Engineering Brains - EP 09 Max Hodak

I was in 5th grade and literally saw The Matrix, and I was like, I don't know if we're living in a simulation, but we should build one, basically. Look, Earth is small. Resources are intensely contested. Space is big. The speed of light is low. There are real limits to building with atoms and doing things in the physical world where there's these conflicts, like a piece of land can only be used by one person at a time. Building buildings is super expensive and complicated. But like in the simulation, none of those limits are true. In the simulation, there's no scarcity. Have you heard of the Jennifer Aniston neuron? No. So there's a well-known result that there's some, in some part of the brain, if you record from these cells, if you show them pictures of different people, they in some cases will respond very selectively. And so there was a Nature paper a while ago that showed that if you showed pictures of different people, there was a cell that would fire just for pictures of Jennifer Aniston. Or if they showed you types of different houses, they would fire for just specific types of architecture. And so the intuition here is that if you look at all of the millions of neurons in that area, encode a representation of architecture or of faces or of people or of places. And this is embedded on a mathematical object called a manifold. And so what you want to do is you want to access that statistical representation somehow. And when these neurons grow into the brain and they wire up, you get all this— when the cells in the brain fire, you're getting all this partial information which represent information about the statistics of those representations. And so you can access this information in and out of the brain through this. This is a very, very deep well of neuroscience that we are like at the very beginning of. Just making these devices at all is pretty tricky. I would say this is like up there with like modern semiconductors for complexity.

Hello everyone, it's Ashlee Vance, your friendly neighborhood host of Core Memory. Our guest on this wonderful episode is Max Hodak. He is, um, co-founder of Neuralink, currently CEO and co-founder of Science Corp, which is an amazing name, and we'll get to that in the podcast. But if you've been reading the site at all these first couple months, we've written a couple stories about Max and Science. They're doing some of the most cutting-edge research and products imaginable in the brain-computer interface field. They're helping people get their vision back, and then they're pushing us possibly in some very weird cyborg-ish directions. So you're gonna have a great time listening to this. We went to Science's headquarters in Alameda, California, and, and Max was full of weird and wonderful ideas as ever. Maybe you're getting this podcast off of Core Memory. I just want to remind you, you can get it off iTunes, Spotify, wherever you like to get your podcast, please do us a favor, leave a review, a like, all that stuff. Um, it'd be a great help. And with that, let us dive right into the brain and Science Corp with Max Hodak. Max Hodak, hello.

Hey, Ashlee Vance.

I would say welcome to my office, but I'm in your office.

Welcome to science.

Yeah, for— so for people who don't know what Where are we?

This is our little pocket of the future in Alameda. Uh, when we started the company, we, uh, the, the brand, we wanted this like white mirror optimistic vision of the future and we created these landscape renders that are on our website now. And we wanted this thing that we, people wouldn't tell if it was like far future or ancient past. I think those kind of blur together as technology disappears. And, uh, when we moved Here we took some photography, updated the website, built some out, and people actually sometimes couldn't tell like what were the renders and what was Alameda. And so it's like this far future vision of utopian future. It's like, or you could live in Alameda.

Yeah. Alameda is a cool place. It's, it's underrated and it's been transforming as a bunch of really cool startups set up shop here.

Yeah. I mean, I have, I think, a fairly encyclopedic knowledge of Bay Area lab real estate. And it's— I mean, I say this, this will of course change very quickly, I'm sure, but like cheaper than a lot of everywhere else. And it's also a good commute for whether you're coming from the East Bay, SF, the Peninsula, kind of all over. And so employees really like it. The— it's its own power grid. It's not PG&E, which is actually underrated for a lot of things, especially if you need a lot of power. It's like an idyllic community. And it is very cost-effective for this type of infrastructure.

And just to paint a picture for people who can't— some people will see us sitting at this table, some people will just hear this, you know, what happens in this from sort of the— what are you guys doing here from wet lab to hardware? What the scope of what takes place?

We do a lot. So we have everything from a large animal vivarium to cell engineering to a small MEMS fab. And we've got some tools here. We also own a captive MEMS fab in North Carolina. I mean, we make kind of— we're like a full-stack neural engineering company. Our main focus is on the retinal prosthesis. We started the company, we thought that the time was right to restore vision to patients that had gone blind due to losing the photoreceptors, the rods and cones in their eye. If you want to do that, there's two layers of cells past that in the retina, the retinal bipolar cells and the optic nerve cells. And we weren't totally sure whether it would make sense to stimulate either the bipolar cells or the retinal ganglion cells, the optic nerve cells. But if you could figure out how to selectively drive one or the other with some other— it's of course much more complicated than that— then you should be able to get vision back onto the wire going into the brain at like the first possible opportunity past the dead rods and cones. And we work on, we work on gene therapies, we work on electrical stimulators, we work on these really cool, pretty complicated cell-based neural interface technologies. But yeah, I mean, we're especially excited about the Prima retinal prosthesis, which is getting submitted to regulators for marketing approval now. We finished a clinical trial last summer in Europe, 38 patients, as far as we know, first time ever that these blind patients have been able to read again. And so not approved yet, but the clinical trial results look really, really great. And we're hoping to have that in like through the regulatory process soon.

All right. I want to, I want to deep dive on all the technology. But, you know, ever since I've known you, I mean, you're— lots of people are interested in the brain. I would say it— I could feel it exuding out of you, your interest in this field. I remember we were talking about you actually wanted to bring something like The Matrix make it real. You were one of the first people I had ever heard who, who just like willingly admit this out loud and, and, and had this passion.

Is that weird?

I think it's, it's maybe a little weird. I don't even know if I care about that. I think there's like people, obviously people who would see The Matrix, I think is like a dystopian sort of thing.

But even just, yeah, I mean, certainly like the, like the human batteries machine state, like that's not the part of The Matrix that I'm into.

Right. Yeah, yeah, no, so that part I got. And then I just think, yeah, I don't know. There's not a lot of people that say something like that with such conviction, I think.

Yeah, I mean, the— like, your entire conception of reality is based in electrical activity in your brain. That's like all of your perception, all of your thoughts, all your memory, all of your reasoning. This is all really emergent from the electrical activity of these billions of neurons. And so it just seemed like very clear to me that if you could, if you really believed that engineering with that substrate was possible, this would enable some really incredible things.

And you know, this is what I was curious about. Is this something young Max was into forever, or is this something you came to in college?

And oh yeah, no, no, this was, I saw that like, this was a path I was on really since 5th grade trying to figure out how to, how to work on these things.

Like what happened in 5th grade?

I mean, I mean, the story was I was in 5th grade and literally I saw The Matrix and I was like, I don't know if we're living in a simulation, but we should build one, basically. Look, Earth is small, resources are intensely contested, space is big, the speed of light is low. There are real limits to building with atoms and doing things in the physical world where there's these conflicts, like a piece of land can only be used by one person at a time, building buildings is super expensive and complicated. But like in the simulation, none of those limits are true. Like you can, like it is, I think, like in the simulation, there's no scarcity. And how, like it is, of course, like anything's not that simple, but that at least early on, I mean, I think I now have more motivations for working on this stuff. It is not just building the simulation. I also think AI is real, making that not an other, merging with that. That's the thing that I didn't necessarily, it was not really in my initial motivation for working on neural engineering, but is increasingly so. And so I think there's many different paths to this, but that was part of my origin story.

