Hypersonic Physics, Deep Sea Life & Princeton’s Millisecond Qubits (EP. 17)
1 hr 33 minAudio
Hosted by Lester Nare and Krishna Choudhary, this episode dives into three breakthroughs stretching across aerospace engineering, astrobiology, and quantum computing. We start with a Nature Communications paper from Stevens Institute that experimentally validates a 60-year-old hypothesis underpinning hypersonic flight modeling. Then we head 3,000 meters below the Pacific to explore a newly discovered cold, ultra-alkaline biosphere near the Mariana forearc - a finding that reshapes the search for extraterrestrial life. And we close with Princeton’s millisecond-coherent transmon qubit, a materials science triumph pushing the quantum hardware frontier toward real-world quantum advantage. Summary Hypersonics without supercomputers - Stevens Institute validates the Morkovin hypothesis up to Mach ~6 using krypton-tagging velocimetry, confirming that “simple” turbulence models still work in hypersonic regimes and opening the door to viable, inexpensive hypersonic aircraft design. Life where it shouldn’t exist - University of Bremen researchers uncover evidence of a chemosynthetic biosphere in the cold, pH-12.6 serpentinizing fluids of the Mariana forearc, offering the clearest Earth analog yet for Enceladus- and Europa-like conditions. A millisecond qubit breakthrough - Princeton’s tantalum-on-high-resistance-silicon transmon hits 1.7 ms coherence, 15× the industry norm - drop-in compatible with Google/IBM architectures and a major step toward practical quantum computing. Show Notes Hypersonics - Nature Communications (Stevens Institute) Deep Sea Life - Nature Communications Earth & Environment (Univ. of Bremen) Princeton Millisecond Qubit - Nature (Transmon Hardware)
Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.
Transcript
From the show. Tap a time to play from there.
Guten Tag, Internet. This is your captain speaking, Lester Nare, joined as always by my co-host and our resident PhD, Krishna Chowdhury. My friend, how are you today? Doing well. You know, not as well as I could be because it's like raining and cold in Los Angeles. It's miserable. It's a little Seattle right now in LA. Oh my gosh. If this is what it's like in the UK, I understand why everyone's miserable. Yeah, yeah. It also reminds me of why I moved out of the East Coast after college and back to my beautiful home of LA.
And I'm not one of those guys who's like, we need this. You know, maybe like California needs this, but like in LA, the water is just going into the ocean. Yes. Like the stuff that's coming outside of the studio is just going. There's no benefit. Like the rain needs to happen where the reservoirs are. Yes. To fill the reservoirs. Yes. Not here. Yes. Right. We don't need this. There is a 6 to 12 inches of water flood on the road. Yeah, I had to jump over that. In the driveway?
Yeah, yeah, yeah. And there was a timing battle there because the longer I stood to figure out how to get over that puddle, the wetter I got. So I'm like, there's a calculation I was making. Like, where am I okay getting wet? I want people to understand like houses don't have gutters here. Yeah, yeah, yeah. Like literally there's no gutters. So everything just falls off the side of buildings. It's – Yeah. That's so funny. It's just – as I was mentioning earlier, rain has done more chaos and damage than any other culture war issue you want to bring up about Los Angeles.
That's true. Rain is our number one enemy. But I think – so we have a couple of good stories this week. Yeah. Yeah. We're on episode 17. We're about to wrap up season one, which I'm arbitrarily saying is every year is a season. Yeah, that's good. So we're getting close to the end of season one. This is the pilot season. Pilot season. It's kind of like Breaking Bad. It's going to be a cult classic. People thought the first season was slow and then all of a sudden it was the greatest. Yeah. Yeah, that's going to be us. This week, we're going to touch on three stories.
Yeah. Our first story is about hypersonic aerodynamics. So there's a breakthrough that might enable planes to fly 10 times faster than the speed of sound. Yeah. We're going to talk about some modeling stuff there. Uh, it's, it was in nature communications out of Stevens Institute, Jersey, Jersey's in two of our stories. Jersey's Institute. Yeah. I know Jersey gets a lot of flack, uh, union County born and raised, not born and raised, raised, raised. Yeah. Not born. Uh, But no Jersey slander will be tolerated.
Story number two, we're going to go deep in the oceans for a deep sea life story. We found some life in these sort of very extreme areas of the deep sea that we did not think was possible. Yes. And there's potentially some implications for exoplanet research and the search for astrobiology and search for extraterrestrial life. That's out of University of Bremen. And our third story is out of our favorite story. It is our favorite institution. Can you blame us? You can't. It's the best. So like, we just have good taste.
Yeah. A Princeton team has built a millisecond trans mom cubit breakthrough, which is going to allow us to actually do stuff with the quantum. I'm using the quantum as a term of art. The quantum. We should trademark that. Better than we have before. So again, this is out of Princeton in nature. right so our first story was the nature communications our second story was nature and earth environment but nature nature yeah now it's nature nature the real one yeah this is from first principles
So let's dive into our first story on hypersonics. So for me, as the UAP guy, Mr. UFO guy. I thought you'd like this one, actually. Hypersonics is an interesting... So the idea of these things that travel very, very fast. Very fast, yeah. Obviously, the number one arena that this operates in is military. Because hypersonics is a delivery vehicle for... nukes especially, is kind of like the thing that is sought after by any sovereign nation that wants global dominance and power.
This breakthrough, so the headline on this story is, a hypersonics breakthrough could enable planes that fly ten times the speed of sound. This is in nature communications from the Stevens Institute of Technology. But what specifically about this, what is the breakthrough that they're alleging is happening here? Yeah. So, you know, we want to travel faster. Yes. Right. This is something that'll give us access to faster air travel. It would be really nice if we just, you know, got to go to, if I could go to India in an hour and visit my grandmother, that would be really awesome.
Right. Pretty chill. But the age of the Concorde is gone and now it's sort of coming back. Supersonic travel is sort of coming back. Hypersonic travel is something that is on the docket. but it's a little more non-trivial compared to supersonic. Okay. So in order to get into that, there's some fundamental physics involved that I think is very, very cool. Let's talk about the Mach number. You know about the Mach number, right? It's the ratio of the speed of your object divided by the speed of sound. The speed of sound is about 750 miles per hour.
Planes usually operated around like what? 200, 300 ish. Yeah. Right. Um, And at the altitude that planes are at, you're a little bit lower. The speed of sound is around 600, 650, right? Because the density is lower. But at the end of the day, the air behaves differently when you are traveling at close to or... Above the speed of sound. Okay. And one of my favorite movies of all time is the right stuff. It was a movie about the early U S army test out in Edwards air force base.
And there's a great scene where Chuck Yeager is trying to conquer the demon in the sky. And the demon in the sky is the sound barrier. Okay. Because back then no one had gone faster than speed of sound. And it seemed non-trivial whether you could. Now, engineers were like, of course you can, because a bullet goes faster than the speed of sound. And, you know, Ernst Mach had that famous photograph of the bullet going through the speed of sound that we had covered in some earlier episodes. Yes.
So, It's obvious that stuff can go faster than the speed of sound and the atmosphere permits it. It's less obvious whether something as big as an airplane that's carrying a human being can go that fast. There's a scaling issue. A potential scaling issue. There's a potential scaling issue. Because what ends up happening is when you go faster than the speed – towards the speed of sound – the disturbance, the pressure waves, the disturbances that you're creating with your object are going to move away from your object at the speed of sound.
That's what sound is. Sound is the speed at which some disturbance moves through air. But if you're piercing through that... barrier, right? Then the disturbance is sort of catching up with you. Or you're catching up with that disturbance, right? And you're piling up air in front of you. And air goes from something called incompressible, an incompressible fluid, which is a way of saying that basically air doesn't change its density as we change its pressure and temperature to something where it does.
Now you can squish air, okay? And what ends up happening is at the speed of sound, you get these things called the mock cone, where the wavefront of all of that disturbance becomes a nice little conical wavefront in your device, like right around the device that's going through. This picture here is of a model airplane in a wind tunnel where the wind is moving faster than the speed of sound. So then relative, the object is moving faster than the speed of sound, and you're getting that really nice mock cone.
And the angle of that cone tells you how fast you're going. So like it'll be wider or more narrow based on how fast you're going. It's kind of like the wake on a boat, right? It's the same principle. It's like the waves in the water are moving at a certain speed, but your boat is piercing through that disturbance faster than the waves can proliferate. Yes. Right? And what ends up in water, you're not getting this compressibility thing because it's… It's just water. It's a two-dimensional sort of surface in three dimensions, so it can dissipate its energy in this, like, Z direction, right?
But in sound, the energy has nowhere to go, and so you're just piling up these air particles one on top of the other, right? So that's the pressure wave that you get when you go supersonic. Okay. Okay? We figured that out, obviously. We have the F-16s and all of these really nice military jets. We have the Concorde that used to be a— The only commercial supersonic jet that has not been replicated at scale yet. Since. Since. Yeah. I think there's recent things coming out of Boeing and NASA that are trying to resurface that—
And they're trying to make a version of the Concorde without the Sonic Boom. Without the Sonic Boom problem. Which is going to be very cool. I do know, because they're abusing me with advertising on X, there is a startup that is doing supersonic jet. They're building their own supersonic jet platform outside of the Primes. I'm not saying they're accomplishing it, but that's what the advertising is putting forth. I mean, nowadays with the access to computation and AI and all that stuff, I wouldn't be surprised if a small ragtag team of engineers cracks certain problems.
