculty member's office, you're kind of nervous. Like, um, I want to work on this or that, or really anything that you think is exciting. So I talked with a few different faculty members, and I ended up working with Harry Atwater, who is known for work in photonics, more recently in quantum photonics. And I just said, okay, this seems practical enough that we're gonna— like, somebody's gonna build this device, something's gonna happen, it's measurable. It seemed a little scary from some of the other very, very abstract IQI conversations. It wasn't IQI at the time, it was just the same set of faculty. Anyways, so I wanted to do something that I thought would be useful. What I didn't realize is that as soon as you start doing theory, it's kind of a little bit separated from something that will be immediately useful. And it turns out I'm better at theory than experiment.

Okay.

It's just how it happened. So I wanted to focus on things that were like, how do we describe dynamics of these collective excitations, plasmons effectively, where a photon couples to matter, creates this joint excitation and said, well, you could always talk about it in a classical limit, but what if we talked about it in a quantum limit? What would that look like? Started kind of building up what I eventually put in my thesis as like a framework for describing such collective excitations. And that then brought me closer and closer to condensed matter. And that, you know, eventually both of those things became part of the— we do quantum information to understand and simulate many-body phenomena. So that's, that's kind of the the journey.

Okay. I mean, it seemed like you, from everything I was reading the last couple days, it just seemed like you are fascinated by the edges of matter and, and what is happening when you do strange new things to, uh, to atoms. And, um, absolutely. Is that fair?

You know, so you start putting photons into some of these, uh, systems, you can create non-equilibrium states, which is a fancy way of saying the state that really isn't easy to describe by every equilibrium theory, which turns out most ways of describing stuff is in equilibrium. So I got really fascinated by, you know, can we develop a non-equilibrium framework for some of these types of correlated quantum matter? So correlated, again, you know, something that has long-range interaction. I was very, very surprised when, you know, I would ask, hey, has, Has someone developed some code or way of simulating it? Can I just use it? And realized, well, folks have done some things, there are demonstrations in some of their papers. And then there's another set of folks who are out there developing these general codes, right? And whether they come from condensed matter or quantum chemistry, they're kind of developing these for large supercomputers without thinking about a specific process. There was like two completely different communities.

Okay.

And the second, the community that was developing these methods and codes for supercomputers actually is now the same people who are developing, you know, the algorithms for quantum computers. And that is just, it's its own thing, right? How do you work on a part of the code? Somebody else does some other part of the problem. And it goes all the way from you formulate an equation that you think is a good description of the type of system you want to simulate. Then you go from that to functioning scalable code. And this sounds straightforward, but actually it's far from because you start looking at things and say, well, that's not really a problem that can be a finite element problem. Can't really use this kind of math solver for this equation. And so there's a lot of different fields that come together to go from that, like, I have formulated an equation to I have code that runs on, first it was Titan, then the Summit supercomputer, now Frontier. I'm actually really proud of the fact that I've accessed each of those generations of computers. And it just kind of became like, how do you scale this thing? What is efficient code? And not just is it efficient from a software standpoint, but actually there's a big difference between I can write something for a toy problem and it will still take the entire supercomputer's time for a month versus I can do it in a slightly different way, maybe make a different approximation, and then I can do a real system, something people care about in, a reasonable amount of time. So, and what's there?

And so, and as you're digging in all this stuff, I mean, in this, this— I mean, maybe this is a stupid question, or maybe it's not one or the other, but you, like, in this moment, you've done all this research, like, trying to push things to their limits and, and, um, exploring the edges of these fields. And then separately, there's just the promise of, of quantum computing and what it could unlock. For the world and different sets of problems. Were you being just pulled mostly by your interest in the nature of matter and what you could do with it, or you were— you were buying into quantum, you know, as a whole and where this field was going to go?

I actually was more in that first category, thinking, how can we leverage these devices It seems like I know a fair bit about how to use conventional supercomputers, and I'd say it was pretty early to the game on adopting GPUs in the context of computational physics. Sounds weird to say that now, given that no one would do it any other way. But at the time, if you came from a traditional physics program, it was not obvious that you wanted to use the largest GPU systems, which is what Oak Ridge hosted at the time. So I thought this is another tool, another approach. Maybe if the same problems can be just written slightly differently, you could have some cool algorithms. You know, didn't really think of it as quantum computers are here, let's, you know, not do classical compute. It was more how do we simulate some of these systems? How do we simulate these systems better than you could do on, you know, the biggest, baddest classical computer.

So I think I remember this moment. I mean, you're talking about like the early days of NVIDIA starting to sell these, these GPU processors for the supercomputers.

Exactly.

It is. It's funny how all this started.

All of those, you know, that, that quantum chemistry community I mentioned, I mean, they were writing and I think this still exists like this, you know, old Fortran and folks were like, oh, we got to migrate things over. And this is, eons before a Claude code type situation. So, you know, folks are trying to figure out, okay, CUDA, how do you take advantage of some of these new approaches and do that for simulating the same problems that the community has cared about for a while?

Now, I remember, I remember all this moment well. So I was at the New York Times and I was talking to Jensen and, you know, he had just He was really pushing this CUDA idea hard. It was far from a mainstream thing by any stretch of the imagination. I'd been covering the supercomputing race for ages. I used to always write every year the, the top 500, you know, how it was shaking out. And you started to see this rise in the top 10, top 20 systems of, of these massive supercomputers that then had like a bank of GPUs that you could tap into and I remember pitching this to the New York Times. You know, I was like, there's this kind of interesting phenomenon going on. All these scientists are starting to use these video game chips to speed up scientific code. And, you know, and Jensen, this guy from NVIDIA, nobody— I mean, they kind of knew what it was, but not really. Yeah, it's really fun. I have a story from 2008, I think. And the Times was like, I don't think anyone cares about this. We're gonna put this on like B8 of the business section. And then, you know, and then the, you know, what are we then almost less than 20 years later is the most valuable company in the world based on this bet. Yeah.

Well, you should dig up that article. I'm sure it exists somewhere in the archive.

I've got it.

I've got it. But, you know, also there was this sense that these very, very complex physics problems, physics, chemistry, I kind of use that interchangeably when it comes to many-body quantum simulation. Nobody really thought that GPUs would be the route initially. Yeah. I definitely felt like some of the more distinguished members of the community would look at folks like me and say like, oh yeah, cool. But it should be elegant. You should have an analytic solution for the whole thing. And I just, eventually thought, maybe the physics doesn't support that. You know, have we thought about that? Never said that publicly till some of these distinguished colleagues have now retired. So now I feel bold and I can go out there and say this. But, you know, they were reluctant to kind of think of this new compute architecture. And that story kind of repeated itself with the era of quantum algorithms as well, where folks said, well, look, this is not really going to do something practical.

Yeah.

And I think quantum is now at that point where, well, we are actually showing something that you can't trivially do using, you know, most of the classical resources that you can get access to via a DOE proposal. So, you know, we're at that point, I'd say very similar to that. You know, first DGX-1 kind of moment.

