The Future Of Our Brains And Bodies - EP 77 Max Hodak Live Event

Thanks everyone for coming. Thank you, Max, for being here. This is our first ever live in-person podcast event and huge thanks to all the subscribers and the other people who came. I still appreciate you, maybe slightly less than the subscribers, but still very thankful. Thank you to Brex for hosting us and obviously being such a big sponsor of our show. And thanks to all the people that helped pull this off, including the Core Memory crew back over there. And yeah. Our guest tonight, I think based on the registration list that I saw, lots of you will be familiar with Max Hodak. He's the co-founder and CEO of Science Corporation, which is a fantastic name for a company, and previously was president and co-founder of Neuralink. That's kind of where we got to know each other. Thank you so much for coming.

Thank you for having me.

You're our first repeat guest.

Cool.

Yeah, it's cool for me. I always enjoy talking to you. I think the first time we met, we— I think I was your first official interview back in the Neuralink days. You guys had never really talked.

This was so long ago. I mean, yeah, very— I mean, you'd obviously worked closely with some other people there.

Yeah, I don't think you guys had done a story. I just remember you telling me how much you hated reporters. It was tense for a moment. Now here we are, all these years later, hanging out.

I'm just bad at text.

And then the other thing I remember was that you're like, we don't really want to get into the primate trials yet. Let's, you know, we'll get to that eventually. And then there was an event that night, the Neuralink had— and Elon got upstage, was like, we're putting implants in monkeys. And I remember you grimacing in the back. Yeah. No. Okay. I was going to take a crack at explaining science, and then I was going to have you tell me sort of how you picture this company. But is it just in the couple days I was refreshing myself on everything you guys are doing. You were quiet for a couple of years, and then over the last year, it's just been, I feel like, one revelation after another. For people who don't know, Science has a— it's like a product, a category called Prima, where you guys are helping restore sight in people who've had degenerative conditions through a sensor that goes in the eye and some glasses. You have this Super fascinating technology. It falls into this category called biohybrid, where you're putting neurons in a substrate, a hole in the skull, and then the neurons have synapses that go down and fuse into your brain to give you extra brain power or to regenerate things. You have a pretty new product line called Vessel, which is around keeping organs alive outside of the body. And then you have a whole set of products, software and hardware, that helps other people try to explore brain-computer interfaces and associated technology. And so I kind of like watched each one of these things come out. And then when I was sitting back this morning, I was texting you. I mean, I guess— I don't know why. I just never pieced everything together before. It just strikes me. It seems like you guys are the most ambitious or have the most products aiming for a general-purpose augmentation of humans, like a new era of attaching, fixing human beings. Is that how you think about science?

We have a couple of different programs, as you mentioned, where The furthest along is our retinal prosthesis to restore vision to the blind. Then there's biohybrid as a general-purpose brain-computer interface technology, which is a little bit earlier, and then our work in perfusion. I mean, our goal is to build a big, like, an independent medical technology company. There's— we don't— at the heart of this, there's really one program. I think this is sometimes seen as like a basket of different projects. And it's like they're just kind of duct taped together. And why is this one company? Why is this not 3 companies? Are there other things that we're doing? Why are we bothering to sell hardware or software to others, to other researchers at all? But really at the heart of it, there's the really one idea. So the kind of the mission of lowercase s science is to use a differentiated knowledge of the universe to improve the human condition. And that's really how I think about our mission at uppercase S science. And the, like, the video we just watched of looking at these stars, I don't think people realize that if you have enough, if you have enough, like a sufficient propulsion, you could visit everywhere in the known universe within about 60 years, like within one human lifetime. All of these places are theoretically accessible, but I don't know that the things that evolved for Earth are going to match the that. And so at some point you have to go further. But really right now our focus, we're not working on products imminently that are designed for augmentation. Our goal is to restore function to people that have lost it. And when I think about our roadmap, we're not really thinking about healthy 30-year-olds or healthy 40-year-olds getting our products. Our perspective is that many people, possibly almost everybody eventually becomes a patient if these The human condition is complicated, and then sometimes it's great and sometimes it's terrible, and kind of as you get older, it gets trickier. And the humanity, I think our worldview is really oriented around this idea that humanity just isn't very good at drug discovery. It's still really mining for oil, and every now and then you find a GLP-1, or just this week, there was major progress in pancreatic cancer, and every now and then you find these things, Very commonly in drug discovery, you'll work on a project for a decade and then at the end you'll get a readout and the answer is no. And then it's like, where do you go from this? You can't get an engineering gradient on that problem. And on the contrary, neural engineering, if we work with the brain directly as a computer, this works really well. If you put a cortical motor decoder in a paralyzed patient or an ALS patient, an hour later they're playing video games. Or if you turn on a deep brain stimulator, in 10 seconds a patient goes from not being able to hold a cup of water to being able to write cursive. Or with our retinal prosthesis, which stimulates the cells just past the rods and cones, I mean, there's been tens of billions of dollars and many decades of work trying to find drugs to stop the degeneration of the retina, or much less restore vision using drugs to really no avail. There's been very, very minimal progress. And with our electrical stimulator that again, just treats the brain basically as a computer, we can take patients from being unable to recognize faces for over a decade to one of our patients in Europe just finished a 300-page novel. And so humanity is just, like the brain is just more amenable to this type of engineering and it is simultaneously the central object of medicine. Like when you get a vaccine or you check your cholesterol, at the end of the day, the thing that you are really, trying to preserve is that the brain activity continues and is interesting. Like, you can get a heart transplant or a liver transplant, but you cannot even in concept get a brain transplant. And when you deal with the brain directly, you get these effect sizes that you just don't really see in medicine. And so at Science, the company, we have a broad mandate to develop devices kind of in that worldview that restore significant functionality to patients that are in trouble, whether that's because they've So when you think about our perfusion platform, initially we're looking at applications around transplantation and there's some really cool research there, but it's really a platform that you can imagine using. This takes you on a path towards things like total artificial hearts or improving dialysis to something that a patient could live with for decades. The BioHarbored is similarly a platform. Initially we're looking at things like stroke, but there's a range of— there's some really cool research going on where that unifies really deeply with what's coming out of the artificial intelligence community, where I think we're rapidly gaining this really cool knowledge about how do brains work and how to engineer them to restore lost functionality that we just could not have talked about 5 years ago.