Which is fair. So... That would be very interesting. So, okay, we've gotten to supersonic, right? Yes. Then there's something called hypersonic. The definition of hypersonic is greater than Mach 5. So five times the speed of sound. Five times the speed of sound. Got it. And there, the air behaves even more differently. Okay? The physics turns into a different animal altogether. It's kind of like levels in a video game. And like when you get to level two, it's harder than level one. Yeah. It is harder than level one. And now we're here in level two, right?
What ends up happening is the physics of the air starts mattering. Before, you could just treat it as an ideal gas. Ideal gas meaning a bunch of point particles that are bumping around. Sure, like even the speed of sound, you can derive that. Actually, you know, in physics, at Princeton Physics, one of the... on the final for statistical mechanics was to derive the speed of sound from first principles.
And I remember being like, thank goodness, because like half an hour before the test, I saw this derivation and I was like, I'm just going to memorize it. And then it came up on the test. I was like, dude, I got this. And I just like, and just like wrote it down. And some of my classmates afterwards were like, what the hell is that? And I was like, you know, I just randomly happened upon that page. And I was like, this is something that I should probably just know how to do in case.
And it was the exact problem. Oh, my gosh. I love that. That's so amazing. It's a really cool problem where you basically treat the particles like billiard balls, like an ideal gas. And you start asking, like, how the dissipation, how these billiard balls will... bump into each other to create the effect of sound. Yep. And then how fast that sound is going to go in terms of temperature, density, and things like that of the air, right? And the mass of the particles even. Yeah, yeah. There's like, there's terms that go into that derivation.
Yep. So all that is fine before hypersonic. Yep. Right. Okay? When you get to hypersonic, then you start caring about what is the air made out of. Yes. It's not point-like particles. Most of it is nitrogen and oxygen. Okay. And those are diatomic particles. Okay. Right? Diatomic molecules. Yes. So diatomic molecules means you've got two atoms. Nitrogen is an N2. Oxygen is an O2. Yep. And those guys have other degrees of freedom.
They don't just move. Right. They also vibrate and they also rotate. Okay. So there's other ways in which energy can affect them. And so the ways in which they interact themselves and with each other, And with the object. And with the object. Start mattering. Because there's more complexity now when we're starting to get into above Mach 5. Above Mach 5. Because the energy scales correspond now. The kinetic energy that you're putting into the air is now at the same regime as the energy of this and the energy of this.
Of the expansion and rotation. You see what I'm saying? Yes. Yes. You see, that's if physics is always about energy scales, it's like, what energy scale am I operating in right now? The energy in the room is at what? Maybe 20 degrees Celsius. Right. So 270 degrees Kelvin at that scale, the vibrational modes aren't doing much. So you can, you can, you can just use PV equals NRT and be like, this is an ideal gas. Yeah, it's fine. It's fine. Fine. When you get to higher temperatures, higher kinetic energies, that's when the molecules start doing this. Yep.
Right? And then you've got to start caring. Because there's emergent properties of the particles doing this. Yeah. That matter in the overall equation of what is happening. Exactly. Exactly. At a low velocity, when I'm just moving my hand around... the molecules are just getting out of the way. But if I start moving my hand around at hypersonic speeds, the speed at which the energy that I'm imparting on these molecules is now making them go like this. I mean, this is the whole idea of the point that you can't just go infinitely fast because at some point you start the surface of your materials...
Yeah, it'll start melting. Melting and breaking apart. Yeah, like if I go really, really fast, I'm going to start imparting energy on these materials. Yes. Those materials will start imparting the energy back into the surface, the metal that I have. Yes. Right? And then maybe that energy is enough to start melting the metallic stuff. The volume of that energy. Exactly. Is getting so high that it is not just... And there's reasons why this makes a lot of sense, but... Yeah, yeah.
I mean, you might have read about this with a lot of your UAP research. I mean, it's this idea. It's why all of these advanced aerospace platforms, when you talk about... you know, the M22, the M35s, sorry, F32, F22, F25, 35s, like the B52s, like all these super advanced platforms. Material science is actually one of the most important aspects of these platforms because we're trying to have them operate both at altitudes and at speeds and with levels of maneuverability.
That are pushing these limits. Like physical limits. Physical limits. Of like the stuff that it's in even. And the understanding of hypersonics. The reason why hypersonics is still the realm of massive global power nation states is because it's very complex. Yeah, it's very complex. All of a sudden you can't just rely on the Navier-Stokes equation to like tell you everything. Right now, now, now it's, it's no longer just viscosity and density and pressure. There's like molecular effects.
You got to get down to the molecular level and that's why it's hard. Right? Yes. So when it comes to, when it comes to like making something that wants to move at a hypersonic scale, there's two tyrants that you have to deal with. Okay. Okay. There's two types of drag. There's a pressure drag, which is basically my, my plane is going through the air and There's higher pressure in the front because I'm piling up all the air. Yes. And then I'm creating an effective vacuum behind me. Yes. Right? So that's going to create a force because there's high pressure here, low pressure here.
Just like a piston. Yes. There's going to be a force that's going to be like, no, I don't want you to move that fast. That's the pressure drag. Yes. Yes. Then there's also skin friction drag, which is the air is moving over my body. Yes. Right. And I'm like tearing the air apart in some sense because the part that's stuck to my vehicle, my flight, is going to be moving with me. But the air away from me is stationary. Right. So there's going to be that like velocity difference. Yes. Right.
And then there's another thing, which is the heat. All of that energy that I'm imparting, if I'm going hypersonic, that heat is now changing the chemical structure of the air in front of me. Yes. And that heat load, that aerothermal load from the viscous dissipation is going to be something that I need to care about. Right. Right. So there's all of these like little things. Yes. And the source of most of my problems comes from something called the boundary layer. Okay. So imagine you've got an airfoil. It's like a wing.
Yes. Yes. What I was saying earlier is there's a thin layer where the velocity shears from zero relative to me. So I'm moving through the air. I'm moving through the air. All of the molecules that are stuck to my wing are moving with me. But all of the molecules that are even a few millimeters away from me are far gone. There's got to be some continuous way... That I go from zero to my speed.
Yes. Right? Yes. And that's that boundary layer. Got it. Okay? That boundary layer turns from laminar to turbulent. Okay. Okay? When it fully attacks the air, it's laminar, meaning it's quite smooth. It's very smooth and it's very nice. Yes. Okay? There's a continuous sort of transformation from zero to whatever my velocity is. Yes. Later on, that laminar flow is going to turn into turbulence. Yes. And the turbulence is going to be, there's going to be these eddy currents.
There's going to be like weird, like circulations happening in the back of my wing. And that transition from laminar to turbulent is a very poorly understood event. I gotcha. Okay. Okay. Yes. Okay. Like seeing where that happens in my wing, how that happens when I change the deflection of my wing. Yes. You know, all of these things are, you know, we've got models for it and we've got ways of understanding it. But at the end of the day, this is an extremely nonlinear phenomenon, which means that, you know, there's not a simple, nice thing.
Like differential equation that I can just like do-do-do-do-do solve. And like I have like a solution that is like in closed form. Okay. This makes sense. And is it also the case that because of the lack of – to be able to do hypersonics is extremely expensive? Yeah. And it's also classified? Yeah. And so the amount of actual source for data to inform the models – is not really accessible in terms of having a feedback loop for how you iterate on your model's predictions.
Yeah. And so you're kind of extrapolating from what you can get, but it's an environment because of this laminar flow to turbulent flow and these new problem sets that happen when you start going hypersonic. Yeah. Yeah. the ability to have a feedback from real world, real world data sets is kind of a chat. It's a little bit of a challenge. It is a little bit of a challenge. And you know, I will, I will go out on a limb and say that even the classified programs haven't really crap cracked like greater than five mock.
Okay. Okay. Yeah. Because, because some of this fundamental physics is just like so foreign. Yeah. That like, there hasn't been fundamental research that's happened to figure that out maybe they've cracked it for small things yes right missiles okay right but like you're putting a human in something yeah that's going hypersonic yeah no that's different no that's that's actually no right so that's a difference you could be going at mach 10 you could be going at mach 10 in the upper upper upper earth's atmosphere yeah right yeah there's not that many particles the density of air is low
I mean, okay, is the International Space Station going at Mach 90 or whatever the hell? Right, right, right. It's not really a thing. Yeah, that's fair. Right? That's fair. But when densities are high, it's a different— It matters. It's a different animal altogether, right? And even—I mean, one of the things that I would suggest is— The space shuttle, for example, right? The orbiters, they had these over-designed tiles, the heat-absorbing tiles on the very end. They were over-designed because predictive models of that turbulence, of the laminar from turbulent background, were so bad that they were just like, okay, I'm just going to make this like 10 times what I need, right?
