JuanBenet

LubaI'm so excited to have Juan Benet with me today. He is the founder of Protocol Labs, and we had such a rich conversation about the future, neurotech, his path into crypto, into thinking about knowledge systems, and how to document the knowledge that we have in the world to impact scientific progress for the betterment of humanity. Juan, thank you so much for being here. Welcome to The Luba Show. I'm so excited to have you.Link

JuanThank you for having me.Link

LubaI want to spend a lot more time talking about what are the problems that you're focused on right now. But I actually want to start more at the beginning. Tell me more about what you were like as a kid. I've heard that you read, or inhaled rather, encyclopedias, and that led you to your interest in knowledge expansion and human knowledge. Where did all of those interests come from?Link

JuanSo, yeah, totally right. When I grew up, I would read encyclopedias for fun. As a kid, I learned to read fairly early. I think at, like, three or so, I started being able to, like, sound out letters and start reading and so on. So I grew up with older siblings, and they were already learning through school and all that kind of stuff. And so I, as the baby, kind of saw them grow up and them learning a lot. And so I kind of wanted to be part of that, and wanted to understand what was going on. And so I was able to fairly quickly become super interested in books and so on. And when I realized there were entire stories written on paper, I really wanted to access that.Link

And so as soon as I started being able to read, I started reading various things. And, you know, first started with, like, literature, fiction for kids type stuff, and then pretty quickly found these illustrated encyclopedias that were incredibly cool, where, you know, they had just a bunch of really cool stuff to learn about, but it was all illustrated to give you a perspective. And then from those, I jumped straight into pure-text encyclopedias, where you think of, like, the Britannica and stuff like that. And I remember just reading. I just wanted to learn what everything was. And I was just fascinated by how much stuff there was to think about and learn about. And so I remember going and reading various articles and just trying to make it through the encyclopedia. And it just gave me the sense that there was so much that you could find out and discover about the world. And thankfully, a bunch of other people had already done that and written it up for you and packaged it up in a really easy, fast way to consume it. That, for me, was fascinating.Link

And the next jumping-off point was probably being able to access the internet. And I got onto the internet early, probably eight.Link

LubaHow old were you?Link

JuanI don't remember exactly, probably eight, nine, ten, maybe. I got to the internet through video games. I used to play a lot of video games, both consoles and PC games and whatnot. And so I used to play a fair amount of online games, and through that, I started exploring the internet. And I remember the early Netscape browser. This must have been, like, '95 through '97 or so. So I remember at least some version of Netscape, and, like, clicking on a page and waiting for it to load, and the little Netscape icon was animated, and then a new page.Link

LubaWas that on a phone line?Link

JuanYeah. And so it was fascinating for me, because I remember going on, I think it was AltaVista at the time. I would search for stuff, and then I would see the number of pages of results. And that was both super exciting, of, like, all the stuff that is now at your fingertips, but also daunting. Because before that, whenever I had found a room full of books, I remember doing a calculation of, like, "Oh yeah, you know, I read this fast, I should be able to read every book in this room eventually, if I wanted to." And the moment that I saw the volume of information on the internet, it was this crushing moment of, like, there's just no way that I could possibly ever read all of this. I was like, "Oh no. What am I gonna do?"Link

LubaBut why was it so important to be able to grasp all the information in the world?Link

JuanI think it's because, I guess I became very keenly aware that knowing about something gives you, like, all kinds of things are super interesting to know about. And I'm just a very naturally curious person. I'm interested in various phenomena and in understanding how they work. But also I noticed how the more you learn about something, the more you benefit. Like, to me, the Francis Bacon quote of, you know, "knowledge is power" is tremendously correct. Power is probably not the right word, because it tends to cast it in this, you know, power-differential thing. It's more like knowledge is capability. If you know about something, you extend your capabilities of being able to do more things. And that's not about power structures. It's just entirely about what you are able to do. Think about it as, like, hey, if you don't know how to swim, then you can drown. But if you know how to swim, then you're good. Or think of humanity before knowing about bicycles or whatever. Or not just bicycles, the wheel in general, just being very limited. And the moment you learn to do this additional thing, you expand your capabilities.Link

And so that was a very formative thing for me, and that translated later on into all of this thinking around knowledge and knowledge systems, and what knowledge actually is. And so one piece is information. I think the last 50 years or so have been described as the Information Age, of being able to digitize everything and distribute it through the internet and all that kind of stuff. But I think it's not just about information and giving access to information. There's a separate layer of this stuff, which I would term actual knowledge, which is, you know, information that represents some model of reality that is either correct or not correct. And you have to test it to find out whether things are incorrect, and, by process of elimination, find out what is more likely to be correct. So this is the entire scientific process. And here I would maybe credit David Deutsch and Karl Popper for pioneering the thinking about knowledge systems, and how do we as a species get to know what we know.Link