Okay, okay. Follow-up questions. I forgot there will be a Q&A portion of this in about Like 30 minutes. I just wanted to time check myself. So just anyway, start forming your questions. Okay, so kind of along the lines of what you're talking about, I spent the last few years covering you guys, covering Neuralink, covering Onward, which makes a spinal implant to help people move again, all of these companies. And I would have the similar experience every time, just about, which is I go to Switzerland, I see someone with a spinal implant walk. You know, they were completely paralyzed for 12 years and now they're taking steps across this room and you're like, this is miraculous. And then it sinks in for a second. It's like, oh, they're not walking at all like they used to. It's fantastic that they can exercise, stand up. It changes their lives in some ways, but then it also It doesn't at all. And when I talk to people in these fields, sometimes it's really unclear to me if it's just early days with this technology and it, like so many things we're used to around Moore's Law, it's going to get better and better and better. Or if this is a different category and it's just we're quite gated. And I didn't know if Primo was a good place to talk about in relation to this about what's possible with the implant today and then where you think it's going, or it seemed like you had broad thoughts already.

Yeah, I mean, it's— I would say broadly, these fields are not like Moore's Law. And although Prima might actually be an exception to this, that really is a thing where the electrodes get smaller, the implant gets bigger, there's— you can cover more of the atrophy. That's an area where really we can't think of it more that way. But I'd say broadly, a lot of this stuff, like, you're right, like, there's the spinal cord injury patients that have BCIs right now. Like, they're not walking, they're not running, they're not, like, swimming in pools. This seems very incremental, but that's the way that a lot of this progress in medicine is. It's not these big flashy— it's not like suddenly there's like the problem is solved. Over the last 20 years, for many cancers, survival has doubled, but each of the results looked very incremental. But it's that kind of compounding incremental progress where each time you learn a little bit more. And to some degree, there's a divergence between what's possible in the animal models and what you can do in patients just because the feedback cycle of human subjects research is so slow. If you have to freeze your design for multiple years to put something in a patient, then it really limits the rate at which you can learn things separate from what's possible with the technology. When you— BCI aside, when I look at broader medicine, one of the areas that I'm most excited about these days is n-of-1. So this is like the conventional approach to drug discovery is you're looking for a thing that you can prescribe to many patients and for like some broad indication, the more patients the better, that makes it more venture fundable. But then you're averaging over the natural variability and complexity of these diseases, and biology just often doesn't really work that way. But if you have one patient and you have the full toolbox of modern science and you can think about them, you can reason about that one case as a really deterministic biochemical system, what's possible looks totally different than what people think is possible in medicine. And so there's to some degree a divergence between what's possible scientifically and technologically and what is entering practice because of that translational gap. And that I don't know that there's really easy answers. I think for us, one of our thoughts on this is that I think medicine is often too concerned with statistical significance and not concerned enough with effect size. If your effect size is small, then it will be tough to translate, and then you'll get into arguments with payers, and then patients will be thinking, do I really want the surgery? But again, if you can take a patient who can't recognize faces and allow them to read novels, it's like everything is downhill from there. It can still be slow, but those results are still very meaningful, and you can really see how they compound to places that people couldn't imagine.

Well, I mean, can we focus on Prima for a second just to go through that? So, okay, so you've got a chip today that—

Yeah, so let me give a little bit of background on Prima. So we're looking at diseases where the— let me back up a second. So in the eye, the eye is, you can think of it almost like a camera. It doesn't It's not quite a camera in the way that the brain, like, works with it, but light shines in through an optic that forms an image on the back of the eye. In the back of the eye, in the retina, there's really 3 layers of cells that, I mean, it's more complicated than this, but you can think of it as having 3 layers of cells. The image hits the rods and cones. These are the kind of conventional light-sensitive cells most, many people have heard of. The rods and cones turn the light into, they release, chemicals that then connect to a layer of cells called bipolar cells, bipolar because they've got two ends, and those connect to the retinal ganglion cells, which are the optic nerve. So if the rods and cones have died, the optic nerve might still be there and the bipolar cells might still be there, but the eye is no longer light sensitive. There's a bunch of different families of diseases that cause this macular degeneration. The most common form is age-related macular degeneration. It's one of the most common forms of blindness, retinitis pigmentosa, diabetic retinopathy, Stargardt's disease. There's a bunch of ways that the rods and cones can die. And historically, people have been trying to understand molecularly, like, what's going wrong in these cells and how can we develop a drug that will stop whatever's going wrong in the rods and cones from happening, or even kind of there's like maybe they haven't totally died, but they've gone dormant, maybe we can reactivate them. And those drugs broadly haven't had big effects. There's a gene therapy that— so in some patients, there's a mutation in a gene called RPE65. When that goes wrong, the rods and cones start dying off. There's a gene therapy a while ago that would fix RPE65. This was acquired for $5 billion by a big drug company. Now that, I mean, it very marginally slows the rate of degeneration. It has some impact, a small impact for some patients, but definitely is not restoring the lost sight. In macular degeneration, there's drugs called complement inhibitors that will slow the rate of degeneration by 15 to 20%, but you're still going blind, and it's definitely not restoring any vision. The neural engineering view of the world is, I don't really care why the rods and cones have died. I don't need to understand molecularly what exactly is going wrong, and I don't need to care the difference between retinitis pigmentosa and macular degeneration to some degree. What I need to do is get a signal back into this cable. And if I can do it in the bipolar cells, then that's probably like that. I mean, it was an empirical discovery of our clinical trial that if you activate the bipolar cells, you'll get an image in the mind's eye. They'll see things. There had been lots of research before trying to stimulate the optic nerve directly where they didn't see things. They just got these diffuse flashes of light., and those didn't really add up to vision. The brain doesn't really kind of bind them together into a coherent image. You'll see these— we call these phosphenes. You'll see these artistic representations of maybe the outline of a face with these flashes of light, but you seeing that through your photoreceptors is very different than the patient seeing that stimulated into the optic nerve, where you've gone from about 100 million photoreceptors to like 1 million optic nerve cells. And so you've done this 100-to-1 compression, And when you stimulate that like 100 million channel signal into the 1 million optic nerve cells, you just don't get an image, you kind of get trash. And the bipolar cells are somewhere in between. And so there is a discovery, we have figured out how to stimulate the bipolar cells in order to get that coherent form vision image in the mind's eye. Now Prima in its current form, so it's a 2mm by 2mm chip that has electrodes that are 0.1mm in diameter., and that is able to get that image, but it's small. It's like looking through a straw. And it's implanted under the retina. Yeah. And so this is a— it's a 1-hour outpatient procedure. Come in through the front of the eye, leave it under the retina. This type of surgery is done all the time. It takes a little while to get used to looking at the videos of this, but you can adapt to it. It can be done under local anesthesia. This is not the final form of this product. We need to make the electrodes smaller. We need to make the implant bigger, cover more of the area of atrophy. We need to potentially individualize it to patients. But it is, I think, a very powerful existence proof that success is possible, that when you do this, you can get a restored image in the mind's eye.