All right, pod friends, I am here to tell you about SendCutSend. They are a manufacturing phenomenon. Whether you're working on your own car or you're a giant automotive company, a rocket maker, if you need metal cut, machined, or bent, you're going to SendCutSend. They'll do it for you right, and they will do it for you fast. Go to SendCutSend.com/corememory for a 15% discount on whatever you're trying to make. We love SendCutSend. Go visit them today. Well, let's dig into that a little bit more, and I'm trying to keep this at a level where we're, we're nerding out. I want this to be as accessible to as many people as possible, which is always, it's just like inherently hard with quantum because it's just, I think. It's hard for me, even though I've covered it for years, to fully wrap my head around some of the concepts. I get the gist of things. And then, so I just always find— I've done a video about this. I just, I find it's quantum is like especially hard, I think, for people to, to kind of latch onto. But, but so let me, let's dig into this like a little bit more and let's go broad maybe at the beginning. But, you know, if I go back 10 years, um, the Even though there are these signs of where the GPUs are heading, classical computers are by far dominant. And quantum really is sitting there as like our next best option. And it hasn't been fully developed yet, but people can see where it's going and it's like, okay, if there is a massive future of computing, new things to be unlocked, it's almost definitely going to be through quantum computing. But we have a lot of work to do in those 10 years, clearly everything we've talked about, the, the GPUs have risen as this extra engine to do a certain class of new computing work that happens to be like kind of the dominant or close to the dominant, you know, computing work that's happening now, AI. And then, and then we've made huge strides with quantum computing where we have dozens of startups, we have what we were talking about before, where qubits used to be rare, the science experiment where a lab would be like, oh, we did 4 last year and now we have 8 qubits. People are making qubits all the time. Yeah. So, so we've kind of dialed that in and you've got startups claiming different things. But there is part of me that feels like over this 10-year period, because AI has risen so dramatically, because GPUs have risen so dramatically, that I don't know if it's like some of the luster around quantum computing has, has gone away. It just, it somehow seems like, oh, well, this other thing appeared that's really interesting and it's doing a lot of stuff. And maybe we don't really know where AI is going to go. Maybe it's just going to solve all these problems, you know, that were hard anyway. And so, yeah, I mean, like you sitting there in this chair, you know, in 2026, how should we be feeling about quantum in this moment where, where it's kind of realer than it's ever been in a lot of ways. But it's, it feels like the attention of, of maybe where the future of computing, everybody's kind of looking over in this other spot now. Is that, does that seem fair?

Yes. And maybe.

Okay.

So, you know, a few years ago, the most popular question on a panel used to be, is AI or quantum going to be first, right? And it was framed as a competition. What we've seen in the last maybe 3 years is that everyone thinks of quantum and AI as going together as part of this next wave of compute. In fact, you know, we don't really care if you're dealing with AI chips that are special purpose silicon, is it photonics, is it a different architecture? Most people out there not interested in how the chips are designed, what's the underlying hardware. And that has actually helped quantum a lot because you can think of quantum as just being part of this set of computational tools that we have, these computers. And some of them look like they have giant lasers in them. Some of them look like they need to be cooled to really low temperatures. But our understanding of what looks like a computer has actually changed because of AI. That's one thing that, you know, I'd say, you know, the attention has not shifted away as much as it's not that you are solving the problem using a quantum computer, you're solving it using some form of next-gen compute.

It's like we have these new tools at our disposal, which makes a lot of sense. I mean, the x86 architecture that most of computing was based on is quite old.

Right.

So this makes sense.

So there was that. And the other thing that I think has really, really helped quantum, and this is where my maybe answer came in because, you know, we care about the energy of compute so, so much, right? And the first set of results in quantum were all about time to solution, right? If you go back to the original Google Quantum Supremacy experiment, they had both the metrics for time to solution and the energy to solution. They published time to solution 'cause they were like, well, that's going to be the more striking result if you say it's gonna take the age of the universe on classical compute, and we can do it. Now everyone's asking, what is the energy to compute? And how are we, you know, in this era of bring your own power, if quantum can bring that workload down by say 10%, 20%, it just becomes part of that solution as part of that, you know, AI supercomputer that you deploy.

What do you like, what do you think that number has to be for somebody to go to all the effort to put their energy towards quantum? Like, is 10% enough or does it have to be 10x?

I think in this case, a few percent is sufficient. And, you know, some examples of that are where folks are thinking about neutral atom quantum computers that are very efficient from an energy standpoint being incorporated into a larger classical, you know, AI-focused supercomputer or data center. Okay. And they're actually now places building such systems out. Again, as recently as about 2 years ago, there was a question of if you co-locate these devices, is there a compute advantage, right? No one's talking about that anymore. It's more, is there an energy advantage at all of taking some of these devices and deploying them concurrently? And if there is, we should do that.

Okay. And then, I mean, Every time anyone talks about quantum computing, the problems that tend to get listed are things like it will help us deal with materials science. You can, you can look at molecules, the interaction of chemicals in ways that you just cannot do with a traditional computer. Encryption, solving these very difficult problems that would take decades, if not hundreds of years, with all the compute power in the world. You can do this on a quantum system. It's always— it's this I always feel like it gets lumped into this, this bucket of complexity, and there is just part of the natural world which just cannot be represented in a traditional computer. And quantum gives you access to play because you are dealing with, with entanglement and quantum properties. It lets us tap into a world which does have quantum properties as a huge part of it. And so it lets us model these complex interactions of the way the world operates. Is this, is that still like when you're talking about having these systems to solve working alongside traditional computers, is that still kind of like the bucket of problems we're after? Or have things, have things narrowed as GPUs have gotten better and AI has come along?

So I think it's always the problem. That we could address using quantum computers has actually broadened.

Okay.

And part of it is that, you know, there's hybrid quantum classical compute, which is very popular for a while because the quantum devices weren't, you know, at this fault-tolerant threshold. We'll come back to that threshold and some of the encryption results that have recently come up. But the idea was, okay, you do some core part of your material and molecule simulation that is the most quantum, on the quantum processor and the rest you can do classically and you have a workflow that essentially brings it all back together. Now, of course, we have quantum devices that, and these are across different modalities where folks have shown that you can actually, you know, simulate a very, very complex quantum system. You can do many-body dynamics on these systems and that's just been really helpful, but You can also think of that as generating kinds of data that you can then train models on, which is not originally something that folks had anticipated, right? So synthetic data you can generate all day, every day, but it's data that models have already seen. It's not actually solving the data bottleneck in AI. Whereas you do a simulation of a spin liquid on a quantum computer, and suddenly you have an exciting new dataset. And I think that is unlocking the, you know, some of the things in physical AI that folks have articulated on, on your show in the past.