And so when I think about the scientific process and the technology-building process, these are two sides of the same coin. The scientific process is how we as a species get to know what we know. Like, how do we expand the tree of knowledge? How do we build up more and more of our understanding of reality? And then the technology-building process is, how do we translate those learnings into concrete capabilities? Whether that's, you know, software technology, meaning different cultural practices, or different ways of doing things, different methods, different processes. Or actual physical technology, where you have some device that you're making, whether it's a lever or a plow or a wheel or a computer or a telescope or what have you. And each of those technological components, again, whether it's hardware or software, expands our capability set, which then enables us to do more science, to discover even more. And this process just feeds back on itself. And that's kind of the story of our species. We've been able to do that first in small groups, and then on larger and larger scales, to flourish broadly across the planet.Link

LubaYeah, I love the fact that you're tying all this knowledge to actual capabilities and to doing things in the world. Because, you know, there is a lot of information flow, and it's very easy to just know a lot of stuff and not do anything with it at all. I'm curious, when you were young and you wanted to know all the things, to have different capabilities, what was the dream? What would you want to achieve with those capabilities? And how has that changed?Link

JuanYeah, I think, I don't know. I kind of, like, wanted to learn about everything and try everything. I was just very naturally curious, and, you know, I would simultaneously be super interested in, like, I don't know, how lasers work, or how computers work, or how musical instruments work, and how to play a piece of music, or how paintings are made, and, like, you know, what distinguishes the great works of art from others that aren't. And so I just had this very natural curiosity about everything, and I just followed more what I was in the mood for that day. And I just had a very wide exploration.Link

I also was fascinated by ancient cultures and what they knew at the time. And I just remember being super struck by the fact that, you know, when you look at the timelines and you see how many thousands of years some of these cultures lived for, and what progress meant for them at the time, what they got to understand and be able to do and be able to predict. Whether it's key patterns about the weather to be able to do better agriculture, or being able to do basic irrigation systems to be able to grow more and more crops, or, like, granary systems to be able to keep their grain and whatnot.Link

LubaWhat's the culture that you're most fascinated with?Link

JuanSo, fascinated with many, probably. I went through a period in elementary school where I was super fascinated with, like, the ancient Egyptian culture, and it was just super cool.Link

LubaYeah, I think it's insane how much of it is lost. There are these different methods and ways of living. And just even thinking about how, you know, we have this mindset of, like, nine to five, seven-day week, or, like, five-day work week. Obviously, that came with industrialization and whatnot. But I bet so many things were different in, say, Egyptian culture: how they treated the calendar, how they, I don't know, looked at different rituals in life and their representation.Link

JuanYeah, so much of our daily life is shaped tremendously by base-layer cultural processes that we put in place. Like, you know, how long is the day? How long is the week? You know, how many days per week, how many hours per day, the weekend, holidays. Some things are, you know, derived from other phenomena, like the seasons. The seasons very much impacted when you had to, you know, plant stuff versus harvest and whatnot, and that shaped a lot of calendars. But so many other processes are totally arbitrary, where they could have gone in a totally different direction.Link

Fascinatingly, in the French Revolution period, they tried to kind of, so our current calendaring and time system comes from the ancient Sumerians. And, you know, the number six was a big deal, and the number 60 was a big deal.Link

LubaWhy was the number six a big deal?Link

JuanI don't remember the reason why six and 60 were important. There must have been a reason. I think it's a different culture that did, like, the number 12 because of the 12 knuckles, because you could count up to 12. And our base-10 system just comes from the fact that we have 10 fingers. I mean, it's easy to count to 10 that way. I don't remember why the Sumerians cared about six and 60, but we ended up with this calendar based on that. And so, you know, 60 seconds to a minute, 60 minutes to an hour, and then 24 hours. And it's just kind of clunky, doing the math on it.Link

LubaIt feels like using the British system.Link

JuanYeah, yes. It's the same as feet, yeah. And so then during the French Revolution, they instituted the metric system, and that one stuck. But they also tried to do decimal time, where you had, I think, 10 hours in a day, and then, you know, each hour could be divided into, like, yeah.Link

LubaIt's fascinating to think how those decisions affect so many different things in life, and how the decisions that we're living upon right now are affecting how we treat our days and weeks and whatever. Based on your curiosities about different ancient civilizations, and these different, maybe minute, decisions at the time and their impact on the world, what did you think the ideal way of living and the ideal world creation would look like?Link