And success today looks like— couldn't read a book at all, and now I'm making out letters, I'm making out entire words. I mean, you talked about somebody who read what, like a 300-page novel?

Yeah.

Yeah.

Yeah, so I have to be— so we're not approved yet, so all I can talk about is kind of what happened in the clinical trial. I can't say that it would be like it would do anything particular for some other patient. But in the clinical trial, you're able to take patients that really had— like, if you— so the eye chart that you're used to with the letters, normally this is put 4 meters in front of you, and then you read off as much as you can. For these patients, they would put it 1 meter in front of the patient. And the patients in our trial, they could read a little bit of the first row, the very biggest letters, with their remaining peripheral vision. They're not in total darkness, but the whole central area of high-resolution color vision— and many people don't necessarily know this, but only the central 6 or 7 degrees of your visual field is high-resolution color vision. The rest is low resolution, more black and white., but your eyes are constantly darting around like several times a second. Your brain is constantly trying to figure out like where, if I glance, is that going to best update my uncertainty about the scene? And when I said the eye isn't exactly like a camera, you don't— what you perceive is not the image that falls on the retina. First of all, the rods and cones are actually in the back of the retina. It's backwards for boring evolutionary reasons, but If you were to get that, you'd see shadows of blood vessels. There'd be all this other stuff occluding it. But what you actually experience is the world model constructed by the brain. And this world model is this highly semantic model of reality. This makes it so that when you look at something, you can kind of intuitively get a sense, like, is it hard? Is it going to be cold if I touch it? Is it soft? When you look at a person, you have a mental model of them. And of course, when you're talking to somebody, you're not experiencing them directly. You're experiencing your simulation of them. And so it's about updating this world model through the cable of the optic nerve. And so the patients, they've lost that high acuity central color vision. They have some residual peripheral vision, so many of these patients can walk around. They won't necessarily walk into walls, but they've lost the ability to read. They can't recognize faces, definitely can't drive. And what you can do is restore a patch from that central vision, just that central area of— right now it's black and white— of black and white high acuity vision. And so it's not individual letters, it's a couple letters at a time typically that they can then string together into words. So they have to kind of look at it in parts, but it's an image just like you see.

And you think this gets, unlike some of the other things that we might have been talking about, you think this gets better because you just make the sensor bigger, better, cover more area, denser over time?

Yeah, well, I mean, if you think about it, the goal is you've got millions of retinal bipolar cells, and if we can selectively stimulate those, then that's all vision is to the brain. And we, And so those cells are 10 microns, maybe 20 microns. The electrode is 100 microns. Every time we stimulate, we drive dozens of them, maybe hundreds. And so that's why it's black and white right now, because color is rolled out like— there are different bipolar cells that encode different colors. And so when we stimulate these in bulk, that gets averaged together to white. Like if you think about it, there's no such thing as a white photon. Like there is no— There's no white laser. When you see white, this is a loss of resolving power. This is you're getting different— like if your vision was infinitely fast, you'd always see flashes of different spectral colors. But when we average together red, green, and blue, then you get white. And so it's seen as high contrast black and white right now. As the electrode gets smaller and we're able to stimulate fewer cells at a time, we expect to start getting color percepts.

So, okay, if we look at the BCI field more broadly, things that people would maybe be the most familiar with— Neuralink putting an implant in the motor cortex to help with movement, other people trying to— people doing similar thing for movement, Synchron, Precision Neuroscience, other companies. At various stages, maybe trying to help with speech, these different functions. Going back to what I was talking about before, I mean, it always feels like there's this limitation because you have to pick a region of the brain. You can only go so deep. Even if you're on the more extreme end like Neuralink and you're doing a surgery, you're still only going millimeters into the brain. So we're like, We've been in this the last 8 or 9 years have been sort of this gold rush of these BCI companies. This used to be an academic field. Now there's startups galore, many of them very well funded. Do you, unlike Prima where you kind of see this path ahead, do you think like all of those companies are fundamentally limited in how far they can go?

I mean, I don't see— BCI is not a product. I think people compare us to other BCI companies all the time, but I think this is And every now and then you'll talk to VCs and a fund will say something like, oh, we have a BCI bet, like we can't, we're not gonna invest in this other one. This is the wrong way to think about it. BCI is a category in the same way that pharma is a category. It's not that VCs will have a BCI bet, it's that there'll be BCI specialist funds that do lots of different things because it's really just this alternative approach to these big effects in healthcare It's different from drugs. And so in the same way, it's like, imagine if like 50 years ago, it's like, oh, like the venture firm is like, oh no, I've got a drug company. I can't fund more drug companies. There's a huge long list of indications that these things can be helpful for, but we're at the early days of it. I mean, there's some really amazing results from UCSF that deep brain stimulation can be incredibly effective for end-stage depression. And now those results come from patients that are considered like, so— and their suicide risks, so end of life that it justifies these surgeries. But as these devices get smaller and the surgeries get easier and the effects get more specific, like Parkinson's, depression, addiction, epilepsy, motor disorders— when we looked around at the beginning of— in 2021, at the beginning of Science, we asked ourselves, like, what is the most valuable unmet need that could be addressed with a BCI-type technology, it was kind of apparent to us that the answer was restore vision to the blind. And if you want to do that, there's a couple different places in the brain you can look at doing it. You can look at the retina, the retina connect, the first stop of the retina in the brain is a structure in the thalamus, or then from there you can go up to visual cortex. And for a bunch of reasons, we thought that the right place to intervene was in the retina. But there's a lot of detail in this. I mean, this is medicine. Like, this doesn't work for patients that have lost the optic nerve, which is glaucoma. But for these other, for hundreds of thousands, for millions of patients, this comes with the potential of a really significant impact.