And you're like, you're super bright. How did you— when you had that impulse in 5th grade, did that mean you started reading tons of science fiction? Did you get into computer science? Did you get into biology, or did you not really take action as a 5th grader on that?

Well, I mean, I started programming when I was really little. My parents tell me that I sat on the floor of a Barnes Noble and cried until they bought me a Learn Visual Basic book., or like a, like a BASIC.NET book or something, or well before .NET, like a, like a QBasic book. And, um, they were very skeptical I would actually make anything. Um, and now 30 years later, I've written many programs. Um, but I've always been around computers. I think, um, when I was a kid, it was like the compiler was the thing that didn't care how old you were and would take you seriously regardless. Um, But no, I mean, there's not a lot that a 5th grader in the late '90s can do to really work on neural, like to work on brain-computer interfaces. So it was reading, it was being an enthusiast. I mean, I certainly wanted to make stuff, but that wasn't really practical. Ultimately, one of the reasons I went to the college I did, Duke, was because there was a lab that at the time was doing some of the best work in the field, Miguel Nicolelis's group in Duke Med Center. And When I showed up to Duke as a freshman, I like pretty quickly figured out how to get into that lab. So the, uh, I remember I, my freshman year advisor, there was like, is there a lab that you want to work in? Is there some internship? And I suggested, uh, Nicholas Ellis's group and got immediately rejected. They're like, they don't take undergrads. It's like, that's the medical center. That's not accessible. And so I eventually figured out that I could do an independent study through the chemistry department. There's a class called Chem 26S. But didn't actually require you to work in the chemistry department. And I was able from there to get into the Nicholas lab. And that was really where most of my education happened in college.

But okay, so but even at that point, did you— were you thinking— you thinking I'm going to be an entrepreneur or are you thinking I'm going to get like a PhD in neuroscience?

I don't think at that point I had a particularly strong sense of that. I was always kind of an entrepreneur when I was in high school. I had like did web design consulting with some friends, kind of figured out how to make some money come out of the internet to buy myself a laptop. And then I didn't really enjoy school. I was never really like— I was a fine student, went to a fine school, graduated with a degree in biomedical engineering from a pretty good program for that. But I never like liked college or liked school for the matter. Um, and so I'd always been somewhat entrepreneurial, but I don't know. I mean, the reason that I ended up going into industry was really more driven by the logic that, um, if in academia when you're applying for grants, if you apply for a grant, these are these philanthropic funders that if they don't fund you, like, fine, it's their money. Who are you to say that they're wrong? Like, they can pick what they're interested in. And in industry there was— if you— it just made more sense to me why someone would give you money deterministically, because if you can create this closed economic cycle, create this return, it didn't— it just— it made me feel like I'd have more agency over my life. And ultimately, when things work industrially, you can marshal much larger amounts of resources than you can in really any other way. Although back in 2011, 2012, when I was graduating from college, it wasn't— it was clear that BCIs were still away from aways from translation. I figured when I was graduating from college that you'd probably need like $100 million in 10 years to build a BCI company, which I didn't have at that time. And so I started a company in something I thought was more accessible and building automation for drug discovery, which was also interesting. It had some—

was that transcriptic?

That was transcriptic.

Okay. And just for people who might not know, a BCI is a brain-computer interface. Yeah. Yep.

Yeah. And I considered going to grad school, But I figured that if BCI is 20, 2012, BCI is probably 10 years away from clinical translation. You go get a PhD, graduate 6 or 7 years later, random postdoc, don't have access to real resources. BCI would probably still be 10 years away from clinical translation. And so I just, I thought that I'd have more agency over my life learning to build companies, get access to those resources, and then be able to deploy them towards figure out how to do it profitably, basically.

Did you grow up in North Carolina?

No, I grew up in Manhattan.

In Manhattan. So was obviously an entrepreneurial place where these ideas might be floating around. But had you— had you thought about Silicon Valley? Were there people you kind of looked up to or anything like that?

I was growing up, I was like vaguely, dimly aware of Silicon Valley. I was a kid through the dot-com era. I was— I think I had like a pretty dim awareness of it. New York was a totally different culture, subculture. No, I mean, I drove out, so I did 2 years of college. I actually took a year off. I had— so when I was in high school, I was concerned that, like I think many high school students, I had no future, was never going to get into a good school. I got a bunch of statistics from my guidance counselor's office, entered those into Excel, built a model, and kind of— you can— college admissions were this, at least used to be this kind of interesting like there's some game theoretic interesting elements there. And so I built a model that predicted like if I applied early decision versus early action and made like what was the best school that I could get into that aligned with my interests. And one of those interests, of course, was wanting to work in a good BCI lab. Um, ended up working out that my— what I thought my optimal move there was to apply to Duke early decision where I thought I'd have a very high chance of getting in. You can of course only do that. I got in, I was committed., and then that eventually, that model eventually turned into a small SaaS company, um, and then I, I took a year off college, moved out to California, ran that for a little over a year, um, that got pretty quickly a very small acquisition, um, but no, I showed up, I kind of drove across the country as a sophomore in college with a friend and, um, got an apartment in Mountain View and didn't know anybody, and it was like very lonely for the first year and eventually started making some friends and then like 15 years later, you look around and like your friends are everywhere. But that was not where I started.

So in this year off, you just took a year off from college, started a company and sold it.

Yeah. I mean, it wasn't like— it wasn't like a big success, but it was a really interesting— that was— I mean, that was really my introduction to Silicon Valley. And I think I remember my co-founder from that was— he's like a great— he's a product guy. But we ended up at this house party once in I want to say somewhere in South Bay. I forget exactly where, but when we went, it was, it was advertised as like a hackathon type party and I didn't really know what to expect. He, he thought that there's like no way that people are going to be on laptops. It's like you're going to be like, it's going to be in the backyard by the pool. Like people are going to have beers and you walk in and there's just like fold-out tables of rows of laptops. And I was like, these are my people.

Okay. One last question along those lines. Because I was wondering this as you were talking, I mean, are your parents engineers at all? Were— was there any reason for you to be wired this way from your background?

I guess like my, my dad has a degree in aerospace engineering undergrad, and then they met at business school. They both went to— they met in their MBA at Wharton. And so I grew up around like a business sense.

Okay.

My dad's a management consultant. My mom does like ESG and philanthropy stuff. She was actually at Lehman in 2008, which was a really interesting form. I have this list in my mind of like impossible things and like the failure of Lehman was like entry one on that list is like there's things that you never think. I mean, this sounds simple now because it's so long ago, but like as a kid growing up around these institutions, you see these things happen that you're like, there's no way that would happen. But the— yeah, I mean, my, my dad could program. They— we grew up building model rockets and doing stuff like that. So I was, I was a bit exposed to it.