JuanI think that probably, for me, got shaped way later. Just going from what we've done in the past to what we ought to do in the future, that's a large leap. David Hume has this perspective, like, you can't get an ought from an is, meaning you can't look at what has been happening and then from that derive exactly what you ought to do. I don't exactly agree with Hume here. I would side more with Sam Harris, that our scientific understanding of ourselves, of our species, of other species, should give us a base-layer sense of morality, of what is correct to do. Like, I am not a moral relativist that says that everything can, you know, according to your frame of reference, be justified.Link

I think at the end of the day, we're, you know, a biological species. We're in connection with the rest of our ecosystems. We have a tremendous impact on the lives of other beings. You have a wide spectrum of benefit and suffering, and our actions can cause either more benefit or more suffering in ourselves and our species and other species. And so I think you can indeed get an ought from an is, but it requires a lot more. To really do that, you need a depth of understanding of how all of the systems work, and what is the nature of these systems, and what are they trying to do and trying to achieve, and why.Link

And you end up with very difficult moral landscapes. Like, you know, a nature photographer, for example, documenting the lives of, say, a lion hunting in the savanna. They go and hunt a gazelle or whatever, and that's, you know, terrible for the gazelle. You definitely don't want to be the gazelle being hunted by a lion. At the same time, you know, there's been millions of years of evolution that have shaped the lives of these different animals in some kind of biological equilibrium and balance. And, you know, the lion also needs to be able to eat in order to survive. And so the morality there is hard. On the one hand, you want to potentially save the gazelle. On the other hand, you want to save the lions. And it's a hard decision-making thing. And the principle that we've arrived at is not intervening in that situation, of letting nature take its course.Link

However, a similar-sounding but totally different phenomenon is one where you end up in situations of being able to have medical interventions for a disease, right? So the natural state of affairs is that some virus or some bacteria, or some other kind of pathology, or accident, or whatever, is leading people to have some bad medical condition. Or it could be a person or an animal having a bad medical condition. And we have the capabilities today to be able to greatly improve the outcomes for those people by giving them some medicine or some treatment that helps cure the disease. And there we are stepping in against the biological equilibrium that was achieved over millions of years. But we think that that's the right moral thing to do, because, you know, dying is bad, suffering is bad, and maybe we can alleviate a lot of that suffering and a lot of that death by intervening.Link

And certainly our species today is just dramatically different than where we were 1,000 years ago, or 10,000 years ago, or 100,000 years ago. And so we've been beneficiaries of this just tremendous progress. And our moral landscape needs to understand that and understand that role. And at the same time, we need to grow up, in a sense, and understand the level of impact that we're having, not just on ourselves, but on the rest of the species on this planet. And we have to take our great power, you know, the Spider-Man line, "with great power comes great responsibility." We have to make sure that we are good stewards of the planet and all the other lives in it.Link

And that doesn't mean, to be clear, I'm a very pro-technological-progress techno-optimist. So I very strongly agree with, you know, increasing our capabilities in terms of energy output and production and so on. But at the same time, we should do so cognizant of the level of well-being that we're causing for ourselves and the other species on the planet, right? Just because we have current mastery and control over nature today does not give us the right to treat other living beings as if they're worthless. Instead, we should be careful in how we treat many other species.Link

And I say this as somebody who, you know, I'm an omnivore. I eat both plants and animals, and these are other lives. But for example, I would strongly side with farming practices that lead to a much higher quality of life for the animals that are being bred for livestock, versus the extreme factory farming situation that probably has an enormous amount of suffering, right?Link

And so I take into account all of this thinking, and I come at it with a perspective of trying to deeply understand how we got where we are. What have been the hundreds of millions of years of evolution that got us here? When did our different lines of species break off? And what is the different degree of capability that all these different forms of life have for fulfillment and suffering, right? So for example, a mammal with an advanced nervous system is going to be able to access a type of suffering and a type of fulfillment that is just very different than what, say, a plant needs, right? A plant needs a very different set of conditions for well-being and balance with its ecosystem. And so there's just a totally different way that you should approach both of those. And the same goes for humans. The things that are going to lead to really great flourishing and well-being for humans globally just look qualitatively very different than what's going to be good for, say, a mouse in the savanna, or a plant in the rainforest, or something like that.Link

LubaSo do you think that people who are responsible for creating different types of technological and scientific and health solutions right now have good sets of principles for how to make those decisions, and to understand which position and perspective they should take?Link

JuanYeah, it's a great question. So I'll give you the quick answer, and then I'll give a longer answer. So the quick answer is yes. I think for the most part, most of the people that I've ever met and read about, read from, in a range of the scientific fields involved, or the conservation fields, tend to approach their thinking and their decisions with a super deep, well-considered perspective that is very informed by as wide a swath of information as they have access to.Link

That's not always been the case. And in fact, there are some very clear examples in history of people that abused their knowledge. Like, there's at least one case of one scientist who did this terrible stuff, where he's a person who tried to defend putting lead in gasoline and tried to defend CFCs, which then caused the hole in the atmosphere.Link