Because I don't know where that transition is, and I don't know the physics of that. So I'm just going to over-design for safety reasons, right? Yep. So it shows you that – I mean I'm sure like the top echelons of NASA have access to some of the hypersonic stuff. But I don't know actually. But a key point is as far as we know at this point in time in the public data realm, hypersonics are exclusively for – payloads like missiles not for human travel exactly yeah those are very different things like like like a human life is worth way more than some nuclear thing that i just like made as i'm testing yes you know no no but that that that is the that's the key thing that is a key thing that's the key thing and that's where this paper comes in okay okay
The central challenge of hypersonic design is predicting and modeling that turbulent boundary layer. And this paper is doing an experimental model. addressing of that problem. Okay? And that's why I think it's very exciting. Okay? What it's doing is basically testing a 60-year-old hypothesis. Okay? It's called a Morkoven hypothesis. And what it says is the essential dynamics of this very high-speed, high-mock turbulence
is not fundamentally different from low mock turbulence. Okay. Okay. Morkoven came up with this in 1962. He was looking at some preliminary data, and he said that this seems to be true. Okay. It's essentially the same. What he's saying is, At high Mach, error is going to become compressible, and there's going to be all of these density fluctuations and temperature fluctuations. But what I can do is I can take the physics that I know from low Mach, and I can do a transformation.
Basically, I just adjust the parameters for density, and I adjust the parameters for temperature. And I'm going to get the turbulent effects that I get at high Mach. Okay? This is a non-trivial statement. Okay? Because what it's assuming is the essential physics at the high scale, at the vehicle level, is going to be very similar. Right? It's saying all I have to do is adjust some parameters. And I'll get the right physics. There's nothing like weird. That's happening. That's happening.
If it's raining outside, I can just put on a raincoat and I'm fine. Yeah. There are no other derivative things. Like the addition of a single item that is trivial. Yeah. Yeah. Yeah. Yeah. Yeah. I can just adjust some parameters. Right. It's not that rain is like acid rain. That's going to melt my raincoat and kill me and all this other stuff. Right. Right. You know, which is which is like it is it is it is filling in and it's not in this. It is saying that this maps in both domains in a way that what we're saying is that is a big deal that we're saying that we're saying that.
Yeah, we're saying that there's some there's some trivial transformation. Yeah. That can scale the really complex compressible stuff. Into just a couple of knobs. Yeah, into just a couple of knobs, right? That's crazy. And why that matters is because right now we have simulations that deal with this kind of stuff. They're called computational fluid dynamics. They basically churn physics equations in a computer and tell you... how the air is going to flow around an object going at a certain speed, given the air density, given the temperature and things like that.
Formula One people do it all the time. This is what video games use for like water. Yes. You know, like water animations and you're uncharted for or whatever. Yeah, exactly. They need to use models like this. They use models like this, and people at aerospace companies use probably more sophisticated models. But at the end of the day, it's just like it's resting on this hypothesis, right? Because there's two ways to do it. There's a direct numerical simulation called DNS. That's the gold standard. You're basically solving Navier-Stokes for tiny, tiny little packets of space, and you're getting really granular with it.
Yep. Yep. Doing a direct numerical simulation for something like an airplane, we don't have the computation. I was going to say, like, we barely have the computation for me to get an image of me making eggs in the morning out of Chachabiti. Yeah, yeah, yeah. Let alone this multipoint. Yeah, this is, like, imagine every single little part of air that's interacting with every other little part of air, right? How many supercomputers are there that you're going to, like, rent time on? Correct. To, like, deal with every single iteration of your design?
It's not going to work. The point is classical computing does not really have the horsepower to do like physics-based or physics-level models without getting to supercomputer scale. Exactly. Yeah. Yeah. And there's not that many supercomputers in the world. There's not one like just here. Right. Right. Right. And then the other – so what we usually do is we use something called a computationally cheap RANS. It's called Reynolds Average Navier Stokes. It's basically a way to like blur out a lot of the physics.
So on the left, you've got direct numerical simulation. Yes. And on the right, you've got the Reynolds average simulation. You can see that it's kind of a blurred out version. Yes. Of the direct numerical simulation. Yes. So it's giving you a good enough answer. Yes. Okay. Yes. Yes. But here's the thing. Every single one of those simulations, especially when you're trying to go into this Mach 5 and beyond level, it's assuming explicitly Markoven's hypothesis. Okay.
Which is that the physics up there is basically equivalent to the physics down here up to a certain scale factor that we can calculate. Mm-hmm. Okay? Mm-hmm. The key thing is this. No one's actually done the work. I love it. To show that this is equivalent. Yes. Right? Right. That we can assume these things. We've just sort of been like, yeah. Yeah. Probably. Right? We've got other things to do. And then came some experimental data that suggested not.
Right. Okay. Okay. So there was a conflict. Experimental data came out of something called PIV, which I'm going to go into. But there on the bottom, you're seeing the red is the experimental data. Yeah. And the green is the DNS, which is the direct numerical simulation, the really high fidelity thing. Yes. Yes. And there's a discrepancy. Yeah. You see that discrepancy? And it is statistically significant. It is statistically significant. You're going from something like 0.4 to like 1.1 on the y-axis, right? That's a 2x almost.
Yes. So. Okay. Right. Yes. Now we're in a conundrum. There's a problem. There's a problem, right? If we want to trust this thing, if we want to trust Mark Hoven's hypothesis to beyond Mach 5 in this hypersonic regime, we have to answer the question. One, is Mark Hoven's hypothesis correct or is the experimental data correct? Right? Right. One of those two is the thing. Is the thing. Is true, right?
Right. So... That was the question that these guys at Stevens Institute were trying to handle. Shout out Jersey. Shout out to Jersey because here's what they did. They said, okay, direct numerical simulation is like... Kind of the gold standard. Sure. Understandably. How did they get the experimental data? They use something called PIV, which is particle image velocimetry. Okay. It's a state of the art optical experiment. And here's what's happening. You got some fluid. You put in some particles.
It can be dust particles, oil particles, alumina dust. And those dust are going to track... Where the fluid is moving. Right? Got it. That's the idea. It's kind of like when you put the dye in animal testing to make the thing show up when you look at it. Exactly. In certain areas. Yeah, yeah. Because the fluid is just moving around on its own. But you need to be able to visualize it somehow. But you need to visualize it somehow. So you're going to put some tracer thingy that's going to track where the fluid is moving. Okay? And it's going to be seated with these microscopic tracer particles.
Okay. But here's the problem. Those tracer particles are quite big compared to the air itself. Okay. Right? The air is made out of nitrogen and oxygen. These are single molecules. The particles are made out of, let's say, 10 to the 5, 10 to the 10, something like that, number of these. So there's going to be a particle lag there. There's going to be, the air is going to be pushing this particle and then the particle is going to move and then I'm going to image it and create this velocity field.
Right. So there's a delta between the thing we're tracking and the thing we actually care about. Yes. Because of this scale difference. Yes. That's not going to matter at low speeds. Yeah. But at Mach 5, Mach 6, now all of a sudden, maybe it matters. That's the question. Does it matter? Does it matter? At high speeds, the fact that my tracer, my dust particle or whatever, is so much larger than the fluid particles themselves, does that matter? Yes. And here's what they did.
They said, okay. Instead of having tracer particles, I'm going to use something called krypton tagging velocimetry. Krypton. They're going to use krypton gas. Okay. And the krypton atom, which is a monoatomic gas because it's a noble gas. It doesn't make molecules. That's going to be my tracer. Because it is closer in size to the thing we are trying to trace than the other aluminum dust and other things. And so we can... The point is, if we do it this way, we can start to rule out the lag problem as the source of the issue.
Because the Krypton is just going to be going with it. Yes. It's going to be moving at the same level and pace and all these things as the actual air. And this is the part that was like a little bit non-trivial to me, right? It's like with dust, it's really easy to set up a 2D laser field. and image the dust. Because the laser... Remember, from first principles, how do we image something? We've got to capture light. So what do we do? For the dust, for the particle image velocimetry, we set up a 2D laser field
The dust is moving around. The light from the laser bounces off the particle, and then it goes to our detector, and we take an image. With a single Krypton atom, how are you going to do that? Here's what they did, and I thought this was very cool. So they got a laser, and they used the laser to ionize Krypton. They have this two-photon procedure where they focus all the laser beam into a tiny spot where the krypton is getting seeded.
And if two photons interact with a single krypton atom, it's going to bump up in energy level and ionize itself. And then you're going to be able to track it. And now this krypton is an ionized thing. And when it goes back to its normal state, it's going to release a photon. The krypton itself is going to create... The photon that we're going to use to track it with. So there's like a read. So there's a write mechanism, right? We're writing the Krypton. We're tagging it. We're tagging it, yeah, yeah, yeah. And then we're reading it later. When the photon's released.