So when you were a freshman, what year, what year did you go to college?

That was fall of 2007.

Fall of 2007. And you're already thinking about brain-computer interfaces because, you know, they were not the most exciting, very academic But you, you, this field is a lot older than that.

Like the Utah Array was invented in the mid-'90s, right? But you'd even, cyberkinetics had been a startup.

Yeah.

Yeah.

You'd been following even stuff like that.

Oh yeah. Yeah.

You would, you, you would, I mean, I definitely never heard of a Utah Array, you know, until, um, I would say relatively recently, but so you, you were like genuinely into, to even as a teenager pursuing brain-computer interface stuff and, and Yeah, I mean, it wasn't—

I mean, I also had like— I also had played sports and did other things, but yeah, I mean, I definitely knew about— was following the field and reading about it. Like there wasn't a lot that I could do at the time on that. I was reading and being interested in it, but— and dreaming about it. But I mean, this is pretty cool. I'm like very— it is pretty cool to be here in 2025 getting to do this. That, yeah, I don't think was necessarily a given.

So that was my next question. I mean, you happen to be— coming of age is probably not the right phrase, but getting into your professional career right as money starts to come into this technology, as it starts to transform from a mostly academic pursuit to lots of brain implant and BCI startups. I mean, that's like you just said, that's sort of fortuitous for somebody who's good at this stuff.

Yeah, I mean, I think that this is also, I mean, this is a, like a narrow slice of just the craziness that we all seem to be here right now getting to do this in what appears increasingly to me like to be the takeoff era. And you can see how in some people this kind of tugs at the sense of like, are we living in a simulation? Like we're here getting to do this stuff in this time when every like AI is happening, BCIs are happening. The internet has, uh, has matured. Um, it's, it is definitely like, I think the most exciting time in, in human history. And I think that that will continue to be true.

So I can never tell because I think like you, you probably hang out with even, I don't know, I spent my whole life hanging out with people at the frontier of science. And so it starts to really skew my sense of reality or, or what other people get to encounter on a day-to-day basis. You probably even have a more extreme version of that. So I, because I have a similar feeling to what you're describing sometimes as this takeoff point. And then I try to put myself in the shoes of someone, I don't know, whatever it is, like 40 years ago. It's like, it's like, if you're inclined this way, do you always sort of feel like this to some degree?

Well, I mean, have you seen, like we talked, we refer, like we think of it as like the graph. Have you seen the GDP over time going back 2,000 years?

Yes. Yeah.

Yeah. So, I mean, it is, it's It's exponential, it's a hockey stick. And I think the thing that I only, I think I realized maybe a year ago, but I hadn't really appreciated before, is this acceleration didn't start 5 years ago or in 2001. This started in the late 1800s. This really started with the Industrial Revolution. And the acceleration has been continuous. And the, And the doomers have always been there. Like the Malthusians were kind of the OG doomers on this because they look at this and they're like, this can't possibly be sustainable, but it all just gets so much bigger. And that is what is possible with technology basically and engineering applications of it. And that has all just gotten done so much further. I mean, it just breaks intuition constantly because it's tough to think that way. But yeah, I mean, I think in the early 2000s or like in the 2010s, people were saying like, oh, there's a generation that went from horse and buggy to man on the moon. Are we ever going to match that? Because can you imagine going from like pre-car to jet planes routinely? And I think now the feeling of like, yes, that has not stopped coming back.

Yeah, I remember distinctly somewhere around 2012, there was a paper from a physicist, Who I think worked at China Basin Lake, and he wrote this paper looking at patents over time, and he was arguing that all the big ideas had been found. And then, you know, since then it does feel like we, and this was like in an era when, when like the iPhone had come out, but then a lot of tech was kind of boring because it was like Facebook and things like that. And so it sort of fit.

By the way, these are only like, it's crazy that we thought that these were boring at the time. Because the, like, the two things that you cited, the iPhone and Facebook, have dramatically reshaped the world.

Well, I meant the iPhone was a big one.

That was a big one.

But then things had kind of gotten a little, a little, uh, I think the iPhone even is more dramatic in retrospect.

Yeah, like we talk about like opioids and the effect of, um, the just how strong an effect on a human brain a connection to a phone is. It's like, it's not it's not that it's under— people get this, but we still don't really get how dramatic that is. Oh no, I mean, I think it's like people get so anxious being away from it for a couple minutes.

Yeah. And then you— we were just with a group of 4 people today and everyone's had this experience. It's like you have small talk for a second, everybody gets bored and they check on your phone, check their phones together.

That is a crazy technology. Yeah.

Yeah.

And then Facebook, I mean, it, not Facebook specifically, but social media, it was less of a technical innovation as a social one, but it has profoundly reshaped the world.

Right. Right. No, but I mean, okay. No point taken and agree. I think it was that sense of like, like you were talking about of the light bulb or, um, the internet.

We wanted, we wanted flying cars and we got, uh, 140 characters, but the 140 characters reshaped Earth. Yeah. And I think this is really underrated in, in this communication is worldwide. Yeah.

Yeah, yeah, no, that's true. I think sometimes like hardware things are easier to— for people to immediately identify that yesterday we didn't have a semiconductor, you know, now we do, and, and it unlocks all these things. Obviously not, not instantly, but, um, okay, so but you feel like in this particular moment, because I think there are people who would argue, yes, something like BCIs are exciting, gene therapies are exciting, um, all kinds of stuff that I cover, energy, All of these things are exciting, right? But, but still feel like not guaranteed at all.

I, I agree with that. I mean, I think BCI in particular, there's been a lot of capital and a lot of interest, but there's still no BCI company making a couple hundred million dollars a year or more in revenue. And until that is true, there will always be a risk of winter. I think that that risk has passed. AI is through the looking glass on this. But BCI is still on that side, which is, I think, one of the reasons that our focus is so strongly on get the visual prosthesis to market, get profitable, and then have that like actually start to have capital markets start to see, oh, there's actually economic return here. Because the flip side of getting to deploy these types of resources is you have to close the economic cycle. That cannot be forgotten.

But if I go down the list of, of DCI, vertical takeoff, you know, taxis, cell, partial cell reprogramming. My sense is you're, you're like net positive on, on this kind of bucket of things that we're on the cusp of.