LubaLike, he obviously knew that it's bad, but he was still trying.Link

JuanTotally, totally right. Like, I can't claim to know what was in his head, but the actions speak very loudly, and it's pretty clear that I think he understood what he was doing. And so there's clearly some bad people that abuse power.Link

LubaSo it's kind of like, is your intention good, having the information that you have at the moment? Obviously, looking back at history, we might have more information and be like, well, that decision was bad, because now we know more. I don't know, like, opioids are actually bad, but in, like, the 20th century we were heavily using them for surgery. So.Link

JuanYeah, totally. So there's mistakes of lack of understanding, which are more understandable. It doesn't mean it's good. It's just that that's a very different level of harm than if somebody is knowingly doing something really bad, right? Those should be treated very differently, and they are. In our society, we treat those two cases quite differently. But there's also something that I think is very pernicious today, which is, we live in a world where the science tends to be broadly misunderstood and misappreciated, and just.Link

LubaHow did it happen?Link

JuanThis is a hard social question. I think it happened through the undermining of authority figures in media. So if you wind the clock back to, say, the '60s and '70s, there was a lot more, actually '50s and '60s, there was a lot more respect for authority figures. And as somebody who strongly believes in, you know, scientific principles of, you know, take no one's word for it, try to figure it out for yourself, and strongly autodidact and so on, I strongly agree with not just trusting authority figures for authority's sake, but instead deeply understanding what they're saying and trying to connect that to what you know, and so on.Link

However, you know, in the '70s, '80s, '90s and beyond, there was just this progressive erosion of trust that happened, where now there's this suspicion towards authority figures of any sort. And not from a perspective of well-considered alternatives, but rather from, honestly, an ignorant perspective, where, like, people just don't know any better, and distrust people that do know better. And that's just a terrible position to be in.Link

I think that's changing. I do think that the advent of social media, which in some respects gives credence to a lot of people who are able to just broadly broadcast stuff, is also leading to a lot more knowledge being disseminated and a lot more people leveling up their understanding. A lot of people from a more academic orientation, or from these authority-oriented landscapes, worry about this kind of stuff, because, at least in the past, it used to work that you would have to develop deep expertise in something before you're able to speak to it. And instead, now, because of the internet and social media, anyone can say anything about anything.Link

LubaAnd then it's about discernment. Can you actually tell?Link

JuanExactly. And I think the key is that, in higher and higher numbers, every year, more and more people are developing that level of discernment and ability to understand what is correct and not correct. We're not yet doing great. I think it could be dramatically better, but it is much better than I think a lot of people give it credit for. And you can see it in the views for really deep topics. So you can go to, you know, channels like 3Blue1Brown, which is an amazing math explainer, and it has millions of views for really deep, niche mathematical topics, which shows you there's now millions of people learning that material. And so that is just way stronger than it was, you know, 10 years ago or 20 years ago.Link

And so I do think that the internet making all this information available has enabled people globally to level up their understanding and level up their discernment. That's not to say that some bad memes and bad perspectives aren't also getting propagated and increasing in numbers, right? Like the Flat Earther thing grew thanks to the internet. But at the same time, I think many more people globally got to understand why we know that the planet is round. What is the background behind that? What are the experiments that you can run to figure this out for yourself, and so on? And so I think, in general, we have this net positive, even though the internet has been used to propagate bad memes too.Link

LubaSo you've evolved your own work to now spend time exactly on this: promoting science, making science cool again, funding different projects that otherwise wouldn't be able to be funded with traditional capital. I want to split this into two. One, understanding how you've evolved your career from working just in crypto, on Filecoin, and then into Protocol Labs. And two, you know, what are the main problems that you're focused on right now?Link

JuanYeah. And so maybe I would start a little bit earlier. This deep interest in knowledge, and this quest for knowledge, led me down a path of trying to understand concretely what knowledge actually is.Link

LubaWhat is knowledge?Link

JuanWhich, you know, surprisingly, you would think it's an easy question. It's surprisingly difficult. I'll give you one framing, and I'll punt the rest of the explanation to David Deutsch, specifically, because I think he does a phenomenal job in a book called...Link

LubaWhat's your favorite book?Link

Juan*The Beginning of Infinity*, I think, is extremely good. And I think *The Fabric of Reality* also does a great job. But I think both of those together give you a very accessible, well-popularized description of what knowledge is, with good grounding from many different perspectives in scientific understanding, and in particular a computer-science-oriented perspective, which I think is a key piece here.Link