Yeah. And we've basically created read-write level permissions. Yeah, on like an atom. On an atom. And we're tracking it as it goes through this Mach 5, Mach 6. I hope people who are watching and listening to this understand how outrageous what you just said is. That's so... And it took him 10 years to figure this out. Obviously. Credit to them, dude. That's such a cool idea. That is a really... That's actually really... Right?
Because you got to think like... That's really clever. Yeah, it is clever because I want to trace how the gas is moving around, but I want to image the gas itself. Right. Right. How do I image the gas? I can't image atoms. Right. I can image dust particles because they'll scatter light. Yes. But if I make the atoms themselves like this weird light bulb, yes. This reminds me of the story we talked about in episode 16 last week about how we were used, the camouflage story about how we're manipulating biology in order to generate an outcome based on its existing evolutionary processes and building into that infrastructure in order to get the outcome we're looking for.
Just conceptually, it's like using the fundamentals of physics. It's like we're eliciting... the reaction yeah that then enables us to do the thing we want to do based on the fundamental principles yeah uh we're not just like we can't just force the answer we kind of have to like work with the universe a little bit yeah yeah yeah yeah yeah and so and so and krypton was the way to do it down was the way to do it that's i think that's i think that was really cool that's really and so what they did was in the experiment they have something called the stevens shock tunnel which is at Stevens University.
It was funded by the Office of Naval Research. It's an impulse facility. And what it does is it creates these shock waves that move through that tunnel at Mach 5, Mach 6. So you've got some apparatus where the krypton moves through the tunnel at Mach 5, Mach 6. You can now measure. You can ionize the krypton in one part. You can track it. And then you can get that level of fidelity that was never possible with the older method. Yes.
Right? And the results speak for themselves. You get the mean velocity curve, and the krypton tagging KTV data perfectly aligns with the DNS. Right. As opposed to what we looked at earlier, where there was that delta, that discrepancy. Meaning we're now actually tracking in a modeling context close to the direct... uh actual like this the the the direct value we see in that really high fidelity gold standard yeah we're now able to kind of replicate that with like less with yeah yeah and now it's in the real world and we can say okay experimentally we can the model is actually tracking the model okay that's that's you know yes it's like it's like yes the the particle stuff that we were doing earlier with the dust particles that wasn't good enough at these at these velocity regimes
Above Mach 5, we now have a regime, a velocity regime that can work with lower without needing a supercomputer. Exactly. We can now, we can rest easy that the simulation framework that we've been using is valid. Is lore accurate. Yeah. Yeah. It's like, it's... It's definitely less fidelity, but we don't have to, we don't need a supercomputer to, let's say, test a bunch of designs before we like go into production.
This hugely reduces the scale of engineering, right? That makes sense. I can now try a bunch of different things on my computer. It can be a big computer, let's say, but it doesn't have to be a supercomputer. It could be an M5 Mac. Yeah, exactly. Markoven's hypothesis is now experimentally validated at Mach 6. At Mach 6. Right? Yes, got it. And that's kind of like the key point is like this 19, was it 62? 62, yeah.
Sort of hypothesis. Yeah. has now been validated a scale of high fidelity measurement above Mach 5. Yeah, and we hadn't gotten that far before, experimentally. Experimentally, we hadn't really known. Is the hypersonics regime, as opposed to the supersonics regime, have something where Markovin's hypothesis is no longer valid? valid above this limit. Yeah.
And we've now said, no, no. No, we're good. We're good. We're good. So now we can start actually pushing these measurements. Yes. Right? To greater fidelity. And we can also start, engineers can be a bit more confident about... the designs that they produce in silica in the computer. Yes. And they, and they try to validate it. Yes. Right. You can now be a bit more confident about, okay, this thing works in the computer. Let's try to actually make it right.
This is, this is, I'm looking forward to it, dude. Hypersonic travel before I'm, before I'm out of here. That'd be great. That'd be dope. That'd be fantastic. That'd be dope. That'd be fantastic. I mean, it's probably going to be like $10,000 a flight, but. As we know, they always got to monetize. Yeah. Before we move on to our story number two, I just want, it's funny, the timing of this, because literally earlier today I saw, or I think it was today. Yeah. Sam Altman made this post about. Yeah.
Anyway, long story short, there's a company that has reached criticality on a nuclear reactor infrastructure that is venture capital funded. It is the first BC funded entity that has been able to show the ability to reach the splitting of the atom in a new reactor architecture. And they did it in computer. Oh, okay. And then this was the first non-heat, no heat, no power, but like, I guess, physical test of that simulation.
And they were able to theoretically reach that point. Yeah. It is just another – all these real-world examples of simulations moving to real world. And I don't think the everyday person understands how much we depend on modeling and simulations. It's insane. To be able to validate ideas before we even start putting – No, because that stuff is expensive. It's so expensive. Making stuff in the real world is super expensive. So whenever people are building new chips, the first thing you do is validate it in silica, right? You have CAD designs and you have all of the software.
Whenever NVIDIA makes its next H4000 or whatever the hell, right? It's going to create so much testing within the computer because we just know physics. at this level so well, right? We're not going into CERN trying to find the Higgs boson. We're like, we need to, I don't know what the mass is. I know what the mass of the electron is. I know what temperature, like, you know? So we know physics so well that it's just way cheaper to do stuff in silica, in a computer, before we...
Rig up a whole factory to actually make stuff. Yeah, you're right. Which makes total sense. And it's an important... I got to check this out. That's pretty cool. It kind of took off on X earlier today. And I'm not... I did try to do some very brief validation. It does appear to be the first time that a private... venture-backed like investor-backed like private capital has been able to do so which is that's that's something and it was i always look at the comments on x it was part of a doe program to get three uh private uh versions of of uh reaching criticality by 2026 so it's a part of a sort of an ongoing doe program yeah um which made it a little bit more
Yeah, definitely a little bit more legit. A little bit legit. I mean, this was funded by the Office of Naval Research. Same idea. A lot of federal government funding. It's all... Another reason why we need more federal funding. All the beautiful things we love. You know, it requires... Yeah. We're going to move on to our second story, which is about microbes. Yeah. And a new biosphere where we thought life...
was impossible on Earth. Yes. And it has some implications potentially for astrobiology. So the headline on the story, Life Found in Place Scientists Thought Impossible. Yep. This is a paper out of Nature Communications, Earth, and Environment, a biomarker evidence for our serpentinite species. chemo synthetic biosphere nice at the mariana for arc this is out of the university of bremen german university uh many people are familiar with the mariana trench people sort of through a variety of sorts might be like oh yeah like there's like these hot vents and there's things that live in these hot vents and that's the most extreme place but but there is something really interesting here about life defying the limits of
that exists in the deep sea in this paper. Yes. Yeah, I think it's very cool. The core concept is life is constantly evolving in places that we never thought was possible. This one is happening near the Marianas Trench. And it's very different from the deep sea thermal vents that you were just looting to, actually. Okay. Perfect. Because the deep sea thermal vents usually happen at places where the oceanic crust is diverging.
Like in the mid-Atlantic ridge, where the crusts are like splitting apart. So there's a lot of volcanic activity happening, right? Right there? Yep. Yep. And so there's a lot of like sulfur gas and all sorts of stuff coming out. And then all of the bacteria are capitalizing on that. This one is a subduction zone. Okay. The Marianas Trench is something where one plate is going under another rather than two plates splitting. Two plates splitting.
Right. In the Mid-Atlantic Ridge, they're splitting. Here, one is going under another. Right. Mm-hmm. So it's a little bit different, and I think that's what's really cool. That makes sense. So life is constantly trying to redefine boundaries, right? We've got extremophiles everywhere. Yellowstone, the life in the hot springs, that's something that's in every single biology textbook. You've got hypersaline ponds in the Atacama Desert that somehow manage to have bacteria. You've even got like nuclear reactor cores of power plants.
They've got weird bacteria. And I want to cover that in another future episode because those things are very cool. This new paper, it's a landmark 2025 paper, and it provides biomarker evidence for serpentine chemosynthetic biosphere. Okay. In the Marianas Trench. So the setting is something called the Mariana 4 arc. Right under Japan, east of the Philippines, is the Marianas Trench.
the sort of crescent of the Mariana's Trench. The Captain Hook? Yeah. That hook is the Mariana's Trench, right? That is one plate subducting under another, okay? And right, not in the trench itself. The trench is like the bottom of that subduction zone, right? Where the ocean is like... a few Mount Everest deep or whatever. It's very, very deep here. This is a part of the, of the subduction zone that's at the top. Okay. It's only about 3000 meters below the surface, but it's still impossible conditions.
Okay. Hyper alkaline pH. So you've got 12.6 pH. Jesus Christ. Okay. Water is, Is that seven? And we're going to get into that a little bit. Okay. It's really cold, 3.5 degrees Celsius, right? So that's about maybe, what, 40, 30, 37 degrees Fahrenheit. Not very cold. Very cold still. I mean, I might have surfed in that kind of temperature, but I didn't have a good time. Yeah, exactly. But somehow these bacteria are living there.