Yeah. I mean, I think for a lot of these, it's really the, they're really held back by regulatory decisions. And I don't mean that dismissively. I mean, like, we can make decisions that we give, that give us a clean, healthy society with all kinds of nuance and trade-offs there. But absolutely, if you had no, like, no environmental or airspace regulations, we would have flying cars all over the place. Um, you'd have, uh, eVTOL everywhere. Um, but we, uh, like, these are societal-level conversations. I think that you'll start seeing these things deployed differentially in different regions that make different trade-offs about these things. But I think it's, um, it's more than just the technology. I think that the technology is great. And even on the drug discovery side, like gene or cell therapies, I think that that you've got a knob. Society is a knob, which is how quickly and how much do you go, do you do in human subjects research versus like how fast do you get new drugs? And we have largely turned that down to -5 societally. And this isn't like the, this isn't like the fault of the FDA. This is, this is the culture of, of It's— this is— there's no like magic bullet here. There's— it's tough because nobody— you don't— you as a patient, you want things that are safe. You don't want anything to be made worse, especially if it's a family member. Like, you see how you get there. I can tell you on the other side of this, at the point where you're doing a brain surgery for the first time or you're dosing a human with a new thing for the first time, that is a very stressful experience. Like, that is not something that is done lightly. And so there's just a lot of concern around that, which is justified. But at the same time, right, like if you look at what is possible in the animal models, like we have amazing technologies. It's that translation because of our concern for other humans creates a lot of friction. And I don't have like a magic bullet for that, but I think that's the tension.

And you're, you outlined some of your, I want to get to your idealistic, your optimum future, whether it's even like in your lifetime or possibly a couple more generations. You, you said some of it when you were talking about the Matrix of living in this virtual world that's bountiful. And, and if we were watching, you got me to start watching Pantheon, in which, you know, in, in that show, people are able to upload their minds and into a computer and, and except for a glitch, otherwise would be able to live forever, maybe with their personality. I don't know. I was just kind of curious.

Yeah.

I mean, where you want to see this all go.

It's tough to talk about this stuff without like sounding like a lunatic. Um, I'll be honest, I don't at this point have like a concrete vision for how I think 2040 or 2045 is gonna look. I, there's like, I, I always like growing up for a lot of my life, I always felt like I could kind of see the future and this guided what I did. And now I just have this event horizon at like 2035, maybe even 2030, and it's just impossible to see past that. Um, And I think part of it is when you look at a lot of the depictions of, of like the future in science fiction, like another, another one of my favorite stories is The Expanse, which is fundamentally a political drama. And the thing that is missing from so many of these stories is AI. And that is often intentionally left out because at the point where you've got AGI, it is no longer a story about humans and the humans want to see stories about each other. But that is, that can't be left out. It won't be left out of the future. I think the only story that I think really gets this, that I feel is compelling, is Iain Banks' The Culture. And that is a very optimistic vision, but a very, like, also grounded one, where it's not like this utopia, but it's not all good. Like, they could spend a lot of their time dealing with really complex moral and ethical trade-offs that you get in that type of society. But it's, I think that there's, that huge transition from a pre-AGI to post-AGI society. How does that impact on scarcity? Obviously, I mean, one of the lessons of Bitcoin for me is like, there's no scarcity information. The internet can make as many copies as you want. And so we created a new source of scarcity. And that is a transition that I, I think is just so out of distribution that is still pretty hard for me to reason about. But But I remain optimistic. Like, I think that there's going to be a lot of uncertainty, but success is absolutely one of the possible outcomes here. Like, it is very possible that that future is amazing, and we have to figure out how to get there. But that's going to be way beyond— I mean, I'm just often feel buffeted along by much larger forces.

I feel like you're holding back. I know you— I've heard you riff on the future. When you started Science in '22?

Mid-'21.

Mid-'21. We had LLMs and not like—

They were around.

They were around, but not, you know, sort of after that ChatGPT moment. It was before that.

Yeah.

Like why, why start a brain company as opposed to an AI company?

I mean, I, every now and then, I mean, I think that my life would be in some ways easier if I was doing AI. AI has a lot of nice structural advantages. You can do it with a laptop. You can inspect the weights. It's like, you can, it's really more engineerable. Than biology. It grows a lot faster. You don't have these time constants in biology. There's a lot, like, there's something very, it's like the, you often have to like actively prevent engineers from re-implementing cryptographic primitives in ways that are totally dangerous because it's just like you've got an F-16 sitting there and you're like, you can't, you can see it, but you can't touch the F-16. It's like very tough. There's like, it's this like glittering, the, thing. And so working on AI is obviously very alluring, but also I felt like it wasn't the thing. Like these are not, these fields are not just really driven by people that come into them later on. I think the people that have been doing this, like Ilya, like Andrej, like, like the people that are like, like Demis, like they've spent their life working on it in the same way that I've spent my life really interested in BCI. And I think that there's a lot of, judgment and intuition that gets built doing that, that I didn't have. They were clearly better positioned. This is, I felt like, my contribution to the world. It was a thing that I think I know much better than almost anybody else. Um, and that also then like capital followed that. Like you could, I could kind of start my career over switching to AI. But, um, one of the things I found in '21 was a bunch of my friends had seen what I'd done previously and, um, were willing to put up a pretty large amount of capital, like basically $150 million from day one.

Those are good friends.

To, to say, like, we believe in your understanding of, of how to build this. And, um, that wouldn't have been possible if I changed careers. And I also didn't want to. I think that this still hasn't been done. Um, it is important to me that it happens.

Did you know it was going to be called Science Corp? And how did you possibly get the name Science Corp?

Well, I mean, there's no trademark. Like, nobody gets to exclusively do science. Like, everybody should get to do science. Um, The, the .com is not available. It's owned by AAAS, the American Academy for Advancement of Science. They just redirected to the .org. Um, someday we'll get it. Um, the, uh, XYZ was kind of the, like, it kind of gets grouped with crypto.

So it's science.xyz. Yeah.

And people think that's like a crypto company or people call it Science XYZ. It's not science. Definitely not Science XYZ. Um, but that, I thought that that domain was going to go mainstream. I think it kind of has, um, it was good enough, it was available. Um, I'm not very good at naming things, and I just felt like the future needed a big multinational, like, biomedical technology company called Science.

Isn't there— there's like some VC in LA or something that was, uh, Science something, right?

There's an incubator, Science Inc. We're Science Corporation.

Okay.

Delaware had no problem with this.

They invested in like totally non-science. They invested in like, uh, what's that?

I think Dollar Shave Club.

Yeah, Dollar Shave Club. In that, that, uh, what's the water that everybody drinks? The, with the weird can, it looks like beer.

Um, yeah.

Liquid Death. Liquid Death.

Yeah.

I was like, this is not, uh, yeah.

I mean, look, there's plenty of science to go around. Everybody should get to do some science. Um, and I think it's basically like the branding for this is just pure memetic warfare. It's like, if we build things that matter to people that they like, then there will be a, the representation of their brain will split off a little bit and it will become less confusing. And if we don't, then then that won't, and that's— that we deserve it.

So when you start out, there's all kinds of things you could have done in BCI land, and, and you had this very focused first approach. So yeah, tell me the thinking, right, when you're starting the company of what you think makes the most sense.