So, okay, so in that case, you're talking about millions of people. When I look at some of the other BCI companies, and I get what you're saying, I mean, there's going to be different treatments for different things, There does seem to be a problem, at least from where I sit as a reporter, which is that a lot of these companies— this is a field that moved from academia to full-on Silicon Valley tech startup world, and they've raised tens of millions of dollars, in some cases hundreds of millions of dollars. And then when I start talking to them about who's going to use a spinal implant, who's going to use an implant that's in the motor cortex. You're talking about relatively small populations of people, whether you're paralyzed or ALS, something like that, but you've been funded almost like a consumer technology company, or you've at least been funded— really, your technology would have to become like an elective, general-purpose type thing. Do you agree with my perception, and are we on the verge of a whole bunch of these companies, it just strikes me there's not going to be enough money from just medical use to get past this cap.

I don't know. I don't— to be honest, I don't— there's— it's tough to say. Again, I think this is different than— I mean, we're not working on a motor prosthesis. One of the reasons for that is I think it'll be widely available. I think there'll be many companies that are able to produce motor prostheses, and there'll be lots of options for patients to choose from. One of the challenges, like one of the things I think you're getting at is it's very expensive to develop these technologies. Right now it costs hundreds of millions of dollars. Nobody's ever gone from a blank page to putting a brain implant in a patient for a new indication for less than about $70 million, really more like $100 million in most cases. And as long as it is that expensive to develop, then you have to charge a lot to have it. And then you get in, then there's questions like, right, who's going to pay for it? Are there enough patients for that to be worthwhile? So I think it is very important that the cost to develop these things comes down. Now this is, I have a whole healthcare catastrophe theory of everything here, which is, I mean, like 20 years ago, we bought far fewer TVs and laptops and phones Each one was much more expensive. Now they're cheaper. We spend many times more total. This is a good thing. Investors can look at this. They can be like, cool, phones, 7% CAGR market, can invest in that, going to get good returns. There's much more money spent on it total. This is a technological growth industry, and everybody's happy. In healthcare, healthcare reimbursement dollars are more or less a fixed bucket of money. It grows at about the rate of tax receipts. And so the problem that we're running into is that there's more— people are living longer and there's more products that are able to provide value to patients. And so as patients live longer and there's more stuff that they can buy to get a high quality of life, there's just more stuff to spend money on. And when we spend 10 times as much on laptops, this is great. If we spend 10 times as much on healthcare, this would be a catastrophe. And the only way out of this is to bring down the cost of development so that you can charge human-scale prices for this. Like, I want nothing more than the ability to sell one of our retinal prostheses for like $5,000. It's not going to be $5,000.

How much is it going to be today?

Don't know yet, but it'll be more than that in the beginning because we'll have spent hundreds of millions of dollars to develop it. We've raised about half a billion dollars so far. A lot of that goes towards the retinal prosthesis. There's a bunch of different reasons for this, but I think one of the things that you have to be mindful of also is that when you look globally, this is now a globally competitive industry, and some of the other countries out there are able to do these things a lot more cheaply. In fact, The first— there's now a company that has a commercial marketing approval for an implantable BCI. It's the first ever, as far as I know. They're in China. They did not spend hundreds of millions of dollars to develop this. And one of the corollaries of that is that they don't need to charge hundreds of thousands of dollars to each patient. And so I think finding ways to develop these in more capital-efficient ways is going to be essential at realizing the potential of these different applications. Because when you look at these other companies, it's not that I think that what they're doing is not worthwhile or that there's not a market for it. It's just a question of, Does the venture math work at the scale that it takes to develop these things? So one of our— it's not really a product line, but one of the things that we do is we license our electronic stack and we own a foundry and we work with other BCI companies to enable them to become capital efficient. So far, all that we've seen are these highly funded, vertically integrated companies that are spending huge amounts of money to develop these things from scratch. The reality is that a lot of it is the same. If you're building a cortical motor decoder or you're building a retinal prosthesis or a cochlear implant or a spinal stimulator, a lot of the electronics are the same and a lot of the fab processes are the same. And our goal with our ecosystem program, which we've actually done a handful of these now, a couple have been announced, there's more in the works, is to make it so that the smaller companies can raise $5 to $10 million and get into humans and prove out in human subjects and you affect in a clinical indication without having to raise $50 to $75 million up front to rebuild all this, make all their own chips, human-rate all their own software, make all their own electronics. And I think that will have a big impact for getting more of these things to more patients because there really is a— there's a very long— I mean, almost anything in medicine ultimately you can rephrase in terms of an impact on the brain.

Gosh, this is going quick. I have all these questions I want to ask you, but I have to turn it over to you guys in about 10 minutes, I think. OK, let's get weird for a second. Your— this perfusion vessel that you guys have, I looked at it kind of quickly over the last day because I didn't even know you guys were already doing this. But so you've kind of— made better versions of devices that get used in a hospital to keep organs going outside of the body. That was my first take. You guys do have a consumer tech flair to your design on some of these products. I'm going to get really weird with the next question. But first of all, why are you getting into this? The goal of making these products?