So maybe I'll try this explanation. I don't know if it's gonna work. But the way I see it, there have been three major knowledge systems developed on our planet so far. The first major knowledge system is the genetic knowledge system. So this is very low-level biological structures that figured out a way to encode information about themselves in a strand of information. So this is genetic sequences, like DNA and RNA, coding for proteins and for assemblages of proteins to produce a cell. And the evolutionary process yielded this pressure for these systems to refine themselves over time and become more and more accurate representations of reality. Meaning, there was this deep selection pressure for representations that correctly coded for correct proteins, but not just correct proteins, protein assemblages and sets of them that would perform better than others. And so you had this very strong process for refining information strands that represented pieces of biological material that had to do well in the environment.Link

So that means that those genetic sequences learned to encode some information about their environment, and what is successful in their environment. And that's a form of knowledge. It's like a program that is run. It's literally, from a computer science perspective, a set of instructions that get computed on by a little computer, which is like the ribosome, and the ability to assemble proteins and whatnot. And they form part of this larger system that then performs better if it is more accurately representing the world around itself, and performs worse if it makes mistakes. And so that is the first knowledge system: being able to encode this kind of information about the environment in itself.Link

LubaSo just to translate it, so that I make sure I understand it: it's basically, the world started with some inputs of different sequences of protein and their quantity, and then it would go out into the world and essentially run, like, an A/B test for how that system is performing against the environment. And if something was off, then it would readjust itself. So it's like a self-readjusting system that would eventually create evolution.Link

JuanWell, so I would say even the world didn't start with that, because those components got assembled out of this selection pressure. So there's a process that happened before. We still don't fully understand how exactly it developed, but as best as we can tell, very early cells were able to form that created a slightly different environment relative to the rest of the ocean. And in those environments, you had these conditions for them to self-replicate and propagate. And there's an open question as to whether that started with nucleic acids coding for proteins, or there was some other representation first. You know, evolutionary biologists would know way more what the current scientific consensus is.Link

But at some point in the early history of life's development, we ended up with what is now known as the core biological machinery of life, which is some nucleic acid that represents the strand of information that represents all of the components that are going to be built, and a set of proteins that form the Lego building blocks that produce the machinery of a cell. And all of the biological processes of a cell, including being able to find its own food and metabolize that food and replicate itself and so on, are all encoded in this genetic strand.Link

And once you have a genetic strand, and it's exposed to mutation over time, which introduces some variance and replication errors in copying, then this replicator process gives you the selection pressure over time for finding those individuals that are going to more correctly match the environment. And so it's a very small pressure, but over millions of years, it has a tremendous effect of selecting for things that are going to be much more effective than others. This is the survival-of-the-fittest perspective applied to just the information strands. So that's one knowledge system. It's encoding information about the environment in itself and making it computable.Link

LubaJust wanted to give it another shot. It's basically just the genetic code that is available to us, that works a very specific way, and that's the first layer of the knowledge system.Link

JuanAnd so over time, our living tree of life, in a sense, ended up building a second knowledge system like this, and that was on top of the nervous system. So the nervous system is a set of specialized cells, which are the neurons, which are able to process information. And this is information of a different type. So this is not genetic sequences, but is instead information represented as action potentials, which is like a neuron pumping ions and then propagating an electrical signal. And so these systems were able to initially do some very basic wiring, to be able to introduce motor control in very, very simple forms of life. So there are these extremely simple forms of life that have, you know, tens or hundreds of neurons that are able to move around and sense their environment. And so you have this very basic wiring.Link

And so these neurons served initially as electrical switches, in a sense, being able to translate some information about the environment into decision making. Like being able to wiggle a tail to follow, you know, one path and go towards the food, or away from predators, or things like that. So very, very basic, kind of robotics-type information processing.Link

And that information processing, over time, again subject to selection and evolution, became more and more sophisticated, and the neural systems became larger and larger, to the point where you had, you know, larger animals. Including something small to us, but large relative to other things, like a fly or a mouse, that have extremely sophisticated neural systems that are able to sense a lot about the environment, create a sense of self, a sense of being able to navigate the landscape. Especially, like, mammals have all kinds of social interactions, and so they need to represent some internal state about themselves and be able to relate to other animals, and then be able to make decisions together. You have all this social activity, like some being friendly with each other, some fighting, or whatever. And all of that comes from a very sophisticated set of constraints that these animals have to navigate. And so their neural systems have to adapt to that much more complex environment.Link

And that system became able to learn about the world by observation. So the much more advanced brain systems were able to build representations of the world around themselves and build ways of understanding what's going on. And so that means generating concepts inside the neural tissue of what is happening, to be able to understand what's going on, to then be able to make better decisions and better plan, and find food and find each other, or run away from predators and whatnot, in a much more sophisticated environment. To the point where, you know, you have examples of transfer learning. You see this in birds, for example. They'll have a bird watch another bird that is able to use a tool, and then that bird can now acquire that knowledge by being able to observe another bird doing it. And so you have a way of transferring a concept from understanding and reasoning about what's going on, and from observation being able to generate the required knowledge representation in its brain.Link