Yeah. So we had— there were kind of two ideas at the beginning of the company. Well, there were a couple ideas. The I, so spring of '21, I didn't really have a plan. I got out of the last thing I was doing, started reading a bunch of papers. I actually would have been the first time that spring was like the first time in my life really that I like hadn't been working. And so I just had a couple months of like not doing anything, which was pretty weird and interesting. By the end of that, I definitely needed to get back to work. But I looked around, I really took it as an opportunity to reset and second guess what I was doing that was a big opportunity for a change. And I looked at AI, I thought about other things. I kind of made a very intentional decision, um, not to do those. Um, and part of that was I became really convinced that it would be possible to restore vision through the retina, um, in cases where the optic nerve was intact, the patient grew up with vision, that they'd lost it in adulthood, and they just lost the rods and cones in their eye. That seemed to me to be, with the technology that becoming available and that, that would be doable. We didn't know exactly the right approach. We didn't know whether this was like an optogenetic cell therapy or an electrical stimulator or something else. And we didn't know whether you should target the bipolar cells or the RGCs. But it seemed clear that that would be possible. And I think that that has basically been borne out. And then there was another idea that I had back in college. This is the biohybrid idea where— Which I love. The problem of like, if you place wires into the brain, there's no space there, you're going to displace some, you're going to damage some tissue. Um, uh, how do you get things into the brain? Like, what can you add? Like, what happens if you culture some neurons on the brain? The answer is they grow in and wire up and form new synapses, and you get tons of them. Um, that was also really interesting, but it was clear that that had to be paired with a nearer-term business. That was too much of a research project for too long to kind of stand alone.

So you're— I want to break this down for people who haven't heard of biohybrid technology at all or seen what you guys are doing. But you're— so, okay, we'll take a step back for a second. You kind of create neurons that are outside of the skull. I mean, you create these man-made, human-made—

Well, I mean, you start with— so you start with a skin punch from a patient. So you start with a skin punch, you turn those back into stem cells, and then we do a lot of editing on those stem cells. Cells. We add in, we try to hide them from the immune system. We add in a bunch of safety mechanisms. Now that you've hidden these cells from the immune system, you have to worry about that. We add a bunch of other stuff to them, make them light sensitive. We make them specific types of neurons. And then you've got this culture of cells. And what we do is we embed those in an electronic device in a dish, and then you engraft that into the brain. So there's no part of the device other than the biological axons and dendrites that grow into the brain. The device just sits on the surface, but these grow in and form millions of synapses.

Right. So you create this. Okay. You, when you talk about taking a skin cell, you're talking about using Yamanaka transcription factors to take it back. You take a skin cell, take it back to a stem cell, then you turn it into this, these types of neurons, this modified neuron. And then you're putting these in a, in a, in a gel, in a device. Make a little hole in the skull and then the neurons in that device are able to connect, send dendrites and axons into your brain. And so you're getting like extra horsepower from this device that's connected.

Well, yeah, I mean, if you don't just get extra horsepower for just adding the neurons, like the intuition here is not like we're adding neurons that are then making your brain bigger, but you're being able to access these representations. So the brain is an information processing organ. It has like, if you record from lots of different, have you heard of the Jennifer Aniston No. So there's a well-known result that there's some, in some part of the brain, if you record from these cells, if you show them pictures of different people, they in some cases will respond very selectively. And so there was a Nature paper a while ago that showed that if you recorded, if you showed pictures of different people, there was a cell that would fire just for pictures of Jennifer Aniston. Or if they showed you types of different houses, they would fire for just specific types of architecture.. And so the intuition here is that if you look at all of the millions of neurons in that area, they encode a representation of architecture or of faces or of people or of places. And this is embedded on a statistic, on a mathematical object called a manifold. And so what you want to do is you want to access that, like, that statistical representation somehow. And when these neurons grow into the brain and they wire up, you get all this— when the cells in the brain fire, you're getting all this partial information. Which represent information about the statistics of those representations. And so you can access this information in and out of the brain through this. This is a very, very deep well of neuroscience that we are at the very beginning of. Just making these devices at all is pretty tricky. I would say this is up there with modern semiconductors for complexity.

Because of all the steps of different things you have to do.

Yeah, I mean, making these cells at all is really annoying. Getting them to survive in to be healthy in the microenvironment of these devices is pretty tricky. Getting them to get through the hypoxic and glycemic injury of surgery is really tricky. Dealing with all of the electronics on these devices is all custom. Um, and a lot of this is like, for the cells, it's not— there's no like file on a computer that somebody can download and like make more of these cells. It's like these cells are an unbroken lineage of cell divisions through these edits for several years from back to the original donor skin cells to make them. And in a lot of cases, the way that you do these edits is you construct basically a test that you can evaluate what's the cell doing and you pick the ones that do the right thing. And we aren't, you don't have total detail on exactly what's going on inside the cell. You just can qualify that they're safe, that they have these properties, they behave this way and they can do the function.

But it's a lot of science.

There's, yeah. And ironically, we really try to do mostly engineering.

Yeah. Well, I was going to say, I mean, there's a lot of engineering, manufacturing.

We very selectively take on our science.

Yeah.

Yeah.

Okay. You, you always, I, we've had this conversation about this a couple of times and I always bring up that you get extra horsepower. I think maybe because I'm just simple and that, that's what resonates with me, but you always push back on it. And then just in your description, I'm not sure if the lay audience would get, um, exactly then what these extra neurons are supposed to be doing.

Have you seen James Cameron's Avatar movies?

No.

No. Okay.

Well, I know what they are.

I think that is actually a really useful analogy here. So in there, the aliens have these blue ponytails with these, like, fibers at the end. And this ponytail is— so all of the information that flows in or out of the brain goes through a pretty small number of wires. There's 12 cranial nerves. The optic nerve is cranial nerve 2. The vestibulocochlear nerve that carries hearing and balance is cranial nerve 8. And then there's about 30 spinal nerves that go down and come out at the vertebrae that get, carry motor commands up to the muscles and sensory information up to the brain. And so the brain, all of the information that flows in or out of the brain goes through one of the 12 cranial nerves or one of the 31 spinal nerves. So the ponytails in Avatar are basically a big externalized cranial nerve with like a connector at the end that you can plug into, in the movie, other brains. And I think that is actually like, imagine if that was a tech product. I think is the simplest way to conceptualize a biohybrid neural interface.

Even this part about like sort of connecting to another person, or you're just connecting to you?

Actually, yeah. I mean, and there's like existence proofs of this. So there's a pair of twins that are conjoined in the brain, Tatiana and Krista Hogan in Canada. I don't, I've never met them, but I've read the case studies. And they, so it's one, it's one skull with four hemispheres linked at the thalamus. And we know that this must be possible. And that, you know, that they share significant elements of their conscious experience. Um, but, um, this takes you to all kinds of really interesting places that I think get at fundamental questions.

And like, one of them was because these neurons a little more accessible because they're external from the brain and you're manufacturing them that I'm sure, please correct me as I butcher this, but, you know, for certain sorts of ideas where we know the shape of those ideas, the neuronal patterns that you could put You could put ideas into, into these, like, sort of like, uh, Inception, sort of like downloading kung fu.

We're like super early in this research.

Yeah.