Okay. Okay. This is a terrible question because I cannot intellectually back it up myself. But I think I was telling you last time, I remember I was sitting with George Hotz, who's, you know, he's quite eccentric in some of his ideas, but we were like having a drink at some, some event and I don't know, we were just shooting the shit and, and, you know, he's, he's just like, I still don't believe in quantum. I do not actually think these people can point to like any problem that we actually need solved that we, we can't do otherwise. I mean, just like a huge skeptic. And, you know, you do get that from—

Oh, I—

yes, yes, absolutely. But so to them, you just say they are incorrect. Like, there is definitely classes of problems that you believe.

Yeah. Yeah. And, and we have, you know, um, this is where the, the mathematicians will say we have strong conjectures, but we don't have rigorous proof. Okay. Uh, speaking of things that are not accessible to the general public, I think that statement is completely lost on a physicist too, where we say, yeah, you know, that's fine. In physics, we're okay with strong conjectures. In fact, we're okay with "We think and we intuit and it'll work out." And that's, you know, most of physics kind of goes that route. Of course, people come from a complexity theory or from a computer science and formal math background will tell you we don't have rigorous proof that this class of problems can only be solved this way. And there's a whole set of folks that are pushing that frontier, right? They're looking for which problems might be, and again, within the simulation of quantum mechanical systems, materials, molecules. They might be types that are easier, some that are harder. And, you know, folks on the complexity side need to chip away on that problem. Really excited for them to do that. And then there's a larger set of folks who are like, let's try and let's see if we can solve some problems that are, we think are well suited for quantum computers. And we'll, be able to make progress. We know we have new algorithms and we can try some of these problems today, even on nearly fault-tolerant but not quite over that fault-tolerant threshold systems. And this is, I think, something took me several interactions with folks in the pharma side to understand. You don't need to do drug discovery 100 times better or even 10 times better. 1 to 2% is still a meaningful advance in that field. And, you know, there are so many— and it's not just discovering the structure of a new drug, but also the kinds of interactions. Can we— and this is something we do in my research group where we think about open quantum systems. So how does a quantum system interact with its environment? How do you simulate that exactly? It turns out that, you know, how drugs behave in the human body, how you think about dosage for folks going through chemotherapy, those are real problems., and they're problems for however you wanna solve them, whether it's GPUs, whether it's quantum computers, whether it's a new computer architecture, they're important enough problems that, you know, we need to go after them and we don't need to wait for somebody to come up with a complexity theory argument for those problems.

Well, I mean, but that, okay, so that is like one part what that I was interested in. I mean, if you look at the results of the last couple years, whether it's like AlphaFold, things that DeepMind is doing, or I mean, I'm just seeing— I never know how much to believe some of this stuff, but it looks so promising in the biotech space where you, you see people doing these, you know, they're, they're changing the structure of a virus and they're coming up with millions of permutations and then they're finding ones that the body, um, that really work when you, when you make a drug based on this virus, like a gene therapy. And now used to take us decades to even find one or two, and they're just sending it off on this LLM and cranking these things out. And I mean, there's still a ton of work to do to see really how well all this stuff plays out. But the results, at least so far, are dramatic. I mean, that was always things in my head that I was told— I felt like I was told, oh, we're going to need a quantum computer to do stuff like that. That was kind of a little bit of the heart of my question, I think, is, is, as you're looking at some of what has poured out of the AI race, I mean, you answered it already, but it does— it just feels like the, the quantum pitch just keeps getting narrower to me. But like, because you were talking about pharma companies, and so, um, maybe I am misinterpreting, or maybe there's just— there's so much work to be done that there is still this enormous swath where they would still want a quantum computer to push that type of thing even farther.

Yeah. You know, you, you see quantum on their roadmap when they talk about next-gen compute. And you know, it's back to that point where it's, it's just not separating out that this is quantum, it's all in this. And I think the word AI has now, you know, become a catch-all for all next-gen compute.

Yeah.

Right. And physical AI, world models, LLMs, you know, you like know all of that from the, AI side. And then there's stuff from the quantum side. There's how do we actually accelerate some of these? It's all one big bucket now. So when somebody tells me now that I, you know, have AI supercomputers on our roadmap, I just assume that quantum is rolled into that unless told otherwise.

Okay. Okay. Okay. And the— oh, there's so many things I wanna ask you about 'cause I really wanna get into quantum sensing and some of these other areas. Let's do the news real quick. Okay.

Yes.

I don't know exactly when this will air, but roughly in the time that this airs, I mean, there was a massive announcement out of Google. Yeah. I mean, please, I'll take a crack and then you can, you can do it justice. But I mean, what I read rather briefly was that according to some of the research and simulations they were running, this, there's always been this, this fear or this idea that these quantum machines will come along and just obliterate our current scheme of encryption and data protection. Protection and that they'll be so good at the math that sort of this very fundamental part of our economy, of the way computing systems works, will get undermined basically overnight. And this would be, you know, quite devastating to money and information security and all types of things. And it seemed that Google was saying, well, we kind of thought this would, everybody thought this would take quite a while. And the roadmap that we're looking at, says this is actually gonna happen much sooner than we thought and we will crack this. Is that, was that the heart? And I remember we were texting.

I know, I know. Oh gosh.

You were getting bombarded.

So, okay, so there, there are two results that came out, but before I even talk about those results, I wanna say that, you know, everything we, we think about for encryption currently is based on either RSA or ECC and all of our—

Which is like elliptical curve—

Elliptic curve cryptography.

Cryptography. I screwed you up.

I don't know, it's fine. It's, I can always say it just once, but if, if, you know, we keep saying it, ECC, right? So all of everything we know from our friends in crypto actually comes from, all of that relies on ECC. And so, you know, when I, we were texting, what folks were sending me was like, is quantum coming for my crypto wallet? And I, you know, just thought to myself, no, let's, let's not, put that out there, let's think about, you know, the fact that there is this thing called resource estimation. Okay. And I'm going to explain this because this is at the heart of this whole issue where somebody, and lots of somebodies, people actually, this is a subfield, estimate based on quality of qubit, reconfigurability, all of the types of gates, types of circuits you can construct, With all of that, what is the total qubit count that you need? And, you know, then you assign thresholds for error rates, et cetera. Think of it as like making a very, very complicated, perfect cocktail, and somebody comes up with, this is their recipe, and then there's a number at the end of it. Now, what has happened is that there are advances on the algorithmic side that have brought this resource estimate down. And there are advances on the hardware side that have brought this resource estimate down. So there's a class of quantum error correcting codes called QLDC, Quantum Low Density Parity Check Codes. This actually, you know, so how we think about error correction was all based on these surface codes, right? And so there was a resource estimate that people could give. And I was like, you need this in order to do this, right? Well, somebody said, well, what if you don't do a surface code? Why don't we do one of these other codes? Or actually, and then this field kind of took off. That plus people being able to do more reconfigurable qubits in the neutral atom community. Now folks in the silicon spin qubit side being able to figure out shuttling. All of those things have brought this resource estimate way, way, way, way down. So you can see a plot in a paper that came out of the Caltech group. They are spinning off a company called Aura Atomic based on this.