So I think when you think about the BCI field, there's a couple different memes out there of the motivation for these things. I'd say there's a couple frontier BCI labs, and then there's a bunch of medical device companies. One of the dominant memes on the frontier BCI lab side is that this is about kind of merging with AI, and that this is a response to the advances in artificial intelligence. And I do think that there's something to that, and I think that that is a thing on a slightly longer timescale. But I think of science really as a longevity company. I think that our real competitors are more like places like Altos Labs or Calico. And so when I think about the human condition, how can we improve this, we have this very strong existence proof in perfusion technology that miracles are possible. So I first got clued into the opportunity and need here about a decade ago when I read an article in The Lancet, which is a major medical journal, which was a case study of a 17-year-old living in Boston who was waiting for a lung transplant. And while he was waiting for a lung transplant, he was being kept alive on this perfusion circuit that was kind of replacing his lungs and to some lesser degree, his heart. And after he'd been on this, so he's in an ICU suite, He's playing video games, doing homework, hanging out with friends, but is totally dependent on this perfusion circuit that's keeping him alive. And after a while on the transplant list waiting for new lungs, he was diagnosed with a complication that made him no longer a priority recipient for new lungs. And so they took him off the transplant list, and this Lancet article was about the ethical dilemmas of like, what do we do with him? Because on the one hand, you've got this 17-year-old who's doing homework, and if you turn off the perfusion circuit, you will immediately die. On the other hand, it's $500,000 a month, and this is like, if this, like there were these aggravating quotes in this article from the doctor saying things like, "This raises fairness issues about if it's fair to leave this patient on ECMO as a destination, then why don't we do this for every patient in cardiopulmonary distress?" And how his family and friends, quote, "Derived benefits from his continued survival, and so we have to like reason about this." This was just like, I didn't, didn't really understand what the limitations were and needed to learn more about that. And so it's a thing we've done over the last few years. But clearly, there's this existence proof that these life support systems are capable of miracles that most people don't appreciate. And it's the same, and simultaneously the same type of technology has revolutionized organ transplantation. 20 years ago, if you needed a new kidney, you were on a 1-hour tether to the hospital at all times. If the car crash happened at 3 in the morning, your surgery happened at 4 in the morning. Now it gets scheduled for the afternoon or the next day with a rested surgical team. In many cases, these organs are coming from patients who are quite sick, which is how it became a donor organ in the first place. And just being on this perfusion circuit out of the disease state kind of brings them back to health a little bit. And there's really exciting research going on where if you could keep these organs for longer, for maybe a couple of weeks, then you could intervene on them maybe with a gene therapy ex vivo where you're not exposing a patient to it, just that organ. You could even immune match it to a patient. And this is all kind of done, but it's not— there are limits to the technology that did not feel very fundamental that we really felt like we got a good handle on, on where the opportunities and pitfalls were. And we thought could be improved. And so I think beyond organ transplantation, I mean, other things we're thinking about are like pediatric like, like if you're a 3-year-old that needs a heart, you're pretty screwed because a 4-year-old heart or 2-year-old heart won't fit. And there's not a lot of 2-year-old to 4-year-old heart donors anyway. But if you could build a system that they could live with, like as a backpack for a decade until they're eligible for, or 15 years until they're eligible for an adult heart, or maybe forever, then that just totally transforms how you think about these diseases. And this, I think is actually like, it's really not that crazy. I mean, we know that a lot, that the fundamental medicine there is potentially possible, but there's a lot of problems to solve. And we got to a point where we had the conviction to bet real money that we could solve them.

So, okay, I think a few months ago I wrote a story on a company called Becoming. They're making a placenta that they can keep alive via a perfusion system in their lab. Like sort of heading towards becoming an artificial womb, more or less. We just did a video and a story on a company called BexOrg. They use a perfusion system to keep a brain, a human brain alive outside of the body for up to weeks, I think they can do. Over the last couple of weeks, there was a story that broke about a company called R3 They're knocking out genes. They're basically growing sacks of organs, or want to, that you could harvest and use for some of these cases that you're talking about. You know, like when I write these stories, there's a group of people who are excited and think it's cool, and then there's another group of people who find all this sort of off-putting. Yeah. How do you think the public is going— it feels to me like we're on the cusp of some quite dramatic changes with this technology, and I'm curious how people will react to this.

Yeah, I mean, look, I think that if your child needs a heart transplant and there's nothing available— and so like in the case of that 17-year-old, what they did is they— there's a filter in the oxygenator in the circuit that needs to be changed every few days because it's It's this gas exchange membrane that clots as the blood flows over it. And so the parents and the doctors didn't want to just unplug this, be like, come into the room, they're like, cool, we're going to turn you off now. So what they did is they stopped changing the filter in the oxygenator and a couple days later the kid died. And I don't think that there's any parent who in that situation is going to be like, no, I don't want the artificial heart system. And I think this is very serious medicine. I think that people don't want to lose their humanity and I think it's important that the future is is of humans and for humans and that we preserve these values and we're not producing something totally different. But I think this is very serious medicine. And there's, at the same time, it opens up, as you mentioned, like new parts of, like new possibilities that weren't feasible before. Like I think you'll see groups work on things like artificial wombs.

Yeah, I mean, do you think, So I mean, check out Core Memory later this week. We're going to have some stories on some more of this stuff. But I now know of like a handful of companies that are working on growing organs, you know, essentially it's like a headless body with no skin. I mean, do you think this will happen in the next 10 years?

I mean, probably. I mean, that's like probably possible. I mean, we're not growing headless people or anything. It's like probably technically feasible. This stuff, it, I feel, so first of all, it's important to remember that everything we do is in dialogue with society. You need to get, you need to bring people on, you have to explain why you're doing it, why it's important. You have to, also there's places that they don't want you to go and you have to be open to that. But the, I think, Where I am, biology is this alien nanotechnology. It is this, like, we're given it by our evolutionary history, we have to deal with it, but we don't really understand it. We're not really good at engineering it. It does all kinds of crazy stuff. And there might be ways to simplify some of these problems. And so I think that for us, like the perfusion circuit approach is probably like better than growing replacement organs and also is a little bit less creepy. But I don't know that the creepiness is going to— I think if you can show real impact on patients, and I think that when people's feelings change about this, when they have a serious diagnosis or a loved one of theirs has a serious diagnosis and they go from thinking like, oh, I would never put a chip in my brain to being like, oh, I can talk to my husband again. And at the same time, there are people out there, because it's a free world, people can do what they want, saying kind of very speculative things, and that doesn't always help. But it is important to remember these are serious surgeries for serious diseases with the potential to really help people.

Okay, I'll do one or two more questions. Do we have microphones over? Can we— oh, in the audience? You guys— OK, Armand's going to help us out. If you guys have a question, put up your hands. I'll just do like a rapid-fire one since I see people already have questions. I get bombarded these days with startups who it's like AI meets bio. We've run our model. We're discovering new proteins, new viruses, you know, and it's like new medicines. There's so many that I— it's impossible for me to tell what's real, what's not. Is this actually happening? Is this not? I always enjoy talking to you because you're well-versed on AI and biotech. Yeah, I mean, are we— like, if you're sitting where I am, I'm being told the revolution is happening right now and AI is having its biggest breakthroughs possibly in biotech. And we are on the cusp of discovering all kinds of new medicines and therapies. Very quickly. Yeah, and I have trouble gauging how excited to actually be about this or not.