Eventually, we figured out how to turn that knowledge into transferable packages. So that means animals, at some point, developed the ability to talk to each other and have very early versions of communication, and that led to more and more sophisticated patterns, to the point where you eventually ended up with very sophisticated languages. It's unclear exactly where those began. There are open questions as to how sophisticated the languages of dolphins are, or where in the primate line that began. Do other primates have sophisticated language, or is it very basic? But at some point, humans, or early hominins, definitely developed very advanced forms of language that were able to transfer very deep representations about the world and themselves and what's going on, and social structures.Link

And that is this second version of knowledge that not only can we refine through our own lived experience, but we can also start transferring to each other. Meaning, we can describe what's going on to somebody else. We can explain some process and put it into speech. And eventually we figured out how to build knowledge tools. Like, the moment that we got writing, which is the ability of taking some knowledge representation that would be in our heads or in our speech and instead encoding it in a permanent or semi-permanent medium. So then you have this asynchrony of, like, now I can take my speech and put it into a tablet and then give it to you, or give it to somebody else, or send it.Link

LubaAnd that's the third knowledge layer?Link

JuanThat's also the second. That's also different aspects of this second layer, which is knowledge encoded in neural networks.Link

LubaTaking inputs, encoding it, and then transferring it from one species to another, or within the species.Link

JuanExactly. So the first layer is just genetic information and genetic knowledge. The second one is memetic: information and knowledge represented in neural systems, which are named memes. So memes and memetic knowledge come from that. And the third one is the digital one. So the digital one, meaning being able to create digital computers that are able to encode representations of reality and transfer them. So this is early digital computers being able to represent documents and information, and then later the internet, and being able to move around packets of information, and eventually artificial neural nets that are able to learn from that environment and learn representations of reality and complex programs and whatnot. So we're in the process of creating the third major layer of knowledge.Link

And so through all of this, even though knowledge is so fundamental to what we do, and fundamental to ourselves and what it means to be human, we don't have a way of articulating what knowledge we actually know. Like, we have encyclopedias and we have compendiums of information, but we don't have a good, systematic way of describing what knowledge we actually know to be true, and what evidence we have behind it. We have very amorphous processes for this. We don't have, for example, a website for every piece of knowledge that we have and what evidence is behind that, or how you might learn that as quickly as possible.Link

And so I became very interested in this question of how you could clearly represent knowledge itself. Both, you know, the scientific process that is expanding what we know, and, if you were to actually visualize the tree of knowledge that we know, how are we expanding it? And then, separately, how are we translating that knowledge into concrete capability expansion, so pieces of technology or products that broadly diffuse that knowledge?Link

LubaWhen did you clarify this specific mission for yourself, that this exact definition of what you care about with respect to knowledge is what you're in pursuit of?Link

JuanYeah, probably this became very crisp for me in, like, late college. This would have been, like, 2008, 2010 and 2012. And it became very clear that this is a very fundamental layer of our existence that is not well articulated, and I think still isn't. In fact, even though we have major-scale systems and processes for how we learn new things, so this is the scientific process, or how we teach a lot of our existing bodies of knowledge, so this means schools and textbooks and information repositories, both in libraries or online, we still don't have a good way of assessing what knowledge we actually have as a species. And what is the knowledge content of a particular brain, right? So what exactly do you know? What do I know? How do we know that? What is the evidence?Link

LubaBut also creating a system for what would be a good representation of what evidence means, what knowledge is. Like, how do you represent the knowledge? Do you ask people who are putting the knowledge down about the experiments, for example, that they ran to come to a conclusion? I guess that's evidence, to come to a conclusion that this knowledge is true. And then I think, culturally, there are probably going to be different variances and nuances for how people interpret it. So it's like, is our interpretive mechanism the same from culture to culture? I'm, like, mind-blown just listening to that. I'm like, how do you fix that?Link

JuanYeah, hard questions. And I think we have rudimentary versions of this stuff. So for example, the crudest version of this is the scientific citation graph, showing this paper citing this other paper, and so they're sort of building on each other. But ideally what you want to know is every single claim, every single piece of evidence that is contributing to a claim or disproving a claim, following the same methodology. Probably you would want this deep graph, the deep representation of all of our knowledge. And then you would want to be able to relate that graph to explanations of this knowledge, to be able to access it. And so, like, you know, if you want to learn something, like a branch.Link

LubaIt reminds me of, remember Roam Research? And the bidirectional linking and graphing.Link