Okay.

Where they show Now this resource estimate now is about 10,000 qubits. Okay. With a reasonable runtime for ECC and a not so reasonable runtime for RSA. Okay. And now you look at that trajectory and you're like, whoop, 10,000 qubits, neutral atom. Hmm. I can get one of those from Atom Computing. I can, you know, place an order. This is like, and it's for a system that's expensive as as a private citizen, maybe. But, you know, as an enterprise, you can buy one of these for less than $100 million, which sounds insane to say it's affordable.

It's doable.

But it is for a lot of the folks who think about, you know, buying large-scale compute. So suddenly folks say, okay, well, you can buy these computers. How do we think about that? And how do we think about the migration to post-quantum? Post-Quantum Cryptography, PQC, right? As a Google paper, two things. First, some of those initial data points that brought this resource estimate down came from Google. There's a, you know, team there that's been working on it.

On quantum for ages as well. Yeah.

There's a whole set of folks who are on the neutral atom side, and then there's the algo folks, including these QLDP codes like Iceberg that are just hammering away on this problem, right? So with all their powers combined, now we, we see a path that looks like, hmm, whether it's 10,000, whether it's 30,000, but not in an order of magnitude higher, it's in that range, you actually have a credible threat to ECC, you have a credible threat to RSA. And this is exactly what, you know, folks would call a technical surprise on the, intelligence, national security side of things, because every— all of these fields existed, but nobody quite saw how it was all going to come together so fast.

Okay. Okay.

Yeah.

Because it sounds like you're saying there were signs of all this was happening and then the Google team just, just sort of put everything.

Yeah. They, I'd say the Neutral Atom folks, two teams, or Atomic as well as Atom Computing, have both put out these numbers and Some of it also came up in the context of a DARPA-run program called QBI, Quantum Benchmarking Initiative, which started to look at these complex, complex scorecards of all of these things, right? I gave this analogy of a perfect cocktail. Well, that still might have 1,000 ingredients if you're really fancy. You know, these scorecards have tracked more vectors than that. And you can see that, okay, Advances in every one of these has actually moved things from the yellow or red to the green category. And that's exciting.

So, well, and so you were, you kind of buy into these results.

I think they're enough for multiple people. Yeah. If it was just the Google team, just the Wordtomic team, just the, pick your favorite, company out there, I'd worry. But the fact that the entire field has been showing this and these resource estimates track what is in the DARPA QBI program, that gives me pause. Right. And again, not panic, just pause.

Yeah. Let's talk about that for a second. What the number weren't they saying it was like 4 to 8 years or something like that?

Yeah. So that again, that's like at 10,000 qubits. Okay. Now you say, well, it's slightly better qubits and slightly more, what happens at 20,000 qubits? And, you know, if you look at the roadmap, say from folks on the trapped ion or neutral atom side, and you know, those numbers are in the next 2 to 3 years. Okay. And if you think about a system that's about 100,000 qubits, and this is, you know, again, comparing qubits across different modalities is hard, but roughly the same number of logical qubits across these. You find that we'll get to that threshold. And this will also roughly coincide with when we hit fault-tolerant quantum computing, which is kind of the threshold that everyone in the field has been waiting for.

Okay. I know you've been asked this question, I'm sure, 100 times, but look, here you are. And we've had this moment of this research coming out and you're a— what is the title? Science Envoy, I think, right?

Yes.

You know, you have to— look at these things for the United States in terms of global competition. But yeah, I mean, this is the thing that always comes up with quantum is, okay, so if they break that encryption, you know, you said maybe not time to panic, but I mean, people have flagged this as a moment to very much panic, possibly because for all the national security and, you know, financial system reasons. So if it is, if it is Now we're seeing, oh, well, this is looking more concrete and it might be going quite a bit faster than we thought. Why would we— why is it not time to panic?

Yeah. So I think—

and what are we doing?

So there's a class of algorithms that, you know, NIST has been validating these. NIST is an organization that sits under the Department of Commerce, and they've been looking at these post-quantum cryptography, PQC, algorithms, right? So they've been looking at ways that you might not rely on RSA or ECC, but you should start migrating to these. And what everyone's now saying is maybe migrating to these should migrate up your to-do list, right? And that's why it's not so much panic, but more like start doing that migration in the ASAP timeframe.

Okay. Okay.

And within PQC, and this will come up, you know, these algorithms, of course, the ones that NIST is looking at,, and these are being produced by the entire community. So it's not just that NIST is coming up with them. They've, you know, got these from the community. It's a, it's a game of cat and mouse. Somebody comes up with something, then somebody comes up with a clever way of breaking it. That is actually what it was like in the early days, at least from, you know, what I've read.

I wasn't around for RSA and, um, I mean, all this stuff is usually open source really, right? I mean, that's kind of the idea, right?

You want that.

Exactly.

Yeah. And so, so that's what's happening with some of these algorithms and people, put something out there and then 48 hours later you see somebody else saying, well, this is issue with it and then it's not as safe as folks claim. And so now I think that process is happening right now and we'll have a robust set of algorithms to migrate to the same way that RSA didn't become the standard overnight, went through a lot of testing and a lot of, I'd say, being punch around, and then folks said, okay, uh, so, you know, even if you can run Shor's algorithm on a fault-tolerant quantum computer, that doesn't mean that there aren't other ways of encrypting information. All right.

I'll embarrass myself possibly here. I mean, just as you're talking and I'm thinking back through this, I mean, you know, for decades there was this interplay between the government and, and technology. But, you know, if you look at the '50s, '60s, '70s, '80s, I mean, so much of this interplay was, was out in the open. I mean, it was, it was developing standards like Unix, like RSA, like TCP/IP, these protocols that all became the underlying force of computing and the internet, many of which have their origins in defense and government work, but, but was quite out in the open and, and nurtured by industry. And then clearly you have to get to some kind of standard in order to have this, this network of networks that we all rely on. Functioning. But at the same time, China has made quantum computing a national priority. Clearly, anything—

Yeah, it's in their 15-year plan. If you've—

Yeah, it's like one of the highest, highest things. And then, and then obviously, you know, if you had some edge on, on information security within your military or government, you could, you could use that. So, I mean, like, is this, this next race kind of begins and we're trying to address these issues? I mean, is the thinking, um, this has to become a global standard because that's how we operate and, and, and we need this for all these systems to function? Or are we— and I'm sure people dive in, I, I just don't know. I mean, or is the thinking, oh, we're heading to some kind of new world where if China gets there first or the US gets there first, then you're, you're keeping that secret for some period of time and, and, um, using it.

Yeah.

Yeah.

It's definitely a race. Right. And I think we maybe in some of the areas that you pointed out previously didn't have near-peer in terms of at least technology competitors the way we do now. Right.

And it was like academic. I mean, it just had different flavor to it. Yeah.