Um, so the making the devices, what we know right now is that you can culture the cells on these devices. They survive, they integrate, you get functional synapses into the brain. That, like, to be clear, that's like where we are. Exactly how you use this and what it feels like and the capabilities, that is at early stages. And, but I mean, if you think about this from like a representational perspective, like the brain has these representations that correspond to like visual processing and language and monologue. I mean, it's weird and interesting that you see the representations that you see in these big transformers, these big AI models look a lot like the surfaces that you see in the brain. And so—

In what sense?

Like the way that they compress information often looks like they find similar statistical structure in many cases. And so there's a question of if you're thinking about using this as like a brain-to-brain technology, you could go between human brains. You could also potentially go between a human brain and an AI model. And it's a very profound sense where this begins to feel like one brain.

And you, I know you said it's early days, but I mean, this is stuff you think about all the time.

Yeah, I mean, I think there's a misconception that like, I don't like, we get a lot of people applying to work here and at the place I was before who are really motivated. Like, I don't spend a lot of my time sitting around thinking about the future of humanity. You spend a lot of our time debugging Linux drivers and like, um, wondering why the cells are contaminated. But, um, yeah, I mean, it's definitely undeniably interesting neuroscience to do.

Well, I, I ask because you said you started the company with these two, two places you wanted to go, you—

Yeah, there's a third, there's actually a smaller third program which has been there from the beginning, but is also much smaller and is a little bit different. But yeah, go ahead.

Well, but you know, two pretty large, massive efforts, one to restore vision and the other we're just talking about, but, but I mean, it's important to you. This is why you kind of started the company was to get to this place, right?

Yeah. And I think that also there's a misconception that these are really very different. They build on, I think, I think one of the other underrated effects here is a lot of the medical device industry and biotech generally, like 15 years ago, all of the drug companies were moved to being virtual biotech. So it's a very small core office. All the research was done through contract research organizations. In medical devices, you almost never see multi-product companies. They're almost always these single-purpose, single-asset companies, an asset licensed out of a university. There's a playbook. You get it through market approval, you go public, you get acquired, and you end up with these big marketing players like Medtronic and Boston Scientific, whose real business is actually sales and distribution and support. And the— this makes it actually really difficult to, to innovate. So it's most of these companies, you'll get a version 1 to market or maybe a version 2, but it usually things get really interesting and exciting at versions 4, 5, and 6. But to do that, you need to be able to iterate. You just can't iterate through CROs, through contract research organizations. If every time you need to do an animal study, you're calling a contract research organization and putting together a, like, $700,000 contract that's going to last for 8 to 12 months, or every time you want to make a new version of your, of your cell line, you're calling a contract manufacturer and sending them $2 million. You're thinking in, like, these multi-year cycles where it's just very hard to move or learn or iterate anything. But you also, get to incrementally use this very expensive infrastructure. This type of work is really limited by infrastructure, much more so than genius scientific insights. And so I, from my prior boss, fully received the gospel of vertical integration. And that is absolutely essential. Being able to go from a design change to surgery in a couple of weeks, or being able to, um, make, uh, like we routinely void the warranty on million-dollar fab tools because you get to a point where like, I would like to put these atoms in these particular locations. And if you call up a commercial fab and say, hey, can you like put this gas and this material in your sputterer? They'll be like, who let you in the building? How did you get here? And so that is absolutely essential. But, um, the, that infrastructure is very expensive. It can consume it. It's like carrying that cost is like hugely capital intensive. And so a couple ideas. One, our foundry, uh, we sell commercial wafer services to others to defray part of the cost of ownership of that. Um, that is complicated. Business that is like, that comes with a bunch of other pros and cons, but we hope allows us to get more scale over time than we'd be able to otherwise. Um, and the incremental cost of these programs is actually quite low relative to each other. If we only had one or the other, you could make either one look more expensive, but it's actually building on the same underlying platforms that we need anyway. If we want to go from, uh, where actually go to, to things that work work really well.

Okay. And so there's this interlinking between these different technologies and the substrates of things that you need, but you, you did. So, but you saw that you needed to make some money and the retinal technology was that, that was kind of the first clearest path for you on, on getting the business going.

Yeah. But I mean, I think it's a little too dismissive just to say we need to make some money. I think the question is like, what was the most value that we could create with the state of the field and our resources, including our team as a resource and what we knew and our differentiated, I think, perspective on the world. And the revenue is downstream from the value. And I don't at all want to— like, I think one of the things that becomes an issue when we talk about the different— our interests, which are not diversification, we think that each of these are really important, um, is it sells short the fact that AMD patients are reading. Like, this is an end in itself. If that was the only thing that Science, the company, did, we would be very proud of that. And so I think that there was an opportunity. We saw that— I mean, I grew up around blindness. My mom's dad had retinitis pigmentosa. When I told her that this was like going to be the first major focus of the company, she was like pretty, pretty excited. Um, and so that we thought what we needed to do was create value for the world. And I think that this restoration of sight is like an incredible opportunity on that. The revenue follows, but with the value at the heart of that, that would, we believe, would give us a platform to go on to continue to innovate and do these other things through the takeoff era that we thought would not— like we thought would be really important and have an impact on the world.

You got more offended for being called a capitalist than I thought you were going to, man.

You have to embrace capitalism. That's like a key Um, that, that's why we get to do this. Don't, like, for step one, don't break capitalism. But, um, but that's, I mean, there's lots of, I don't know, I talk to companies in, uh, the BCI space all the time. I was just having a meeting with one the other day and they, like, we, I think we think, I think we feel strongly that these things are used for good. And, uh, I don't know, there's a lot of companies out there that are thinking like, oh, we're going to get all this recording data. How do we monetize it? Do we have subscriptions? I'm like, this Why are you gathering this at all? Like, you don't need any of this. Like, this should just like run it all locally. It should belong to the patient. And so I think that there's nuance there.

For the, for the layperson, can you explain how the retinal technology works and where we are with your trials and where we will get to with other people being able to get this technology?

Yeah. So the, we have, I think we have two retinal technologies. The one that is for the Sloan that is coming to market next year, hopefully, it's submitting for approval very, very soon.

In the US?

Uh, in Europe also, also hopefully the US. We're waiting on a response from FDA on a question that we sent them recently. Um, the, uh, so the, the product is called Prima. It's a little chip that's inserted in the back of the eye. It's an hour outpatient procedure. Um, just insert it to the front of the eye. You leave it under the retina and the chip has all these little, if you see a picture of it, it has all these little hex grids and each one of these hex grids is a little solar cell. Basically. So the patient wears glasses that have a laser projector and a camera. The camera's looking out at the world, it converts it, it simplifies it, converts it to an infrared image, and then it uses the laser projector to project this infrared, um, image onto the implant. And wherever the implant absorbs infrared light, it stimulates. And so this is actually a really elegant idea because that's the whole implant. There's no wires, there's no battery, there's no cable coming out of the eye. It's, you send it energy and information simultaneously in the laser projection. And so that there's something really elegant about that. And we didn't invent that here. We, um, the original inventor of this technology is a professor at Stanford, Daniel Planck, and the electrical engineering department. And he's, uh, he, he's great, super smart guy. Um, and we're fortunate to be able to, to own and commercialize that, that idea now. Um, and so the, the current device has 100 micron pixels. So each time, so it stimulates that first layer of cells in the eye, the bipolar cells. And each pixel simulates a bunch of bipolar cells. So the next step is to shrink those to make it smaller and smaller until you get towards native acuity, like the native resolution of the eye, which we think we can get to in a couple device generations. Right now it's just black and white because when you stimulate all of these cells, you average their representations together and you, one of the things you lose in that is color. Like that's, there's no such thing as a white photon.