And there's also a nutrient scarcity. There is profound limitation, especially when it comes to phosphate, right? There's not a lot of phosphate there. What's really cool is this paper finds life there and then tries to characterize how that life is doing living down there, which I thought was very, very cool. Yes. The geochemical engine is something called serpentinization. Okay. It's a reaction between water. And whatever mantle rocks are down there, because this is like very close to the Earth's crust, right?
So you've got water, you've got some rocks, and what is output is a mineral called serpentine. You've got a lot of molecular hydrogen. Okay. H2. Okay. And H2 is a very energetic molecule, right? You can burn H2 and create energy and water and CO2. So that's what's very interesting here. Okay. And the analogy is like, you know, when you've got iron and minerals, that iron splits and reacts with the oxygen and creates hydrogen.
Because the water, it's going to strip the oxygen out of the water and create H2. Yep. Yep. Something similar is happening down there, right? The minerals are reacting with the oxygen in the water and liberating the hydrogen. Yep. Okay. And then that hydrogen then goes and then creates methane because there's going to be carbon there too. Because there's a free hydrogen floating around. Yeah, free hydrogen floating around, and that fuel is then going to create abiotic methane. So this is not methane coming from life.
This is methane just coming from the chemistry down there. Got it. Okay. You know what I mean? Yeah. No, that makes sense. Yes, that makes sense. And so that's what's really interesting. We got to understand sort of the tectonic setting here. Okay. In the Mariana's Trench. So I told you, right, it's a subduction zone. Yes. You've got one plate going under and another plate going over. And the fluids in that subduction zone, in the mantle itself, are what create this serpentinization. You've got like a sort of chemical process that like rejuvenates the ocean floor with new chemicals and things like that.
But that resulting chemical is like... The water there is like really, really messed up. Okay. Which makes sense. Which makes sense. It's like way deep. Right. It's near a subduction zone. But the pH is 12.6. Okay. That's like bleach and oven cleaner. Yeah. It might cure COVID. I heard that. Yeah. Oh, yeah, yeah, yeah, yeah. So we just got to go to the Marianas Trench. Right. Get it. Get it. And then drink it. Yeah, yeah, yeah. Got it. Yeah. But somehow, so these bacteria don't have COVID for sure because they live there in that bleach and oven cleaner of 12.6.
And just to tell you just how bad that is, right? So the pH of water is 7. Water is at a pH of 7. And what that means is pH means percent hydrogen ion. Okay. It's a log scale, meaning that water has a concentration of H+, which is 10 to the minus 7 moles per liter. Moles is a unit of the amount of stuff. It has to do with Avogadro's number. It's about 10 to the 23 thingies. Like a dozen is 12. A mole is 10 to the 23 times 6.
So I'm not going to order a mole of cupcakes from the bakery. No, but you would order a mole of water. Actually, no. A mole of water is like, I think, 55 liters, so maybe not. But it's something like that. A mole of carbon-12 would be 12 grams. Got it. So that means that there's about 10 to the 15 hydrogen ions per liter in just normal water that we drink. Okay. Something with a pH of 12.6 means that there's 10,000 fewer hydrogen ions and 10,000 more OH- ions because the oxygen, the H2O can split into an H+, and an OH-, right?
Yep. 10,000 more OH minus ions and 10,000 fewer H plus ions. That is an incredibly basic environment. There's all these OH molecules with a minus, right? They're an ion. It's extremely reactive. And the fact that anything can just like stay stable down there is incredible. Got it. Okay. Yeah. Right. So what we're trying to identify is the base environment in which we are trying to even operate. Yeah. is not ridiculous is not conducive to operation yeah yeah it's ridiculous to like to like stuff that we know stuff that you're right right right right right so here's how they can uh they um they got the sample they have something called a gravity corer okay it's a it's a it's a core so what they do is they they get like a 1.4 ton weighted steel pipe and they just shoot that down
Into the ocean floor. Yes. And they grab like two meters worth of soil in there, right? They bring it back up into the surface. Got it. Then they take the mud out. The mud has this blue hue, which is serpentine, blue serpentine. Okay. There you can see. And you can do mineralogy on it. You can do x-ray diffraction and thermogravity, and you can identify minerals like serpentine. Okay. And brucite and things like that. And what you want to do is you want to find life there, right?
The immediate thing that anyone would do, it's like, okay, I want to find life. What do I do? I look for DNA. Yeah. Easy. Yeah. I look for DNA. This thing has about 10 cells per cubic centimeter. Right. That's really low. Right. Okay. Usually you got something like 10 to the 5 cells per cubic centimeter in soil. Okay. Okay. Soil has a ridiculous amount of cells, a ridiculous amount of DNA. Okay. This thing has only 10. So what's going to end up happening, if you try doing something like PCA, which is polymerase chain reaction, it's basically a way to amplify DNA in your sample.
So whatever little bit of DNA you have, you multiply it and then you try to, you know, put it through your measurement sticks and you try to figure out what the thing is. Yep. If you try to do that, you only got 10 living cells. Most of the DNA that you're going to have is just going to be dead stuff that's, conglomerated on the ocean floor. Right. Right. Because the ocean has just raining dead material all the time, all the time. Phytoplankton, maybe a whale died somewhere.
So you're going to have all of this random DNA and not the actual DNA of of the stuff living in there. Not a perfect analogy, but if you try to take a photo that is really low resolution and blow it up to really high resolution, you're trying to take pixels that didn't exist and make them larger. And this is just going to be a larger version of not a lot of information. Yeah, exactly. Like if you, if you did one of those like AI, um, deep pixelizers of a really bad photo, it's just going to start making up stuff, right?
Because you don't have actual info. It's, this is dead information from potentially a lot of dead stuff. A lot of that stuff that's just been raining down on the ocean floor. Yep. Right? So you can't really do that. Okay, that makes sense. So instead what you do is you go for trying to detect lipids, which are fat molecules. Okay. Lipids are the stuff that make up... the cell membranes of your cell. Every single thing has this. Okay, a lipid membrane. You need a lipid membrane because how do you define life?
Well, you need an inside and outside, right? It's the border. How do you define a country? You need a border. Yeah, exactly. Yeah, but how do you define life? You need a border. You need a 2D border for a 3D object, right? And that is always a lipid, a bilipid membrane, right? Yeah. So every single thing of life is going to have these lipids. Now, there's going to be two types of lipids. There's going to be the living lipid from whatever's living in there, and then there's going to be the dead stuff that's raining down. How do we actually discern between the two, right?
And so here, we're going to have two different types of lipids. There's the core lipid, which is lipids that look like this. They've got a part that likes water. On the top, the purple one, and then a part that doesn't like water. And what you do is you make two membranes out of this. So there's a part that likes water that is facing the outside, that is facing both the inside and the outside. And then the inner part is the part that doesn't like water. So now you've got a way to separate water from water.
Right? So you've got a way to separate the cytoplasm of your cell, the stuff that's inside, from the stuff that's outside, based on a wall that the water really doesn't like to cross. Yes. Those lipids have two parts. There's the part that likes water, the hydrophilic part, and then the hydrophobic part, which is the part that doesn't like water. If you see those intact... That means it's pretty recent. On the other hand, if it's died for a while, those two will have clipped off. And you're not going to see the two together.
So you've got these intact polar lipids. Okay. And that's what these guys were looking for. That makes sense. The point being that by finding these bipolar lipids. Yeah, polar lipids. Polar lipids. Yeah. Both sections still being in close proximity, meaning that they were preserving something inside that was like a living process. Yeah. That means it was as close to living as we can get because they're still in close proximity.
As an organism dies, those things naturally, because they no longer need to maintain the life inside, they will part away from each other. Yes, exactly. And so this is like a tangible way for us to be like, this is... Recently living or currently living versus this is for sure dead. Exactly. Exactly. Right? So that's one way of doing this. Okay. The other way to do this is to actually look at the carbon isotopes in your life. Right? There's two stable isotopes. There's carbon-13 and carbon-12.
Carbon-12 is way more abundant than carbon-13, but both are sort of mixed in together. There's a baseline... Like there's a baseline abundance of them, but life is going to have way more carbon 12 than carbon 13 because it's lighter. And so whatever enzymatic action is happening, the enzymes are lazy and they're going to use the carbon 12 more than the carbon 13 because the carbon 13 is harder to move around. Right. It takes more energy. It takes more energy, right.
And so for whatever little bit of energy you have, you're going to usually, just by probability, fix the carbon-12 into your organic chemistry more than your carbon-13. So by looking at this abundance, we can also tell whether what we're looking at is alive now or has died and is just sort of part of the thing. Yes. Right? And from this, they found that there's... One thing that was really cool is when they look at the mud line, right? They have this two meters, two or three meters of the ocean floor.
Yes. At the very top that's close to the ocean floor... They don't see a lot of these biomarkers. Okay. So a lot of the stuff is dead. But underneath, they're seeing living organisms. Right? So there's a living tissue that's happening underneath. There's a layer of dead stuff. Yeah. And then there's active biology. Underneath it. Underneath it. Right? And by doing a bunch of different... oxidation measurements and like HPLCs and all this other kind of stuff, they came up with a really complicated ecosystem where you can actually like find like, what were the things that are making this happen?