JuanYeah, yeah. So the beginning of the internet is deeply intertwined with similar perspectives of trying to figure out what we know. So the Memex was an early vision of what pioneered hypertext on the internet. This was from Vannevar Bush in, like, the late '30s or '40s. It was a description of a system. It was a visionary paper describing, you know, how you could use early computers to digitize. This must have been the '40s. Must have been the late '40s, actually. How do you digitize pieces of text, and how do you create a global compendium of all of the books and all of the pieces.Link

LubaThe Zettelkasten system came from something along those lines. You know, the system where you write down on a piece of paper your piece of knowledge and then link it to another index card that represents that piece of knowledge. I think it's called Zettelkasten, or something along those lines.Link

JuanYeah. No, yeah. So there were these perspectives about, you know, how could you trace your journey through all of these writings and coverings, and how you keep your own notes, and how you make your own notes accessible. And these were early versions of what later on became hypertext, and precursors of the web and precursors of the internet.Link

LubaSo you basically got very interested in this topic of how do you trace human knowledge, and how do you represent it on the internet or elsewhere, around college years.Link

JuanAnd how do you accelerate the process of discovery. So how can you accelerate the scientific process to be able to know more faster? Because, you know, if you look at history, the more we know and the faster we can know it, the better all of our human experience has become. Like, you know, the more we're able to translate it into better methods for making shelter, better methods for curing diseases.Link

Or, you know, we had the knowledge of how to make a microscope. We had the knowledge of how to clean and make lenses and do glass-based lenses. And then that led to the microscope. And the microscope led to our very direct observation that there were a bunch of little things all over the place. And what are these things? And that categorization, trying to understand what these little things are that we can now see, led to germ theory and our understanding of bacteria and viruses and so on, and that led to sanitation. And so, you know, in the late 19th century you get things like pasteurization and washing hands and so on. And all of that is downstream of people realizing that there were all these little infectious microorganisms that were only visible and understandable once you figured out the knowledge required for a microscope, right?Link

And so if we can just speed up the rate of scientific discovery, and the rate of invention, and the rate of diffusing those inventions into society so that many more people can access them, then our living conditions get dramatically better, broadly, right? And so if we can make more progress faster, then everyone's living conditions automatically start getting better and better and better. You do have to make sure that there's broad access to these, and you need a large-scale global system for how you diffuse the knowledge and diffuse the capabilities. And that's kind of what modern economies are solving.Link

And so, yeah, I became very interested in articulating what actually is the tree of knowledge, how do we speed up the building of the tree, and how do we speed up the translation of that knowledge into concrete inventions and products that could embody that knowledge and increase our capabilities.Link

LubaI love how, now hearing very clearly what is the mission that you're on, or what is the topic that you so deeply care about, there is a very clear through line for how you've tackled this problem with Athena first, then Filecoin, and then now Protocol Labs. Tell me more about what Protocol Labs represents right now, and how are you specifically working on solving this problem?Link

JuanYeah. So the inception of Protocol Labs happened in, like, 2013, 2014. I was a student of history, and had seen that, in order to do a lot of R&D and to do it at scale, you need a lot of capital. Meaning, you need to be able to organize not just thousands of people, but, like, hundreds of thousands of people at scale, just because R&D is difficult and expensive to do. And so I became keenly aware of the Silicon Valley story of how, if you build a technology company and you can create a strong product that has a very large cash flow, you can then fund the development of a lot of other R&D.Link

I was primarily inspired by the great technology companies of the 20th century. So this means things like Bell and Bell Labs. So Bell is the telephony company, and they had Bell Labs, which was this R&D outfit that pioneered things like information theory and solar cells and lasers and transistors. It just was an incredibly innovative environment that was able to produce so much. And they did it at a scale, like, the highest concentration and most scalable R&D single organization that has been built in history. I think, to this day, I don't think any organization really matches what Bell Labs had been able to do.Link

LubaIs it still around?Link

JuanSo, no, not exactly. So there is still an entity, an organization, that claims that lineage. But the Bell Labs of its heyday is no longer around, and, tragically, fell apart because Bell itself was broken up. Bell had this natural monopoly that was built into law, and eventually that got challenged and removed. And then that came at a time of a slump within Bell Labs' history, and Bell Labs couldn't actually recover. And at the same time, you also had Silicon Valley growing and overtaking, and Silicon Valley drained a lot of the talent away from what would have otherwise been Bell Labs' growth.Link

LubaSo at some point in college, you were like, "Hey, I care about this knowledge problem."Link

JuanAnd I need to create an institution that is able to do something economically really valuable for the world, such that we have the ability to form a lot of capital to orient towards creating something similar.Link