Now, you know, we see that at least in quantum from day zero, it has been framed as a global race. I think it's because Shor's algorithm was introduced, you know, mid-'90s and folks recognized this is important and this is important to national security. And that is something that has always been there in the DNA of, you know, people working in quantum. The other thing you're starting to see now is that, you know, every country has launched a an initiative of some sort, strategy, whatever they want to call it. And it articulates, you know, we want quantum computers, we want them here, we want these to be on-prem, we want to think about supply chain for these, we want to think about who and what can, you know, access these. Should we think about a supply chain equivalent for algorithms? There's a lot of different stuff that we're talking about now. That we didn't in previous eras of technology because there just wasn't really a reason to.

I mean, and maybe you're biased, or maybe you can speak freely since you're an academic, but I mean, since you are involved in some of these government discussions, I mean, are you impressed with the level of thinking around this, or are you underwhelmed? And, and it's— this is yet another area where, um, the US, you know, perhaps has a less coherent strategy than, than China.

You know, I've been very impressed by the American strategy in quantum science and technology. In fact, NQI was, you know, part one originally signed into law in 2018 during the president's first administration. There's a big effort now to do a bunch in quantum. There's, you know, lessons from the AI action plan, which also the president introduced early in the term this time. So I think there are a lot of people focused on quantum in the administration. I'm not going to get political about this, but just say that there are so many technologists who are in roles at Commerce, DOE, ARPA-E, and of course OSTP that are focused on quantum. There's a big effort in the DOW, in the NatSec ecosystem. Really looking at the technology. We talked about the draft for QBI programs. So I think folks are focused on it and the right folks are focused on it. Having said that, could more people be more focused on it? Could there be, you know, more support for some of the companies? Companies have raised a lot of private capital, but, you know, could there be more of a demand signal that also comes from government? Absolutely. And I think the new results that we saw should, you know, get everyone, uh, to listen.

Sorry, are we, are we considered to be like neck and neck with China, ahead, behind?

Uh, from everything that is publicly known, I think, uh, near-peer, neck and neck with the PRC is, uh, you know, the wisdom. Again, that is just what is known publicly, and that's it's hard to say, you know, if somebody puts out a PRL immediately following another result that comes from, you know, an academic group, you know, were they sitting on that result? Are they ahead? They're just putting out the result now, or did it just so happen, right? Yeah. So without kind of having that full transparency, we won't know. And I think in this particular technology, we just won't know.

Yeah. Okay. Okay. And then this might not be apples to apples. I mean, When I used to cover supercomputers for the longest time, the biggest machines were sitting in government labs and they were used to simulate nuclear weapons, do weather modeling, these, these national initiatives. As time went on, still some very large computers, but, you know, most of them, most of the world's largest computers are not actually getting ranked and they sit inside of Google data centers and Microsoft and Amazon and the government, you know, can still build very big machines. But I think, I think it's fair to say that these trillion-dollar companies have the world's biggest, most impressive computers at their disposal now. And it changes this balance of power around where's the most innovative work happening. If something like you had this massive breakthrough around quantum computing, would it, you know, traditionally it would have happened at a government lab or a university. Now it's almost definitely going to happen within a company. And I guess in the US, I mean, there is some, you know, there's interplay between the companies and the government, but I mean, it's just a different dynamic in this new era. Yeah.

And all of the companies you just mentioned that have amazing classical compute that's not ranked also have significant quantum efforts, investments also in other startups that, I'm sure have some relationship with supplying devices and supplying maybe the first devices. So I think we'll see a lot of breakthroughs. And the breakthroughs in this field are actually coming from industry, big and small. And I think there's a very productive partnership with the national labs now. There's this big initiative that Undersecretary Dario Gil launched, the Genesis mission. It's actually, you know, it's a big executive order. And it was exactly towards this, right? Like, how do we form these partnerships that bring in the best of what the DOE labs have to offer, what academics have to offer, and yes, what private industry has, which is really, really hard to replicate. Maybe there's no reason to replicate in these other settings. So I think It's bringing the full complement of what we can all do to this problem.

Is the thinking that if some tech company cracked encryption first, I mean, are they required to tell the government they've done this or are they—

Well, Google was very transparent with how they're thinking about this, how they're being very thoughtful about who has access and so that it doesn't turn into a free-for-all. And they have a plan, and I really appreciated that they put this out there in the form of an archive paper. So I think that's good. There are probably other companies that will follow their lead and put stuff out there, IBM, Microsoft, and that's helpful. I think it's something you'd want to know, though, you know, all my data sits with these companies anyways.

Yeah.

They have access. It's like, you don't need to break any crochet.

No, I just, I mean, there's like a sci-fi novel in, in, uh, it's the same thing in the AI world. It's like, you know, if Google or Anthropic got there first, it's like, do you, uh, tell everyone or do you kind of use this super intelligence for yourself? For a while to either make some money or do some interesting things, you know, before, before everyone else.

Yeah, well, nothing interesting happening with my data. Yeah, that's just, you know, you want the academic papers, maybe just read the archives.

Yeah. Well, there's a million places we could go. I just want to be mindful of time because there are a couple of things I wanted to talk to you about. Last time we talked, you were having lunch and you brought up quantum sensing, and I have to confess, confess I, I really didn't know what that was. I did a little reading over the last couple days and it, it was like fascinating to me. I mean, from what I read, and then again, I'll get you to fix everything I, I mess up, but, um, I guess clearly we have sensors in the world to track things like air quality and, and water and, and all this stuff. And, and obviously they're taking some measurements of molecules and and what's going on. And then it just seems because quantum systems are so incredibly sensitive, you know, when we build a quantum computer, you have to like shield it from the outside world in order for it to work. And so, so just from what I was reading, it sounded like these systems are so sensitive that you could actually pick up on like the tiniest change in a whatever it was. If you're trying to track like a virus spreading or some pollution or Is this like the general idea behind quantum sensing?

Yeah, so, you know, crappy qubit makes great quantum sensor. And because it couples to the environment and actually turns out our best measurements of the fundamental constants of physics to the, you know, 10th or beyond digit are actually based on quantum sensors. And there's a field, position navigation timing, that is thinking about how you can use effectively measurements of these same fundamental constants of physics, gravity, magnetic fields, et cetera, so that you can know where you are and you can have really good timing. Actually, our timing standards also from this come from atomic clocks, quantum sensors. And the idea is, sure, you don't need GPS if you can do quantum sensors on platforms. Now, the challenge historically has been that, you know, quantum sensors also, like quantum computers, form factor is hardly portable.

That's like one, you need like a giant, a building, or not a building, but some encased— Well, mini building.

That was all, that was true. And you know, if you had said we're gonna have a network of atomic clocks, these are going to be launched into space, folks would look at you and say, right on, what would we launch? A room. It kind of didn't make sense. But everything that folks have done in semiconductor manufacturing, everything that's happening in all these other fields, has actually brought the form factor down to, you know, poker chip size. There's a company, Mesa Quantum, they're actually based out of Playground here. They are making these and they were at, you know, they started with something that was more like a coffee mug, went to poker chip size, and now they're going to something that is blueberry size.