And they have a hypothesis where the methane comes from the bottom and the CO2 comes from the top and the water comes from the top. And they have all of these different mechanisms that they've actually traced through to create a magnificent ecosystem. Right? There's a circle of life just for this. Just in there. Just in the same area. Isolated from the sun. There's no sun. Right. Right. Which is important. This is important. It's isolated from the sun. It's completely driven by chemical processes that are happening because of the interaction between water and the Earth's crust.
Right? Mm-hmm. And it's completely in isolation from the rest of life on Earth. Mm-hmm. I think that's a foundational claim. Because, and again, I think just to kind of parse the implications a little bit here, part of the point is... For a layman like myself, a lot of times we'll think of sunlight, solar energy, as being a required prerequisite for all life.
And when the sun goes away, you can't have the light life because that's the only source. This is why photosynthesis is so foundational and all this other stuff. What this is saying is independent of sunlight, and this goes back to your point earlier about why it's different than the heat vents. Yeah. Or heat. Yeah, yeah, yeah. Or heat or the sulfur, the active stuff moving out. This is just like slow chemistry.
Happening. In the crust. And out of that slow chemistry in the crust, you still get life. You still get life. They still find a way. Right. As Jeff Goldblum was saying, life finds a way no matter where. And it's incredible how they find a way. Right. There's all these adaptations that they've come up with. For example, in there, in that high pH, those lipid molecules are just going to become like soap. Right? So how do you prevent the soapification of my cell membrane?
Well, instead of the ester bonds, they're using ether bonds. So instead of a carbonyl bond, where the carbon is attached to an oxygen, and then two groups, you've got an oxygen that's mediating that. And that oxygen becomes the anchor for your phospholipids. Not phospholipids, sorry. Glycolipids, because you don't have a lot of phosphate. Right, right, right, right, right. You're using actually like sugar and a bunch of lipids to actually create your cell membrane. The other thing that's kind of crazy is that, you know, at that temperature... Fat freezes, right?
Butter is a solid in the fridge. But that's because it's a saturated lipid. If you use unsaturated lipids like oil, that's not going to freeze. And the reason why it's unsaturated is because you've got these kinks in your carbon skeleton. You've got a double bond instead of all single bonds. And a double bond means you're going to have a kink in your carbon skeleton, which means it's not going to stack up. as nicely as saturated lipids will, right? And if you're saturated, then everything's going to stack up and you're going to become solid.
Yes. But without the saturation, then you're going to remain liquid and you can have that malleability that you need for life. Yeah, yeah. It's so funny that it's literally that double bond kink of, let me actually pull that back up. That was photo 12. That double bond kink of carbon in the middle. Yeah. Literally just that. Just that. Totally changed the entirety of the expanded structure that you scale up to. It's kind of crazy, right?
It's crazy. Like life figures out ways to make it work. Yeah. you know yeah using fundamental it's just fundamental physics right the double bond creates a kind of weird quantum mechanical effect here where like now my bond can't be linear right and which means the kid at this temperature can't just then solidify into a block it has to remain in that liquid it's so cool yeah i like that it's so cool that like life can find this out and this has huge astrobiological implications obviously and for me that's the coolest part um yeah
The cool biosphere analog, right? Like the Marianas Trench is this first high-fidelity analog of cool serpentinization in the cold. In the cold. Okay? Not in the hot. Right. Not in these zones where the plates are drifting apart. And then the heat's coming out and sulfur and all these nutrients. This is cold. That's actually really important. So all of the ice moons become much more interesting because of the existence of proving that cool serpentization can have a biosphere.
Exactly, exactly. And Enceladus is a big one. Cassini actually sampled plumes. This is a moon of Saturn. And Enceladus is something where we know there's an oceanic... ocean. Yes. There's an ocean there. There's a crust on top. And there's so much geologic activity that there's actually geysers that shoot out plumes of stuff into the cosmos. And Cassini, we were lucky enough, the Cassini got close enough that it actually sampled those plumes.
As it was flying by. As it was flying by. And what we measured was hydrogen gas. Yeah. And methane. Which is exactly what we have here. Meaning, and so like, I know I'm extrapolating. This is an extrapolation. However, given what we just walked through about what we've seen at these two plates. Yeah. In the Marriott's Trench on Earth, where we were able to send down our little torpedo, pull up meters worth of...
The top layer was dead, but below that, the same chemistry that exists in that region here on Earth mimics the chemistry that Cassini pulled up. Yeah. And there's a direct match. There's a direct match. That's not to say that there is life on Enceladus. Yeah. However, the chemistry. The chemistry is very same-same. It's very similar. Yeah. Yeah.
A pizza in New York is not the same thing as a pizza in California. Yeah. I get it. Yeah. But it's still a pizza. But there's possibility there. Which is so exciting. Which is so exciting. That's so exciting. The other thing with Europa, right? Models predict that serpentinization is actually more probable than, let's say, deep sea vents on Europa. Which makes sense. Right. And now we've got Europa Clipper going there. It's going to do a bunch of things. If we're going to send, you know, if we get our funding back. And JPL gets its funding back. We need our funding back.
Yeah. But if we do, then we're going to send, you know, let's say something that drills into the Europa Ocean and then tries to look for it. Well, those things should probably have a mass spectrometer, right? Because the mass spectrometer is what was used to actually characterize these lipids and these lipid biomarkers. So maybe let's not look for DNA. Maybe let's look for lipids. This introduces a whole new paradigm for what to even look for when we go to these distant worlds looking for extraterrestrial life.
That's a brilliant insight in that what we look for determines the results. Yeah. It's the pushback that people have given about SETI looking for radio signals. And it's like any sufficient technological da-da-da, like would they be still using it, yada-yada. I mean in an analog – in an analogy – What we're saying is historically maybe we'd always be looking for DNA as the sign of life. Looking for lipids, the presence of lipids is another signature. Yes. That is equally relevant and viable in that search and might be more fruitful than...
Because even on Earth, we can see in regions that we previously believed to be uninhabitable, life is there. But the way we found it was not through the DNA, it was through the discovery of the lipid presence. Yeah, exactly. And like, who knows, maybe they have a different form of genetic material, right? DNA is like this arbitrary thing that perhaps our ancestor four billion years ago came up with, right, to encode DNA. data right but there's there's got to be other ways right it's just whoever came up with it first was like this is dope right and then just like ran with it right yeah right uh we are living with the uh choices that the internet
pioneers of the early 90s decided how a browser was going to work and how graphical user interface is going to work and that's just what it is yeah and now we have to deal with that that is a very interesting story that was very cool huge implications not only for again expanding our idea of what does life mean just here on earth but then informing how we ask questions about the universe around us in that same journey. Fascinating, fascinating paper. We have that one again from university of Bremen.
That was in nature communications, earth and environment. This, this week of stories is fantastic. We're going to move on to our third story. Our last story of the day, our Princeton story. Yes. We've, we've, we've kept a couple of weeks of not shilling Princeton, but, But when you're the best, things happen. I don't know what to tell you. It's a breakthrough. And so we cover breakthroughs. The millisecond transmon qubit breakthrough out of Princeton. This was in Nature. No subjournal. Yeah.
Nature mainline. Princeton's new quantum chip makes a major step toward quantum advantage. Now, in a previous episode, you explained to me the distinction between quantum quantum advantage and quantum supremacy. Yes. In the context of Google's paper claiming that they've achieved quantum supremacy, which received a significant amount of pushback. But this is about Princeton's quantum computing for a new qubit, the qubit being like the fundamental unit, bit in a classical computer, qubit in a quantum computer.
So what is it that we... Literally we as in you and I personally. Yeah, yeah, yeah. Hey, I donate. I donate. I donate every year. The money goes to this year. They called me up like last month being like, hey, are you going to – the class of 2014. I was like, yeah, fine, fine, fine, fine, fine here. As long as I get to claim that I was the one who came up with this quantum breakthrough. So talk – to help me understand like what – because we talk about quantum quite a bit actually. We do. So this is – it has to be actually a big deal. Yes. For you to put it in.
It is a big deal, yeah. Okay. So quantum computers, right, the promises to break encryption, simulate materials, you get a better portfolio for your financial stocks because it's going to figure out the best way to do things. And all of that relies on actually making a quantum computer, right? And making a quantum computer as of now, it's still a hardware problem. Okay. Okay. Okay. It is still a hardware problem. A quantum computer, let me just go through briefly what a quantum computer is. Okay. It uses a qubit instead of a classical bit.
So a classical bit is your transistor that is in a state of either I'm going to let current through or I'm going to block current. So that's my zero and one. A qubit can be in a superposition of two states. It can be in a zero and a one, and it can be in a linear combination of something in between. The other thing that's different about a qubit versus a classical bit is a classical bit, one transistor doesn't really talk to another until it's really told to, right? It's like maybe one transistor flips its sign based on the state of another transistor, and that's an explicit...
sort of thing that we're doing within the algorithm of whatever we're implementing. But in a qubit, the power of the qubits is that they can be entangled to one another. They are not independent, right? So one qubit can be in a superposition of zeros and ones. Another qubit right next to it can be in a superposition of zeros and ones, and so on and so forth. And I can get access to an exponentially large number of states that in my memory without actually having to explicitly encode each one of them.