Yeah, I mean, look, I think that a lot of it is real, not all of it. And I think that it will be— some of it will come down to our ability to actually run these experiments in humans, because until you've done that, you don't really know. I mean, the laboratory mouse is an essential part of the process, but kind of everything works in mice. And so if you've done it in mice, that doesn't really prove anything. And if you've just done it on a computer, that also isn't enough. And the step from going from a computer model to trying something in a human subject is enormous. And so that is gonna take time. Not every, like, there's tons of blog posts being like, we found all these drugs. It's like a small percentage of those are actually gonna go to the clinic and even find out. But I think that the progress that's coming from AI is largely real. And the serious groups that have the resources and capabilities to bring these things through probably, I think, will get some major results. And there's, yeah, I mean, I think it's tough to give a really generic answer because there's such a broad range because there's just so much out there and there's new things every day. But I think that a lot of it is real.

Okay, okay, okay. I see a, oh, where's the microphone? There we go, okay. Thank you. Hi, Max.

Yeah, so I'm curious, like, my question broadly is like, how do we make sure that neurotech does not end up like pharma? Because like pharma has like a huge lobby, and I would argue that like it's not, it doesn't really like think about the patient first anymore. And so like we're kind of on the precipice of all this technology coming into public view, and all these players are entering the scene. But like we're navigating capitalism, and like every, like obviously these things need to make money, and like you mentioned we need to bring down the cost of development, but like how do we steer I guess, the development of this young industry away from what we've seen happen, I guess? So how do we make sure that neurotech doesn't end up like pharma? And I think that the central— going back to the thing I said earlier, the problem with drug discovery is that it's just you don't get these big effect sizes. And a lot of it is downstream of that. There certainly are, there is, I think that your perception that to some degree some of these groups have gotten good at separating financial return from helping people is true. I don't think that, that's certainly not the mindset of like scientists in pharma. I think that pharma in general is somewhat unfairly maligned. I think that they really are trying to find, they want nothing more than to find great drugs that have huge effect sizes and they can like get paid for and help a lot of people. Like when you have that, that's the most valuable thing. Like this idea that they're holding back a cure for cancer because then it's not like you're not getting more sick people, this is just not the case. The problem is that this is just beyond humanity's capabilities. And so healthcare is different than many other industries. Like 20 years ago, we didn't have big autonomous car companies because just like clearly it wasn't there. But we did have drug companies because humans don't like dying and they don't want to be sick and they want to have these things. And so it is a little bit different than many other areas of technology where at any given moment in time, for kind of the most resources that we can allocate to it, we'll kind of take whatever we can get. And in the case of drug discovery, it's just like limited. And I think the way that Neurotech is different than that is— well, it will be different or not to the degree that we're able to actually get those results. And so I would say I'd keep the focus on what is the effect size that you're able to see, to what degree can we bring down the costs of development. And some of these constraints are self-imposed. There are real regulatory costs.

The—

people talk about like, oh, it costs like $100 million or more to run a clinical trial. The direct costs of a clinical trial are actually not that high. Like the real problem is that they take so long to run and while they're running, you're carrying the costs of this big company. And so just like the duration of waiting for this is super expensive. And so if we— anything we can do to accelerate regulatory reviews or if we can delegate authority to local IRBs like many other countries do, or like with the FDA, there's— the FDA is small. There's not that many reviewers there for any given area. And so when you look at what like Europe does, where Europe has a system of notified bodies, where there's these businesses that are basically blessed by government to review products, they can have thousands of reviewers in an area. And so you can get much faster feedback and you can work much more closely with them. Those are the types of things that I think will allow us to really get the iteration cycles that allow us to learn, that allow us to make really compelling products. But I would say pharma is not evil. They just are limited by biology is hard and humans want whatever we can get, and that is the result of that.

Let's do one more on that side and then we'll switch sides. Yeah, sorry. I saw your hand first. Whoops. Sorry, one second. Got to get it for the podcast. Yeah, to this gentleman. Thank you. Great.

Thank you for the talk.

These great insights. In terms of—

with regards to PRIMA, what are some of the fundamental discoveries that you found that others have missed? And then how are you extending those learnings to some of the new products you're developing? Hmm. Why did other people not do this? I mean, this is always like a little bit confusing to me. So there are some, there have been some genuine technical advances. So for example, there's a material that, so when you put an implant in the body, you have to realize there are no truly passive surfaces anywhere in the body. Even bone is constantly getting remodeled. And so there's these, it's all very dynamic. And so when you put a device, it starts getting attacked and you want to keep your electronics like functioning and the metals in there from leaking out and you want to keep the body from leaking in. We call this packaging and packaging is like a weirdly hard problem. This is like one of the things that people don't really think about, but when you like these implants often degrade. And so the way that people have dealt with this historically, like if you think about like a pacemaker, The thing that you actually see is like a welded titanium box because welded titanium does fine in this environment. But if you put in like a tiny little like film, you've put in like a film of electrodes, this is going to break down. And so there's a material, silicon carbide, that is a pain in the ass to work with, but we're now getting experience with that. You can use it as an outer encapsulating layer and this will outlast the patient. And so there's a handful of results, things like that, that are really pushing the field forward. Now, I'd say that the central thing that made Prima more effective than the prior devices was the subretinal placement. So there have been a lot of groups that have tried to stimulate the optic nerves. So when you think about the retina, the light transmits from the front. The first layer of cells light hits are the optic nerve cells, past that, or the bipolar is past that, or the rods and cones. And so if you just like put an electrode array in the eye and you just lay it on the retina, what you're getting are those 1 million optic nerve cells. And it's the easiest, it's like the place to be. You don't risk a retinal detachment so much by putting it there. It's kind of an obvious thing to do. A bunch of companies have done this. 10 years ago, there's a company called Second Sight that got all the way to market with an epiretinal stimulator. But when they did this, they were just able to get these flashes of light. And I think kind of like with pharma, because when you're stuck in these technological local minima, you can rationalize a lot of things. And so there's a whole culture out there that thinks that more phosphenes is vision. And if I just get more flashes of light, then eventually these are kind of like pixels and I'm going to get vision. And neurologically, it doesn't work this way because you don't see the image on the eye. You experience the world model and really have to update the world model. And the phosphenes are not updating the world model in the way that you need. But there really aren't any subretinal implants out there. I think there's like one other device. And so just the placement of this under the retina, like if you go and ask a vitret surgeon who is especially does this, a lot of them are gonna be like, "This is really tricky. Like I don't know how to do that. You're gonna cause retinal detachments." Like this is not true. Like it actually is not that hard and totally fine. But that in itself was like somewhat contrarian. And then it was an empirical discovery that if you stimulate the bipolar cells, you get vision. And so then from when we think about competing in this, can other people see that and make a subretinal implant? Sure. But the real trick in Prima is the fact that it's powered photovoltaically. So let me explain a little bit this device. It's a tiny little chip. If you look at it under a microscope, you see all these little cells. Each one of those is essentially a solar panel. So when we place this under the retina, there's no cable coming out of the eye, there's no battery, it's just this little chip. And so like the Second Sight device, this was a 4.5-hour surgery with a belt around the eyeball with a titanium box on the side with a power coil and a bunch of electronics. Like we don't have any of that because the chip with the photovoltaic array is powered by glasses that project an infrared laser, and wherever that laser falls on the implant, it stimulates it. And so we're— By powering it photovoltaically, we're able to have this device that solves a lot of the historical problems. And this is the area where we have a lot of deep IP. And so there's a lot of different tricks that I think come from having— you have to start from this first principles understanding of what is the anatomy that we're dealing with and what is the goal here and what is the full toolbox of modern technology available for this. And it's taking things like, oh, this looks like an image sensor. We know how to make image sensors. Everybody in this room has a multi-megapixel image sensor in their pocket. This is really easy now. And if you take that same technology and use it in a different way, then you can put it together for this.