So what does that mean though? It's like, it's what, it's just like one quantum chip somewhere in there or something?

—some quantum— It's in their case, it's doing some local measurement of a magnetic field. And based on that and based on some analysis, you can actually, so with some useful edge compute, you can say, ah, I'm hanging out in California. And you can do that without actually being able to, or without needing to access the GPS satellites. Okay. Okay.

And it's, but it's a, And you work on this in your lab, right? You're doing quantum sensing.

So we think about solid-state quantum sensors, which is slightly different from, you know, the types of things that make clocks. So quantum sensors also come in all shapes and forms. But if you think about, say, a spin state, a defect in a material like diamond or silicon carbide, it is very, very sensitive to external magnetic fields. So it becomes a really, really good quantum sensor. Oh, okay. Okay. And you can use that. And we've thought about how you can, of course, it's not as simple as I put the sensor there and suddenly it measures something, but you have to have some way of initializing. You have to have some way of reading out. Same things you want to do for compute, but slightly different constraints. If you can do that, then you have ways of actually being able to do a measurement out in the air. I had for, the last few years, a program with the National Science Foundation, also with DITRA that was focused on how can we deploy such sensors, not just single, but network of these sensors and use that to say do anomaly detection or other kinds of interesting sensing problems. So that is the idea, right?

Is that you'd have, I mean, maybe I'm totally wrong, but you would put these sensors all Yeah, in the real world. In the real world. And then, and then some, um, some bioweapon is, is there's some, some—

maybe just somebody sneezes. Okay, okay. And it's a— and say there's one person that sneezes in a large airport, okay? In the interest of not making people— let's just say you want to know who is the person that, you know, is the was the reason everyone else got the flu. Yeah. Now, you could do that if you have a very high-sensitivity way of doing a measurement, but that itself is not sufficient. You also need to know things like where was the person spatially. You'd want to know, was it that person or the person slightly next to them? And so that actually becomes a problem that is known as data fusion. So I can have all kinds of sensors. Maybe I have some you know, cameras, and it's an airport. Maybe I have a temperature sensor so I can tell one of these people seems warmer than the others. And I have this very high sensitivity thing that picks up when this person sneezed, and that is what that sensor is picking up. So it could be a molecular sensor. I get data from that, and I fuse it with the data from all these other sensors, and something my group has worked on. And then I can answer a question that somebody might have, which is, Who sneezed?

And the idea is just that the quantum sensor is so much more sensitive to these changes that you're getting that initial data read on where something started.

Very, very— fine-grained. Exactly. It's very, very low, you know, in the signal-to-noise kind of language. This is something that would normally be below the threshold of a classical sensor that could also do similar molecular sensing. But you'd need to be above some, you know, number. Okay. And quantum sensors actually operate below what is confusingly called the standard quantum limit, which is the limit of a classical sensor where everything is operating at its best quantum limit, right? So, uh, SQL. And people in quantum sensing, by the way, talk about different bounds and saturating different bounds. But between SQL and the Heisenberg limit, which is set by the uncertainty principle, there's orders of magnitude. So effectively, you know, you can do a lot of interesting quantum sensing and very, very high sensitivity measurements. What's important here is that these are now sensors that can be deployed in the real world. They're not just, you know, sitting in my lab or a few other people who work in the field. And that makes them, I'd say, useful. Yeah.

And so you're saying, I mean, your lab's doing this, there's already a startup doing this.

So like how real is this? You know, there are actually multiple companies, different types of sensors focused on different problems. So Mesa is very focused on position navigation timing. They're targeting, you know, can you put these sensors on drones? Can you put them on other kinds of flying platforms, maybe on subs? That's a different kind of problem space. Then there are companies that are thinking, can we use this to detect a person sneezed? So the form factor and then the constraints of how you package the sensor look different. You could also imagine some of these eventually becoming part of wearables. We're not there yet, not there yet, both from a form factor and price standpoint, but that's something that the field is moving towards. And so you do location.

This is like magnetic field. It's—

you just stick the sensor in. It could be, you know, a response to a specific kind of molecule. Those would be different kinds of sensors. It could be gravity. These are all different looking quantum sensors.

And the atomic clock is a quantum—

Yeah, yeah. It's an example of a very old quantum sensor. Of course, there you're using— and of course, atomic clocks also come in different forms of what they actually sense. But yeah, those are, you know, so we have timing standards that are set based on clocks that NIST hosts.

Okay. Okay. Am I going too deep on this? Like, what is the quantum thing at play in an atomic clock?

You know, it's, it's, so atoms are so, I'd say, stable and also so versatile in how they get used in different things, right? You can think about neutral atoms you're using the spin states to do some kind of encoding. We could also think about using some of the lines. You could think about using the nuclear properties of these. And actually there was a pretty big advance that came out of NIST that was focused on these kinds of nuclear transitions. These atomic transitions are just really, really stable. They don't move. So they set That's everything that you'd expect in terms of drift, anything that is sensitive to temperature, environment, etc. Well, atoms behave the same way regardless. You can use some of these transitions as a very precise sensor.

Okay.

This is me trying to explain 2 semesters of atomic clocks, and I'm confident folks that work on this will say, Oh gosh, my whole field. And by the way, you know, Nobel Prizes have been given for lots of these things. I had no idea.

So I just got interested.

I just know any of my colleagues that I send a link to will say, really? That's what you said about my life's work?

Anyways. How is the— when I was digging around to prepare for this, It also seemed like you were— I was curious if you still are. I mean, quantum biology was a field of interest. Is that—

does your lab? Yeah. So we don't so much do quantum biology and we just do quantum plus molecules. And some of those molecules are relevant in biology. But I have to be honest, I know very little actual biology. I get into trouble every time I even try and name one of those molecules. But you could think of these same sensors being embedded in the human body and— go ahead.

Well, no, I just want to— can we go broad on this for a second? I mean, when I was reading about it, I mean, it was saying stuff like photosynthesis, how birds navigate might have these quantum principles to them. I interviewed someone for this show, and I think Richard Penrose is maybe quite convinced about this, that consciousness has Yeah, oh gosh, yes. Which I know is, is people are divided on that. But, but so there's this idea, and maybe it's not even an idea. I mean, is this just known that things like photosynthesis, like some of these massive processes in the world just have quantum things that we don't fully understand yet? Or this is like a field of, of, we're trying to figure out if that is—

There are certainly molecules that, or, you know, biomolecules that have quantum properties, right? And the question that always comes up is, is it coherence? Coherence by itself is actually a classical phenomena, or is there something truly quantum? Is it long-range entanglement? And so it kind of gets into this like very, I'd say, esoteric world of how do we quantify quantumness in these systems? And people have strong feelings when that comes up. But ultimately, You know, you could think of quantum sensors as ways to couple to these systems and maybe shed some light on these types of questions. I mean, molecules are ultimately quantum mechanical, and there's no doubt, right? So the question really is how much of those properties persist to classical scales, and does that then become the dominant behavior of the system? And that's where You know, there's a lot of debate whether it's maybe mesoscopically it's quantum, which we, you know, have other examples of, maybe somewhat macroscopically, but at a system level you could understand it classically. And then there's some people who say, no, you should really include these effects.