Photon.

Okay. If you look on the color spectrum, there's no white. White is when your brain loses spatial temporal resolution to distinguish what colors the photons actually were. And so when you stimulate a bunch of bipolar cells, this averages them together and you lose the color resolution. So it shows up as white or black. But we think that in the next device generation, we might be able to get red and green. Blue is a little bit trickier, but there's a path there.

What's the difference between what somebody— obviously they can't see. What can they make out? Like, letters on a page.

Yeah, I mean, there's videos online from the trial. There's— patients can intuitively— some patients can intuitively read. So I think it's important to draw a distinction because people have been working on retinal prosthesis, visual prosthesis generally, whether into the retina or into cortex.

Yeah, like when I've seen stuff historically, it's kind of amazing because people can't see and then they can see these blurry shapes.

Well, but yeah, I think it's important to really distinguish what they're talking about because there's— when we talk about form vision, this is vision that is apprehensible to the brain as a coherent whole, like a letter at a time or a word at a time, um, that, that is distinct from these things called phosphines. And all the earlier devices, they would get— like, if you put an electrode array on the optic nerve or in cortex and you stimulate it, you'll get these flashes of light. Those are phosphines. But they, like, you can't look to them. They don't add up to a whole, like, you'll see these artistic representations of these flashes that form a pattern. But when you actually have a patient, we know this because 10 years ago there was a company called Second Sight that had a stimulator that activated the optic nerve that got these phosphine fields. And it was, patients could like kind of read in the sense that they could like look at letters and kind of make out letters. But it was, it wasn't intuitive. Like the brain didn't assemble it into a whole. You'd have to see like, oh, there's like a line and it's like, it's an E. B, then it's like, I see, oh, it's an A, and you'd go that way. Whereas the results from the Prima trial are like, they're reading, they're reading off words.

But like, if, if I had these glasses on and I was sitting here, what would I see versus what I— so this isn't for people who are completely blind, it's kind of degraded condition. Yeah.

So the, the trial was done in age-related macular degeneration. Okay. And so they have some residual peripheral vision. They've lost the— it's only like the central 6 or 7 degrees of the field of your vision are high-resolution color vision. This is a part of the retina called the fovea, and then the macula around the fovea. And so these patients, they can walk around, they can generally kind of navigate the environment because they've got this peripheral vision, but they can't recognize faces, they can't read, they can't— it's really very debilitating to lose your central vision. You don't notice it because your eyes are constantly darting around. These are called saccades, saccadic eye movement. And so even though only the central few degrees of your visual field are high resolution and in color, your eyes are constantly filling in information. And your brain, more fundamentally, has a model of the world which is being updated by this weak sensory input. This is one of the reasons that, like, optical illusions can be so strong, because it's not really— your eyes are not a camera. They're an information channel to the world model in your brain. And that that gets these updates. And so actually in the beginning, like when you lose sight or lose hearing for that matter, your brain tries to turn up the gain to say like, I'm trying to looking for something. And so hallucinations are very common in blind patients. And so the first couple hours with the device are just learning to separate the phantom percepts from the real percepts and distinguish what's real coming in from the optic nerve again. And then once that's broken and they get it, then they can, they can figure it out pretty quickly from there.. But yeah, but to your question, what does it look like? It— the— another, another task that's done, there's, there's videos online of patients filling in crossword puzzles. And this is— so most motor control is not visually guided. It's like you don't need to look at your fingers to type or to catch a baseball. Most motor control is proprioceptively guided. There's nerves in the muscles that give you feedback so that you know where your, where your limbs are when you're not looking at them. And depending on how you try to stimulate vision in the brain, you may or may not get for free this connection to motor control because your brain has to correlate the visual signal with the proprioceptive signal back together. And there's these videos of a patient filling in a crossword puzzle. And this is interesting and really not to be underrated because that is a visually guided fine motor task. They have to see the grid squares on the page. They need to see the letters.

I see.

They need to control the tip of the pen within it. And so that I think is is very strong evidence that it is, it shows up as vision for everything that vision is useful for.

Um, and, but like, and what about just looking at someone across from the distance that we would be at?

Um, so there's different modes of, in the software. Um, we actually were planning to do more of a study on facial recognition that wasn't actually carefully studied in this clinical trial. Um, but it's definitely a thing that we're going to be, uh, characterizing better.

And what, and, and I mean, where does this technology go in terms of, um, somebody who was born completely blind? I mean, it's not really meant for that.

Yeah, that's not meant for that.

Okay.

Um, that's at least, I'll just say a different problem.

Okay.

I think there, for a long time, there's a conventional wisdom that, um, if you didn't form this vision circuitry in childhood, it wasn't, you know, because you missed these critical periods in development. Um, and there are some case studies, um, from like like, I don't know, like decades ago, where patients would get, like, if their retina developed fine, but maybe they had a genetic cataract or something that would be removed, the patients would find it totally overwhelming and their brains were not at all prepared for it. And some of these patients committed suicide. But there's also, but I'm not sure that I find this, like, I think these anecdotes are not totally convincing. There was a case, I don't know, 10 years ago maybe, I don't remember exactly when, where there's a group called Project Prakash that went to, I think, rural India and treated this population of, like, teenagers with genetic— I think it was cataracts— and repaired them, which is a very easy surgical fix. And then these teenagers got sight for really the first time, and within a couple weeks, many of them could catch a baseball. And so I think that the brain stays plastic really throughout life more than is commonly appreciated. The critical periods are definitely important, But anyway, that's like, that's loss of no vision from childhood is just very different in the impacts on the brain.

It was maybe like too specific of a question. I was just wondering if, if this is step one of curing a wider variety of visual issues or how far this technology can stretch.