Okay, I promise this side, maybe do you want to just pass the mic there? Yeah, there we go.

So fascinating. I'm curious just to like a little bit of the future. I would assume the goal is to have like a one-to-one connection between these electrodes and neurons at some point. And so I'm curious, A, do you think that's possible, or is there some sort of, you know, second law of thermodynamics that's stopping that from happening? And if it is possible, is the goal at some point to have coverage of all the inputs and outputs? Like, are we gonna do that, you know, in 2050? Be all-in-vision? Yeah, actually. Yeah, well, so, okay, so two different parts to this. So the first is there, as you scale up, so the implant's small and it doesn't cover the full area of atrophy, much less the full retina. Obviously the eye is curved. It's tough to make chips that are that curved. So now you start thinking about multiple different triplets. As you have a larger surface area under the retina, you do risk a retinal detachment. You don't want the retina to come off the back of the eye. And so there are ways to deal with that. This is ongoing research, but there's, again, at Science, we have a broad mandate to look at how to solve these problems. We weren't founded around a single asset. The Primo was an acquisition. Now we knew what we were looking for, and I think we were very prepared for that, but that was, we're like, this is the best thing that works now. The electrical stimulator, there's a 4x4 grid of these two cell layers, or there's a gene therapy way to do it, there's an electrical way to do it. We kind of mapped out the full space of possibilities here, and we want to provide the best product for restoring vision to the blind. Now I think that that soon will be Prima. In 10 years, it could be Prima 6. It could also be a gene therapy. We have another vision program that uses a gene therapy that makes those cells light sensitive themselves. So instead of having a chip under the eye that can stimulate them with light, what if we could express a protein in the bipolar cells that basically puts the stimulating pixel in each of the cells? And so that would be how you get to kind of every cell you can communicate with. To do that, you need to solve some hard problems. Somebody has to run around. You have to solve some hard problems in that protein design, which I think our group in op engineering is the best in the world and has been able to make real advances in that, but has to go through the whole clinical trial process. Now, to the other half of your question about all of the other senses, I mean, so we've had cochlear implants for decades. We're getting motor decoders. There's interesting research going on in providing sensation back to the brain secondary to motor control. But I think that if you can provide vision, hearing, balance, and a kilobit per second of motor control, then really you're halfway to something really cool. And I think the big missing piece there was the visual processing. We have that now. And so I think that again, this stuff is coming on less than a decade time scale.

I think just for efficiency's sake, yeah, let's go there.

Lucky. Thank you both for the fascinating talk. So going a bit speculative, so how would you envision do we go from human to superhuman? As in, how much are we limited by the bandwidth of the optical nerve? Is the strategy that you mentioned before of having a field of vision that is centered around something and like less in smaller definition at the edges, is that already optimal for the bandwidth that our brain can handle? Or could we just plug in a high-definition camera to the eye and the eye would and we would have somehow a different perception of reality. Yeah, I mean, this is— so these bottlenecks are co-evolved throughout the brain. I think sometimes people think like, oh, if I just— if I have a BCI, then I'll be able to communicate. I'll be able to speak so much faster. I'll be able to like talk to— like have an AI that's really fast. This is— I don't— that doesn't align with my understanding of the neuroscience here. I think writing or language really is thinking, and the visual input to the brain, again, is not a camera. It's about updating this world model with these sparse samples, and the brain's very efficient at this. And so the— now with that said, that doesn't mean that some really cool things might not be possible. It's you have a finite visual field. You can't see behind you. Is it possible to, through a BCI, just give you more visual field? That is somewhat more speculative that nobody's ever done that, but there's some really intriguing neuroscience that suggests that it might be possible. And so that's a little bit further away, but I think we'll see.

Were we at a talk together, like a salon, and people were talking about merging brains from different species.

Maybe.

When we were out near Stanford.

Yeah.

Okay.

I mean, so all communication is about creating correlations between brains. So when I, like, I have a thought, I can serialize this to language, or the language kind of is directly related to the thought. I can send this over vibrations to the air, you see it, you unpack it, and then that lights up a concept space in your brain that corresponds to the idea. That depends on us already sharing these concept spaces. You get these when you learn a language or when you learn math in school. And if you don't have that, if you haven't taken calculus, then trying to communicate the idea will be lossier. You'll have to do it through metaphor. You'll have to find some other way to map it. And one of the really interesting things that has come out of AI over the last last year or two is this empirical finding that when you look inside these big AI models, the mathematical objects that you see look a lot like the representations that you see in neuroscience.

Like neurons, groups of neurons light up for different concepts.

Yeah. So for example, there's neurons in the brain that represent space, like any possible object or any possible face or different ways of making decisions. And when you look in these neural networks, you see things that look very, very similar. And so that was one of the first clues to me that the AI people were really on the right track. And this idea that these things are stochastic parrots is just totally wrong. And so you can get this clearer sense of what these concept spaces are, how they're developed, how they're implemented both in deep learning models as well as in nature through biology. And so when you look at different species, Obviously, we don't share languages, but you can still ask, well, what are the shared semantic spaces and how might you be able to map these? And I think there's been a little bit of work done on this in interpreting whale song and birds. And some of this is very stereotyped. In some cases, there's things that look more like grammar. But underneath, you can think about what are the semantic representations and how can we map these? And I think that there is going to be a lot to that.