Right. As a terrible analogy, because I'm always trying to figure out ways to think about it in my head. If you have a relay race, a four by 400 classical computers, you have to wait to hand the baton to the next runner. Yeah. But in a quantum computer, my first lap, the first runner can have like multiple, uh, runs and like whichever one happens to get there first is like and so it's like you're all four just basically cheat right right go through the whole thing yeah but just as trying to like contextualize like what the difference is when you say that a qubit can be in a superposition of both zero and one there's not you're not waiting for the handoff in order for the next thing to do its next process
Yeah, yeah. You can do a lot of things in parallel. In parallel. And you can access a lot of computational stuff in parallel. In parallel. Yeah, yeah. It's a weak analogy, but it'll work, actually. And at the end of the day, how a quantum algorithm works is the following. There's something that's very famous called the Shores algorithm, okay? This is the one that all the NSA and everyone is excited about because this is the algorithm that lets you break RSA encryption. Got it. But it works like any other quantum algorithm in that here's what happens.
You've got a bunch of qubits that you're going to prepare in some kind of state. Let's just say for argument's sake, I'm going to prepare them in the zero state. So they're all zero. And then what I'm going to do is I'm going to apply quantum gates. To these guys. Okay. And what that means is I'm going to poke this qubit a one way. So it's going to maybe flip from a zero to a one, or it's going to flip into halfway between zero and a one. And then this guy's going to talk to its neighbor. Who's going to talk to his neighbor. And all of these quantum gates are basically pokes in some way of each of the bits that I have.
Mm-hmm. The key bottleneck is the following, okay? My gates and my poking have to be faster than how long it takes for my qubit to forget itself. Here's what I mean by that. In a transistor, if I turn it off... going to stay off. Unless like a cosmic ray comes in and like you know, some proton comes in and knocks exactly that transistor from a 0 to a 1 it's going to remain off. And I can reliably come back and it'll remain off.
This is how data works, right? Why is my data... How do I store stuff? This is the transistor or like whatever thing is in the 0 and the 1 is going to be in the 0 and 1 forever. It's not a transistor, it's solid state or whatever. But At the end of the day, I have some time horizon that is way longer than my computation. Yes. that I know can reliably remember if it's a zero or a one. In quantum mechanics, that's no longer the case because these qubits are made out of incredibly small technologies.
And so there's a lot of forgetting that happens. I prepare my qubit in a zero state. For example, let's take the extreme case of an electron that's spinning in one direction, and that's my zero. If it spins in the other direction, that's my one. Well, I need this thing to remember that it's spinning in that direction. But if there's a bunch of stuff around it, it's going to start poking it, right? There's a bunch of atoms and a bunch of little magnetic fields that the atoms create that are going to try to change the way that my electron is spinning.
So if my electron is a qubit where spinning down is zero and spinning up is one, and I prepare it in the down direction, It's not going to take very long for that electron to just get influenced by all of the stuff around it and forget that it was a zero in the first place. This is called the decoherence time. And one of the central challenges in creating a workable quantum computer is to make the decoherence time way, way longer than the...
algorithm time so so this this actually makes total sense the the point is you don't actually have like a workspace that is active for long enough for you to do like yeah like valuable computation yeah to get an output because the decoherence time is so short similar to our first story because we're at a scale now where the little stuff Changes in very, very, very small things that might not have mattered in a classical system matter a lot in this quantum system in the same way that like under five Mach, you don't have to worry about the little stuff.
But above five Mach, you have to worry about stuff in a different way. Exactly. Or the details matter. And so – If you can increase the decoherence time, that means I can run a long-running algorithm. It's like a context window in AI. Yeah, exactly. You have more time to do the thing you want to do before... Yeah, before it forgets, right? Yeah, yeah. And there's ways to do this to get around it. There's stuff called error correction, where what you do is you, like, replicate the qubit multiple times.
And then you have, like, a bunch of qubits retain that information. So if one of them forgets, you can have, like, the other. But, like, at the end of the day, a hardware solution is... is king it's not it's not clean it's not clean right right you have to do all of these contingencies if you have a hardware solution right well then you can implement that software solution with the hardware solution exactly and even multiply correct beyond yes beyond like what you were so the hardware is always king the king the biggest fundamental limit is still at the hardware is Still at the hardware level, exactly.
And so there's different kinds of qubits. I just gave an illusion of the electron being a single qubit, right, where it's like spinning one way or the other way. The one that we're talking about is actually pioneered by the 2025 Nobel Prize in Physics by these three gentlemen, John Clark, Michel Devereux, and John Martinez back in UC Berkeley. They came up with this idea of macroscopic quantum tunneling, right? The idea was to make a quantum circuit, a tiny little circuit, that had something called a Josephson junction.
Yes. And then you have on the order of billions of electrons that behave like a single quantum unit that are tunneling from one to the other, from one to the other, right? We had a great deep dive on this during the Nobel Prizes, so it is a phenomenally well-done deep dive. Yeah. So please go check out our Nobel Prize episode on this, because if you're curious about this concept of the superconducting circuits with Josephson junctions and macroscopic quantum tunneling, it is really fascinating.
And we have, I think, one of the best explainer videos out that's available. We won an award from TikTok. TikTok did give us a second place for our science content because it was so good. And I just was so fascinated by it. Yeah, yeah. Please watch that video. Exactly. So these guys sort of went and created that thing. Right. And IBM and Google now use something called the Transmon design, which is a charge qubit. It's basically their design, but they've attached a large capacitor on the side.
Okay. And they've made it insensitive to charge noise. But this is basically the qubit that they're using to make the... the quantum computers that they talk about. So currently there's not a lot of quantum computers that are being built. There's very few amount of entities that are doing. No, the ones that are like you're saying with, with IBM and they're using this particular structure for the, in order to be able to like, do the process that we're talking about, which has existed prior to the paper.
Yes, exactly. Like there, there's a platform that everyone's using saying, this is kind of the way we got to go. Everybody, this is the current state of the art. Yeah. And that's what it is. Yeah. Yeah, exactly. But those guys, I, had this problem of decoherence time. This new paper that's come out in Nature, millisecond time skills. They're getting their stuff all the way to milliseconds. And this is a victory against the materials that we're using to actually create that transmon qubit. And this millisecond time skill is about three times...
the timescale that was there before the record. Got it. And it's 15 times what is the industry standard. Okay. Got it. Okay. Yes. It's very long. It is an order of magnitude. Order of magnitude. Better improvement upscale. Got it. Exactly. Yeah. Okay. And what they're doing is they've, The problem is, as I was saying, right, there's these tiny defects that happen everywhere else that sort of make my qubit forget what I am. If it's a zero, it'll just go halfway.
If it's a one, it'll go halfway. There's all these like glassy materials. There's oxides. There's substrates that happen when I construct my chip. that is going to make the decoherence happen. And just to kind of zoom in on the point, the point is in the process of generating these chips, we have to make things at these incredibly small scales. Our engineering capabilities are quite good, but there are still levels of imperfection or just that the scales are so small that things happen in ways... Yeah, this is at the tens of atoms, hundreds of atoms...
So the slightest little thing is going to, again, take our gate of our decoherence time and shrink it with any amount of 10 to the negative big number imperfection, which is like barely anything. Yeah. Am I understanding that correctly? But to the quantum world, it matters. It's everything. Okay. You know? Yes. Yes. So that is one of the core sources of the decoherence time problem. Yeah. It's all of these little imperfections that are coming in.
And so there's been a... history of trying to make things better. Right. Okay. When it comes to getting longer and longer decoherence times, the first platform was to try and use niobium and aluminum. Okay. The sapphire, actually sapphire is just aluminum oxide. Okay. And that was being used instead of niobium in 2021, there was a breakthrough the same group. Actually, instead of niobium, they use tantalum on the sapphire aluminum oxide. Okay. And there they got up to about 300,
Okay. Okay? Yeah. Sorry, 300.3 milliseconds. So 300 microseconds. Microseconds. Right? Yes. And then finally now they've challenged this bottleneck. And instead of using Sapphire, they're using high resistance silicon. Okay. So extremely pure silicon that is high resistance. Yes. And what's happening is you're superconducting stuff that's happening in the qubit. Yeah. It's no longer dissipating energy into the substrate itself because the substrate has such high resistance.
One way to think about it is the energy just doesn't want to go there, right? And so it's staying within the qubit, the part of the circuit that actually matters. They've locked down their border. Yes, yes, in some sense, exactly. That's exactly right. They've eliminated this bulk loss, and using tantalum, which is something that— Not a lot of people use, but these guys were using it. Tantalum turns out to be much better than niobium and aluminum. So it's this combination of tantalum and silicon.