Yeah, I mean, so there's many different types of reasons that people go blind. If you have like cataracts are easily fixed, not a problem. Corneal abrasions also, I mean, you need donor tissue there, but can be repaired. The— so we're focused on diseases where the rods and cones have degenerated, but they grew up with vision, lost it in, in later life, and the optic nerve importantly is still there. When you lose the optic nerve, that is more difficult. The most common cause for that is glaucoma, which is high pressure in the eye, which squeezes the blood supply to the optic nerve, and then you lose that. If a patient has glaucoma or has otherwise lost the optic nerve, Primo would not work for them. Them. But there's a lot of reasons why people lose the rods and cones beyond macular degeneration. There's also retinitis pigmentosa. There's Stargardt's disease, which is kind of a young person's age-related macular degeneration. It's genetically driven. There's diabetic retinopathy. There's a lot of reasons that people can lose rods. There's all kinds of idiopathic photoreceptor losses. There's not as much— so people Neural, we've obviously had cochlear implants for 50 years. Sensory neural hearing loss is much more common than rod-cone degenerations. A common, a major reason that people have sensory neural hearing loss is an autoimmune reaction. So your immune system attacks those cells and then causes them to die. There's just, my intuition here is that there's just clearly a much weaker evolutionary pressure on protecting those. And sometimes people talk about the eye being immune privileged, they're— it's not like there's definitely immune cells in the eye, but they tend to hang out on the periphery. I think evolution learned a long time ago that if you overreact in the eye and kill the photoreceptors, this is really bad and you must not do that, which is one of the reasons the eye ends up immune privileged. And you also don't have as much spontaneous loss of these cells than you do in, for example, the ear. But there's— but still it happens every now and then. There's viral cases sometimes. And so really for any of those, like if you can bypass those dead cells and just get the visual signal back into the brain as early as possible, we hope to provide mechanisms for doing that that get better and better over time. I do think that it, down the road, a biohybrid type device could work for glaucoma. People have done, not us, but like the academic community has done experiments where they've engrafted cells into the eye and they've seen that they've pathfound down along the remaining optic nerve tract into the thalamus. And that is really intriguing, but it's also a big science project. And so is for the future.

Okay. Which, okay, so if you look at something like what you guys are doing there, and then I look at the BCI field more broadly, you know, this technology is first hitting with people in quite dire straits. They have vision problems, they have ALS, they're paralyzed, or they're I mean, well, so I mean, my big question is, is, you know, we hear all the time of where this technology might be going. I mean, what is your— what is— what is your best guess on how long we're in this phase where BCI-type technology is being used really as a medical treatment versus when it gets to something closer to like an elective kind of consumer thing that people are choosing to take advantage of.

Yeah, I mean, I— my view on this is that everybody eventually becomes a patient. I think that this body is great for a long time, but eventually fails. And so I think we take a perspective here of it's like we're going to restore lost functionality, eventually extend. But I think what you're going to see first are people that become patients anyway, get capabilities that are differentiated.

Or that they lost or—

Yeah, definitely restore things that they lost and eventually go further. But I think, I mean, I don't know. It's just so hard to bet against tech progress and technology right now. And I just, two days ago, we had a meeting where a thing that I thought was going to take years, it might, is like, might happen much faster now. And I, they just— the timelines keep compressing. And so I don't know, like, who—

I like that.

Like I said, I've got this event horizon at 2035.

Right. Okay. But the end goal of your company is what? It's to help people in need.

Definitely.

But is it also to give special powers to the rest of us at some point?

Well, I mean, so part of— I think the reason that I'm having trouble with this question is that BCI is not a product. BCI is a field in the same way that like social media is a field. But if you're just like, I'm going to build a social media company, you're like, what does that mean? Like, what is the interaction here that is compelling that people are into? Is it photo sharing? Is it 140-character like updates? Is it family? Like, you got to be more specific and you're going to have many different products within that that are all different ecological niches that are compelling. And so I think when you look at motor BCI, like BCI has become kind of synonymous with motor decoding, I think in many people's minds. I think that generalizes into communication.

Yeah. Like, okay, well, yeah, I mean, okay, tell me next 20, 30 years, sort of like a mind meld, I can express a thought directly to you.

Yeah, probably. Yeah. I think like direct, I mean, so first of all, language is pretty good at this. Like evolution did a pretty good job of this.

Sure. But yes, but I've like, when I am experience when I'm at a concert and I feel this like sort of joy during the best moment of the concert. Yes, I could like describe that in words. It would be like a little slow really for what I'm actually feeling, but I, I am definitely not conveying.

But this is why you have the other like 10 cranial nerves of like your face and your expressions and things. But, but yes, I think that on like 20 years for this technology, I would say is for many of the things that you could imagine is a sufficiently long time horizon. Yeah.

Even words are slow. If I wanted to convey like a paragraph to you in an instant, I mean, I guess that's kind of what I'm— almost like an LLM. Like, am I sending a block of tokens, you know, from my brain to your brain?

Yeah. So generate production of language. I think that there's a myth that we could just think faster. Perception is a lot faster than production. So getting information into the brain can be done at much higher bandwidth. It's like, like you can get vision in and you can see an image, you can read much faster than you can produce the tokens. So I'm not sure that it's just like talking faster, but I think that it's— so all communication is about creating correlations between brains. We can have this conversation because we both speak English. We picked this up during childhood. We can talk about ideas about brains because we both had an education about what our brains— if there's a concept that we don't both share, then we can't convey it to each other. I think the same thing is going to be true over these BCIs. It's no matter what the link is you need the concept spaces present on both sides in order to light them up. Now, where you get into more interesting territory is how do you add, like, yeah, how do you add a new cortical area? How do you add representations to the brain that you didn't have before so that you can communicate these things very efficiently? Or how do you decode into a pre-trained model that allows you to get out more complex— you can get information into the brain very efficiently. Getting information out is limited to muscle control and language. You can't get vision out of the brain. You can't get hearing out of the brain. And so that type of stuff I think might be possible.

Okay. Um, I want to be mindful of your time. I, as we were setting up for this, you're like, oh, there's all this crazy stuff I wish I could tell you, but I, but I can't. Well, so I, I'm probably not going to get those details out of you, but what does that even mean? That just means you're seeing things out in the world that are fascinating.

And yeah, I mean, I don't mean that in like, I mean, I wouldn't over, I wouldn't overthink that. It's just, there's like a lot of cool stuff and I think it's also, um, yeah, I mean, I think that it's the most exciting time in human history and I think that that will like continue to be true. Um, the world is also, um, I think that as these things accelerate, like there'll definitely be some instability. Ability. But I don't know, it's— there's the only way out is through.

So there's the stuff you're working on, but then you're hearing from your friends and you're just hearing that things are clicking and interesting.

Yeah, I mean, I'd say Silicon Valley is buzzing. Like there's, there's just so much. I don't know, like, yeah, 2 days ago, I think Celine's company Loyal announced that they got through a update, a major milestone with the FDA for dog longevity.

Yeah.

Like all, like I just refer, I'm not speaking specifically. There's no like one thing that I'm like, oh man, I wish I could say. It's just like, there's this sense that like a lot of cool stuff is happening and we're going to get a lot of it.

Okay. It seems like a good place to stop for the optimists. Um, thank you so much for your time. Thanks for letting us come to your company. And, uh, yeah, thanks.

Yeah.

Cool. Thanks. Good to see you. Core Memory is hosted by me, Ashlee Vance. Today's guest was Max Hodak. The show is produced by me and David Nicholson. Our theme song is by James Mercer and Jon Sortland, and the show is edited by the tremendous Jon Sortland.