So that's— I mean, that's where I get excited that maybe I'm just embarrassing myself again.

No, no, please. This is like a fun topic.

I know Penrose is controversial on this stuff, but I mean, there is there is totally this chance, right, that we've been— whatever the number is— the last 80 years kind of had this like dominant physics that we understood a bit better, and it was like playing out in the world, and we have these rules we're going through. But yeah, but we know that there's these other properties to things, and we've just sort of had less time to explore them. We've had less machines to poke around. Yeah. On— and like when he's talking about, well, maybe consciousness has this roots in, in some quantum realm, and that's what we need to unlock these new understandings about ourselves. And I mean, this could be true. Like, there could be all these things in the world that we, we start to—

I mean, we certainly, you know, have folks who are thinking about these problems and working on these as serious research topics. I think, uh, there's, you know, how do you— and this is actually in itself a question— how do you go from like a loosely framed problem to something that is mathematically crystallized? And then you can propose concrete experiments that people can try. And, you know, this is the kind of blue sky work. My work on quantum sensing, you know, the more applied stuff was from actually, believe it or not, the NSF. But all of the crazy ideas I had support and still have support from the Gordon and Betty Moore Foundation. And it's all been focused on, you know, are there things here that we missed? To ask exactly, you know, you keep saying might embarrass myself. Well, I, you know, I embarrass myself by design in some of these grant proposals where I say, well, you know, this isn't a serious topic, certainly not something that's ready for prime time, not ready for NIH, NSF, but, you know, blue sky, let's ask the question. Yeah. And maybe it's a naive question, maybe, The question itself is how do you mathematically sharpen it? And that's exciting to look at. And then, you know, can we propose experiments where somebody can unambiguously say that it is or isn't right? And I think that's the whole point of science.

Okay. Okay. Be honest, because I can edit this out. Are you okay on time? Are you— am I torturing you?

No, no, I'm okay on time.

Okay. There's a couple other things I just want to ask you about.

Please.

Okay. Okay. So if we had to If I'm, if I'm a college student coming to UCLA and I'm like, you know, God, I mean, there's such big questions about where the world is heading, but it's like, should I be diving into AI or is, is like quantum, you know, the place I should— you, I'm trying to find kind of the most exciting cutting edge era of, of, you know, where I wanna dive into next. And maybe it's, but it's like certain personalities are gonna go other ways. But do, do you feel like you have to kind of pitch some of these kids on, on quantum, or there's plenty of people are, they're like, man, this is the most interesting thing I could be working.

So it's, you know, so interesting. This is, uh, the time of year where students commit to which school they're gonna go to. And, you know, folks are in that, I'm gonna be there this fall. So the pitching starts around now. And something I've noticed this year, which I had not really seen previously. Folks are scared of tech. They're, you know, worried that they're coming into this very fast-moving field and what if they place just the wrong bet and then they're working on something that isn't relevant in 4 years. And I honestly have not, you know, seen folks have that kind of fear. And what I, I'm a, you know, there's a lot of optimism here, but in this case, I think it's well placed. I tell folks, this is the most exciting time ever to be in tech in all of human history. And we, you know, get to invent new stuff. We get to work on new fields. You might work in a subfield that isn't the one that you look at when you graduate, but as long as it's developing some of these other skills, you'll be able to kind of, you know, pathfind after. So you don't have to think about your major, you don't have to think about, you know, your first research project, all of these things as setting you on some trajectory that you must then follow for all time. Yeah. There's gonna be a lot of pathfinding. And that's of course a very hard answer for students that have done everything right to get admitted, you know, to— Got their life all planned out. Right, and I'm basically telling them that's great, but, you know, we don't have, a crystal ball. It's not even a hazy one now. It's just completely gone. So you just have to actually do the thing that you enjoy doing. Yeah. And that's got mixed reception so far. So if you have ways I can pitch that better, please.

Well, I was just having the same experience 'cause my oldest son is about to go off to undergrad. And we were at a college this weekend with the president of the school was giving her pitch to all, you know, the students that have already been admitted. And then it was fascinating to me to listen the parent questions because I live in such a bubble here in the Valley. And this was a school in the Midwest. And, you know, like one parent was like, she was like, oh, if my kid goes here, you know, what are their odds of getting into law school? Is this like a strong enough school to help their chances of getting a lot? And I just remember thinking in my head, like, I was like, bro, I'm like, please don't go. You can't go to law school. I, in my head, I was just like, that's That's, that's the future I'm not sure you should be worrying about, you know. But then, and then some of the other, yeah, and then some of the other parents were, you know, they just went straight to like the AI, yeah, question. And it was, it's always interesting when you get out of the Valley bubble, and, and I was, I was just the stark reality of how scared these parents were of what was going to happen. And then, and at this school they were pitching like, oh well, we've just started this critical thinking around AI program because 'Cause this is a liberal arts school that sort of pride themselves on trying to shape the politics of the world a little bit. They're like, at least, you know, we'll have a chance to like shape the way the AI is developing. And I was just like, God, what a weird existential time to be going to school.

And then there are some folks asking, you know, what am I gonna get in a classroom? I mean, I think there's, you know, complete reset of how we think about information, and that is going to propagate into academia in all ways, right? It's not just publishing. I get this question a lot. I'm editor of a journal where folks ask, well, how do you distinguish human or AI written? Yeah. How do you think about reviews? There's all that. Folks are asking similar questions in the classroom. I think our relationship with information has just changed. It is, it has. That's, to me, it's very exciting. I can see if you're not a tech optimist or as in the tech bubble as we are, you might not have that view. But that is something I try and communicate to the next generation that regardless of what you learn in a classroom, like your development, your education is very, very important. And it will be important even if you don't end up using exactly what you learned in that, you know, writing class or one of my courses.

Yeah. And just my son going through this process, I've just been battling with that. I mean, I'm pretty pro-college for a whole bunch of reasons. Like it's just sort of the experience of being with a bunch of kids and like having that 4 years to, to, to grow. But I mean, to a lot of what you said. I remember, you know, I'm old enough to remember the dot-com boom and it was like when Google really came along with first the internet and then Google, just this idea that you had all the world's information at your fingertips. But then, you know, I've been playing with Claude agents so much in the last like 2 or 3 weeks and I, I was like, my God. I mean, I'm trying to put myself in my kids' shoes. I'm like, I could learn so much more about what I'm interested in so much faster. I don't care how good the school is at this point, uh, if you're just in— if you know what you're into or you're trying to accomplish, I mean, you just have— we went from the world's— all the world's information to just this like tutor who's just at your disposal, um, to be commanded as you wish, you know.