How are you doing? A couple more? Sure. You okay? You guys okay? All right. I promised this person over here. Yeah. Yeah. Much love, Ashlee.

Thank you. Yeah, I'll keep this quick. So, every agent, whether it's an LLM or a human being or a cat or a mat, like, or a mouse, has a privileged ontology. The only reason language works, as you were saying, is because we have shared manifolds developed over billions of years. Those manifolds were sampled over many, many iterations of evolution. So my question to you is more about future life, artificial life. Let's suppose that I can 3D print not just any brain, I can design it at the cellular level with such resolution that I can choose to create 4 lobes instead of 2 lobe brains. I can choose to put the ventral and the occipital pathways in the basal ganglia somewhere else, maybe closer to some other semantic cluster in the brain. Now we're talking about enhanced or different cognition, different forms of human life that didn't exist before. How much of your current process of brain mapping and so on would stay, and how much of that would have to change?

Now we're getting somewhere. The last question has to top that.

I mean, I don't think that— again, we're learning a lot about engineering brains, biological and synthetic. I think the basic principles are kind of the same. And many things, like to your question, can you have a brain with 4 hemispheres? Nature has done that. There's a pair of conjoined twins in Canada that are one head with 4 hemispheres inside, and it's joined in, like, really interesting ways. And when they were children, there was imaging that was done, and they tried to tease apart, like, what did they feel, what elements of their conscious experience did they share? And I mean, it's impossible for us to imagine, but that doesn't mean it's unphysical. And I mean, this is not the type of stuff that we do in practice. Give me 10 years, we will be doing it. But yeah, I mean, I think that it's all like the principles are the same. I think like the— Yeah, I think one of the— There's one failure mode, which is new ontologies that don't exist in human language but exist in a human biology we created artificially. That's a hard problem to solve.

Okay. One last one. Who's going to take us to sci-fi land? Yeah. Can we pass the mic? Is it still over there? Yeah. Okay. Thank you guys. Also, there's hats at the back. I still see. Can you hear me?

This is going to be a fun one, light one. So you said earlier, I'm curious, what we see is not reality but a world model interpretation of reality, etc. And so I'm wondering, is this something we know? Is this something you think specifically? Do we know that there's a discrete boundary between perception and interpretation? Because if there is, and I don't know about this, I haven't come across that, but if there is, then It would be pretty easy to boost activations or something and then build a night vision app for your stuff or Beautify app for your implants. And that would be a great product, a little fun too. Well, I mean, so like night vision we already have, like you can put on goggles, right? Like we use night vision all the time that embeds into regular vision. The, like the question of could you develop, could you engineer a totally novel phenomenal mode with a different structure, something like a vector sense, something that we really don't have? Yeah, I mean, probably. Evolution has done a pretty good job of spanning the information available. They run on the same hardware, so we don't know the boundary between what I perceive and what I understand. I'm asking you if you know that there is such a boundary, and if so, can we tap into it and add a filter? Because that'd be a great product. Well, I mean, so in a basic sense, like, this has to be true, right? All the brain— like, the brain is in encased in the skull is not magically or telepathically connected to the world. Like when you see the wall, like the room out there, what you're actually experiencing is neural activity, and all the brain ever gets are electrical pulses down a small number of cables, the spinal and cranial nerves. And so in some sense, the electrical activity in spinal and cranial nerves is— that totally defines your connection to reality. And the— like, you all of any of your experience is neural activity. And so if you can alter the brain, you alter your effective reality, like in a trivial sense.

Okay, I'm going to ask one last question. Thank you all so much for coming. Thank you, Max, for doing this. I always feel like you always end up as like a philosopher in all these these discussions, and I appreciate that you're able to go there so eloquently. It's kind of like why I love talking to you. It's so thought-provoking. So maybe this question is— I don't know if it's too easy or cheap. I'm just always trying to get how you picture the world. I'm going to pick 20 years, but you can kind of like—

No, I can't see 20 years anymore.

Well, okay, so pick whatever number you want and whether it's sort of AI or more biotech, you pick. Just try to like, in your head, you know, what do you think if it's roughly 20 years? Like, what do you think would just be the most surprising thing happening in the world at that time?

Yeah, I mean, I've got this like event horizon in my mind at about 2035 now. I mean, I think that the transition, like AI is real. Like I think that is not going back in the bottle and it's not hitting a wall. And I think there's going to be many things like BCI is independent from that, but we're going through another industrial revolution. Like there's, for thousands of years, we were living in a society that did not have a concept of progress. Like if you were a medieval peasant, there was no sense like, oh, there's going to be new products soon that are going to like be better. It was just like, it was static. And that was deeply ingrained in the perception of reality. And now we have a sense of progress because you might have seen the graph of, if you just Google GDP over the long run, the graph of productivity, it's flat for thousands of years. And then in the mid-1800s, it does this. And I think that there's this idea sometimes that this is recent, like, oh, this started in the '90s or this started in the 2000s and it's been accelerating since then. No, this started in the mid-1800s and it's been accelerating since. But there was a generation that went from looms in the living room making textiles to working in factories in the city. And over a short term, their lives got worse. It was not fun to go through the Industrial Revolution. And this led to things like the labor movement and led to environmental regulations and a lot of the modern regulatory environment came from the experience of acutely people's lives got worse and bad things happened and we learned from that and then we learned to thrive in that new world. And then after that, there was a generation that went from not having a horse or riding a horse and buggy to seeing a man land on the moon. That was roughly a lifespan. And we're going to, over the next 20 years, the transition is of that scale. I think it's important to remember this has happened before. This has happened several times before. I don't think there's anything you could have said to somebody in 1900, before relativity, before cars, before jet aircraft that would've helped them understand what their retirement was going to look like.

Yeah. Okay. Well, I hope you guys have enjoyed this. I enjoyed it. Thank you, Max, so much. Like a really heartfelt thank you. When I saw the invite list and I look at like our subscribers and who's reading our stories and everything, I saw so many familiar names, even though I feel like I know some of you. And so, you know, we've— Core Memory has been at this for a bit over a year. So thank you all so much for showing up, for reading and watching everything we make and coming out to meet us in person. I don't think we get kicked out of here. I think there's some more food and drinks. And so if you guys want to continue the discussion, please do. And just thank you. The Core Memory Podcast is hosted by me, Ashlee Vance, and/or Kiley Robinson, 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.