And that was super non-trivial for them to actually grow. Because you've got silicon and you've got tantalum on top. There's all this chemistry that's happening that's kind of weird. They had to use a really ultra-high vacuum before they were using just high vacuum. And it wasn't enough. It wasn't enough. Yeah, yeah. So they had to use ultra high vacuum so that when they were depositing this stuff, no weird oil was getting in, nothing like that. When it interacts with the air, nothing weird is happening.
It's been a process that's been happening for quite some time. But at the end of the day, they beat that millisecond barrier. And they got to the key. The point is, like, they had to cook up. like like breaking bad yeah like heisman they had to cook up the the material yeah that actually went into the the the thing itself the chip itself that is being now used to do this process yeah exactly it's like completely novel it's not like you can just go like hey nvidia like hey you know tsmc no and then they were doing it they had to which like i i want to like make that yeah like that's
big challenge. It's multiple groups in the university. The chemistry department was involved. The electrical engineering department was involved. That's so cool. There's all these different players that were like, oh, actually, I'm really good at tantalum. Yeah, right. And I'm pretty sure you need tantalum. Right. And then somebody else was like, put this on the silicon and you're going to need a higher vacuum. You're going to need a higher vacuum. Someone was like, you're going to need more cowbells. Yeah, yeah, yeah. And there's all of these different factors that are coming in, in this ecosystem at Princeton to create this qubit.
And this qubit is amazing. It's got... The best one, the best one had 1.7 milliseconds. That's crazy. And to tell you just how big this is, you know, before you were getting to 0.3. Yes. The best one is now five times. Right, right. And that 0.3 was in a lab setting. It's not at the industrial setting. Okay. So you're getting almost 20 times to 25 times the industrial setting of stuff. Right? And you might be thinking, you know, a millisecond, that's a thousandth of a second.
A thousandth of a second is not very long. Well, the gates, right, the stuff that we're using for quantum stuff is on the order of tens of nanoseconds. So you can fit in a lot more stuff in one millisecond now. A nanosecond is a hundred thousandth of a millisecond. A millionth, sorry, yeah. It's also because it's like quant, like when you're doing it's parallelized, right?
So it's, it's the volume of what you're actually doing. It's not like, it's not like every millisecond that goes by because you're now parallelized in what you're actually doing. Like the actual amount of work, the way we think about it in a classical sense. I don't know if I'm explaining that correctly. You're exactly right. Like the, the, the, How do we say it? The amount of stuff you can do with this.
The amount of token output is not just like the same as you would get in a millisecond on a classical system. No, no, not at all. The token output is like running ChatGPT parallelized N number of times. And all of those token outputs get you to your ideal whatever output. I'm just trying to talk about how that millisecond difference is extremely meaningful. Exactly. And the other thing is, you know, getting, let's say, 20 times longer in this decoherence time adds up exponentially.
Because here, this is a physical error, right? This is something that's happening at the hardware level. So all of your error correction algorithms can now take advantage of the fact that the thing that they are doing the error correction on is itself not that bad. Yeah, it's less error prone. Yeah, right, right. It has higher efficacy. So you're adding on exponential on exponential. Yes. Right? And it's going to be a really big thing, I think. The other huge thing about this is this chip, which is it's...
It's just an upgrade on the industry-dominant Transmon architecture, which is the same thing that IBM and Google have been using in their current architecture. Right. The recent paper that came out of the Willow quantum chip in Google, it's using the same architecture. So the thing is, this is something like a, I can take that out. Yeah, yeah. And I can put this thing in. Yeah. Right? All of the – outside of the chip itself, there's all of the other – the cooling system, like all of the other infrastructure.
Yeah. The way I talk to the chip. All of that. Yeah. They can hot swap the chip. Exactly. It's not a new – it's not a totally new chip design that requires a refactoring of the rest of the infrastructure. Exactly. That's really – That's huge actually. That's a big deal. Because now this is scalable. Yeah. It's industry compatible. And so if Willow just gets this design, it's going to be a thousand times better. And Willow already is making headlines. So imagine what it can do with this, right?
It's really bringing us closer to that future, possibly, of where something scientifically... standard could be happening with quantum computers, you know, within the next decade, you know? And I mean, there's, again, everyone always talks about all the implications of quantum computers. The one that still gets me most excited is like the ability to create simulations and models with higher fidelity. Yeah. Coming back to our first story. New materials. Yeah. Right. Right. And like the things you can then...
Yeah. Instead of trying a thousand different materials to see which one is the room temperature superconductor, I can just simulate a bunch and then be like, oh, these are 10 good candidates. Let me try making those. Because alpha fold was a lower complexity problem to do all the folds that proteins can do. Yeah. I mean, it's still an incredible. Not minimizing. Not minimizing. But we've already crossed that bridge. So, you know what I'm saying? You already got the Nobel. That's in the past. We're trying to go orders of magnitude.
Yeah, yeah, yeah. Now we're doing quantum mechanics, motherfucker. Which is like, it's a big, that's a big. All of these stories are actually pretty crazy. Yeah, so I think this is huge. The other thing is they're using silicon, right, instead of sapphire? Yes. So instead of aluminum oxide, they're using silicon. Yes. Well, we're, as humans, incredible at making silicon. We have the whole industrial infrastructure for that is actually prime. Yeah, so scalability in terms of that is going to be huge. That makes sense. That makes sense. Yeah, we can piggyback on this trillion-dollar semiconductor fabrication process.
This is crazy. Right? Yeah. Yeah, and that's a photo of one of the grad students probably posing with it. Yeah, they're like, we don't need your face. We just need your beautiful hands. Yeah, yeah, yeah. You're a hand model. Yeah, put the gloves on, though. Yeah, this is expensive. Yeah, exactly. So, yeah, very, very, very cool. I think it primes the new chips for industrial scaling. We could be getting ever closer to that dream. So now I can read the title, The Millisecond...
3 to 5x longer amount of time you can actually run and keep things in memory, for lack of a better term. Transmon, the platform that is already industry standard that's used, and so it's not net new and everyone has to change everything. Qubit breakthrough, the type of thing we do, the tantalum silicon Qubit out of Princeton in nature. We're still number one. I don't care what nobody says. Yeah. Actually, we are, though.
There was a thing that it took the U.S. news rankings and all these other rankings and then averaged it out. Oh, yeah. And we were number one. Yeah. Look, we talked about our leaderboard that's coming in our last episode. That leaderboard, we're going to create our own metric. Yeah. Make sure it prints in the snow. You know, it'll be – the methodology will be public. Yeah, yeah, yeah. It won't be a black box like all these other AI models. And the asterisk will be printing in all this stuff, bro. And it will just rank based on the facts, you know, as we like them.
Yeah. Three super, super six stories this week. We talked about hypersonic aerodynamics above Mach 5. Very interesting. I had Stevens Institute of Technology in New Jersey. That was in Nature Communications. We followed that up with some microbial life in the deep sea, not near hyperthermal vents or anything. This is in a different – This is cold. This is cold. Yeah. The title was Life Found in a Place Scientist Thought Impossible out of University of Bremen.
There were biomarker evidence for serpentinite chemo synthesis. Yes. Yeah. In the Mariana's forearc, nature communications, earth and environment. And we ended with an incredible quantum breakthrough, millisecond transmon qubits out of Princeton in nature. I just, I mean, the stories we've covered this year have just been, and again, it's tough because we weren't doing this in previous years. So it's like. Yeah, we don't know. We have no baseline. Yeah. I mean, I'm learning a lot doing this show.
It's pretty, it's pretty cool. It's incredible. I just hope people recognize and understand how incredible some of the stuff we're doing all the time is. Yeah. And I just, I mean, as humans, we're, we're so sick. We're sued. We're so sick. Yeah. We have problems. Okay. We get it. Yeah. We get it. Those are like individuals and like groups that identify with problematic stuff. Yeah. But when we, when it comes to like science and like resilience and like, just like our ability to like chipping away at that cave of the unknown, you know, it's just, it's just amazing.
It's incredible. Uh, I'm Lester Nare, as always, joined by my co-host and our resident PhD, Krishna Chowdhury, with another incredible episode. We will see you all next week. This is From First Principles.
Do pigeons navigate with their livers? A surprising study proposes that iron-rich immune cells help homing pigeons sense Earth's magnetic field. But other researchers argue that the mechanism does not hold up. Lester...
How can a weather satellite see Ramadan, Christmas and a blackout? Dr. Miguel Román joins Krishna Choudhary to explain NASA's Black Marble: nightly maps of the lights on Earth, from a single band on the VIIRS...
Can sound destroy a tumor without an incision? Dr. Zhen Xu joins Krishna Choudhary to explain histotripsy: focused ultrasound that creates tiny bubbles to mechanically break apart tissue, without relying on heat. From a...
What happens when AI moves beyond the chatbot and into the laboratory? From Claude-designed proteins that bind scorpion toxins to robots that assemble optics experiments, we examine what AI can do for science and what...
Can light restore aspects of sight? Can recalling a cold experience change an animal's metabolism? And what can ghost particles and mirror-image molecules tell us about the universe? Lester Nare and Krishna Choudhary...