And so you could think about research experiences that students can have at the undergraduate level, uh, experiences with types of, you know, hardware, things that that you couldn't do in that way. And I think that involves, of course, universities completely, more or less overnight, revamping the curriculum. And I'm deeply, you know, sorry if folks were getting ready to just run the same class for another 2 or 3 years before, you know, they retired. But, you know, to all of those folks, I'll say, that's just not going to cut it. Nobody's going to come to that class. Nobody's going to sit in. And you know what a student learns from an active research, active or industry active or something, you know, faculty member is so different from somebody who's just kind of doing the same thing that they have for years.

That's— I think you just articulated what I was feeling, which is the old college experience feels so sedate. Compared to what it will have to be.

If I said that, I might get, you know, a tsunami of emails from my leadership saying, wait, what did you say about us?

But I mean, it feels like you will have to be this active experience to really get something out of it, you know, because it's just the idea of—

yeah, just asking questions or solving, you know, the same old problems. That's just not going to cut it.

Yeah. Yeah. Okay. Okay. I've asked I could ask you questions all day. You're so fun to talk to. Is there anything— I mean, there's a million places we could have gone. Is there writ large in the quantum computing field? Everyone always wants to know when it's going to be real. We've talked about this in different ways. To that, I have an answer.

Yeah. Okay. So the threshold— and I promised we'd get to this, right? Fault-tolerant quantum computing. That is the threshold at which you can do meaningful error correction. And all of these problems that are utility scale, let's say on the algorithmic side, become accessible, become available. And that first set of FTQC, right? One acronym you walk away with from listening to this, FTQC, that is on the horizon. And for some modalities, it might be as soon as '28. For some others, it would be, you know, before 2030. And, you know, the first fault-tolerant quantum computer we have will not be the last. There'll be multiple modalities. There'll be, you know, this is another question I get, which modality will win? Yeah. We will have fault-tolerant quantum computers that rely on different physical modalities.

Light, superconductors.

Right. Neutral atoms, of course, you know, a lot of photonics, semiconductor spin. I mean, there's just going to be, and of course, trapped ion. It's going to be different types of quantum computers. They're going to look different. And what we will find out is if they're all good at all problems or if some of them are good at some problems and some at a different set of problems. We just don't have an answer to that yet.

On that. But so, so 2028, 2030, things are kicking off and like a new, a new—

Yeah. And I think the slope has changed.

Yeah. Okay. Okay. Okay. Again, a million things we could do. Anything like super top of mind that we didn't—

I think that was it. Okay. Okay.

And when— so when is your marathon? When do you— when do we find out?

Oh, God. Okay. I have a set of races end of the year. Okay. And then a few more early next year. So I have time. I have to hit this qualifying standard before '28. Yeah. Before there's a fault-tolerant quantum computer. Okay. Let's say. I have to hit this 2:37 standard or at least give it what I think is, you know, my best shot.

And this is a standard. Are you trying to make the Olympics?

Well, this is just to run the trials.

Right, okay. That's it. But you might want, 'cause it's in LA.

No, no, no. No, no, okay. Even though the course is gonna go right by where I live, it is not on my, there's a big, again, between 2:37 and what the female pros run, which is more like a, you know, 2:15. To 18, that's like a world of difference. So I would just, you know, consider myself an athletic success if I do everything I can to hit the standard. And if I hit the standard, I'm pretty sure I'm gonna like cry at the finish line. And if I get to run the trials, I'm gonna be really, really happy.

Okay. Okay. I had one more question. It's out there, left field. Try to picture how your mind works. Oh no. Yeah, no, but like, do you, do you do, do you visualize stuff well in your head? I do.

Is that how you— Yeah, I visualize. I can. Yeah. Very spatial.

And you kind of always been that way?

Yeah. Yeah, I think so.

So like when you were in high school doing all this stuff, I mean, physics just made sense to you?

Yes. Yeah. I think that was one of the ways I knew that proof-based math wasn't the way to go. I enjoy types of math that's closer to the physics, which is more you can see a beginning and an end, and it's something that goes with it.

Okay. So were you like that, like, as a really young kid? I don't know. You don't know? Were you, uh, were you a big reader? Were you— or you— yeah, I was drawn to math and science, and I did a fair bit of reading.

Yeah. But I wasn't like, it's like, I'm gonna read everything.

The Word Cell Shape Rotator, uh, do you know this?

Yeah, yeah, I know. I'm just trying to think back to like I, I don't know. I think the, the other stuff just came easy. I could look at something and be like, oh, we need to end up here. This is going to be relevant. And then there's like, I do this version of something in my, in my head. Okay. Sorry.

No, no, that's helpful. I was picturing it because I just, I was, I always try to put myself in, in people's shoes and just kind of get a feel for them. Okay. Last, very last thing. Last time you were here, you had this bag. Yes. We should give people like a top tip. Maybe all your clan already know this because you guys all use this equipment. But what was, what was the brand of?

So it's Hyperlite, but the material itself is this ultralight material, Dyneema. Okay. And I think it's a feat of material science because it's exceptionally strong and it's exceptionally light. And now folks are incorporating Dyneema into all kinds of things. So You know, it used to just be climbing rope and things that both needed to be very secure and light, and you didn't kind of care as much about, you know, the cost. And then somebody figured out how to do Dyneema maybe more scalably. So it's still not cheap. I would not buy a backpack for, you know, school that is made out of Dyneema, but it's really, really light. So you can, you know, have climbing packs, uh, shoes if you're into, you know, ultralight, uh, running or climbing shoes. And folks are trying to, you know, make everything Dyneema. So I'm not sponsored by Hyperlite or Dyneema, but folks, if you're listening, which you could be, please, merch is welcome.

So, and does everyone in the hiking world already know about all this, or this this is—

I don't know if everyone ever, you know, so hikers also, there's the weekend hikers and, you know, there maybe weight is not as much of a thing. Yeah. If you're gonna carry something for days in bad weather conditions, suddenly, you know, there's this climbing phrase, grams make ounces, ounces make pounds, pounds make pain. So, you know, depends on what kind of hiker, but yeah, I'm excited for, new kinds of materials to make their way into climbing and running gear.

Well, it was super cool. I had never heard of that before. So there we go. We're just trying to give some practical— Practical advice. Yes. In the midst of a quantum computing discussion, I feel like we've done a service to the world then. Pri, thank you so much for coming in. It's been really cool getting to know you just over the last month or so. So I can't thank you enough for all your knowledge. Yeah, thank you. It's been fun. Awesome. Thank you. The Core Memory Podcast is hosted by me, Ashlee Vance, and/or Kylee Robison, or both of us together. It is produced by me and David Mickelson. Our theme song is by James Mercer and John Sortland, and the show is edited always by the John Sortland. Thank you so much. To Brex and Send Cut Send for all your support. And thank you most of all to everybody for listening or watching. We love you. Please leave us a like, a review, a subscribe, all those tremendous things. Thank you, and we'll see you again.