LubaYeah, yeah.Link

JuanExactly. And that's the path that I started on. And the first product and technology that I was pursuing was building IPFS and Filecoin. And so these are two related technologies in distributed systems. And so those became the first major products that I could build, to then create some economic cash flow for Protocol Labs to be able to do this larger, longer-term, broad R&D vision.Link

LubaHow did you have so much conviction in Filecoin? Because Filecoin did end up being very successful, which helps you fund Protocol Labs and its efforts. How did you have the conviction that this is going to make money? Crypto was pretty early back then.Link

JuanYeah, that's a whole other story, and fairly orthogonal to what Protocol Labs is now. But, yeah. One of the things I've been able to do over time is understand certain areas of technology and try to think through when they are going to become, let me try another version. So, like, one of the things I do: I tend to look at a lot of different fields and update myself on what are the major breakthroughs, what's happening in that field, what is the potential promise, the range of applications, what are the fundamental scientific discoveries, what are the bottlenecks. And so I keep tabs on a range of fields and what the progress is.Link

And cryptocurrency was one of these ideas that had been explored before Bitcoin. So, you know, the first cryptocurrencies were not exactly named cryptocurrencies, but the vision for it got shaped in the '80s, '90s, through things like the Agoric papers and the Cypherpunk movement and so on. And things like, you know, DigiCash, and, like, PayPal. The first version of the PayPal product that Confinity built, this was the Peter Thiel, Max Levchin first part of PayPal, had a cryptocurrency in it.Link

LubaYeah. And so I actually even remember, when I was back, you know, a kid in Belarus, my dad had digital wallets across, I think, like, Yandex.ru or something.Link

JuanYeah, something along those lines. Probably that came a lot later, I think.Link

LubaNo, no, it was the late '90s, early 2000s. There was already, like, digital.Link

JuanExactly.Link

LubaYeah. Back in Russia, there were digital wallets. And I remember with my dad, like, uploading some money to that digital wallet, and it would convert it to a currency.Link

JuanWas it, like, fiat currency, or was it a different...?Link

LubaNo, it was a different currency.Link

JuanOh, that's awesome.Link

LubaYeah. I don't remember exactly what it was. But I remember when I discovered cryptocurrency, I was like, what is the difference between this new crypto thing and, like, the digital wallet money that we have?Link

JuanOne of the biggest and most widely distributed forms of early digital currency was WoW gold, like World of Warcraft gold.Link

LubaOh, that's so cool.Link

JuanBecause a lot of people play, like, millions of people play that game. And there was economic activity that you could do.Link

LubaThat's pretty cool.Link

JuanYeah, it's pretty wild. So all of that early history, and the first wave of peer-to-peer systems, and the web and so on, showed that eventually you would have digital forms of money, and Bitcoin was the first very successful version of these. So even though there had been a number of attempts before, what Bitcoin really succeeded at is solving this large-scale, long-term adoption problem of how do you build a large enough currency that a lot of people start using and believing in. And that shaped the beginning of the modern cryptocurrency industry.Link

And so at the time, I actually happened to see Bitcoin fairly early. I probably tested it in '09, '10. And at the time, it seemed like yet another project like this, and it was not clear that it was going to work. It wasn't clear to me why it would succeed. I remember mining some amount of it in, like, a computer cluster, and then.Link

LubaDid you keep your keys?Link

JuanNo. I think we mined a bunch of it, and, like, I forgot about it, whatever. I don't even remember how much it was. Who knows. But at the time, it was just not clear that it was going to succeed. And, you know, I think a lot of us really underestimated Bitcoin's potential. And by 2011, 2012, it started really showing strong signs of success, where now a lot of people were starting to use it and talk about it and build wallets for it and ways of transacting in it and so on. And so by 2012, 2013, it was clear that this had major potential.Link

And around that time, as I really looked at Bitcoin and tried to get deeper into its process, I found this core idea of the mining process of Bitcoin extremely interesting. So the mining process of Bitcoin is this ability of allocating some portion of the currency to funding a service that the network is going to run. In Bitcoin's case, it's just the hashing power that secures the Bitcoin blockchain. It also serves as an exchange, in a sense, in that you can walk up to the network with a computer and effectively acquire Bitcoin by providing your computing power to the network. And so that gives you a way of bootstrapping the entire system and bootstrapping access to the currency.Link

But what was really bothersome to me is that the computing process that Bitcoin used was effectively this useless piece of work. Though it's securing the blockchain, it's not doing anything more useful beyond that. It's just hash rate power that's wasted. And it seemed very clear to me that you could do something actually useful with that process, that you could weave a more useful process into the consensus. And that's where Filecoin came from. So it came from me wanting two things. One, a currency that I could use with IPFS to have a market, to be able to pay people for storing data. And second, a crypto mining process to weave into the block reward function that would do something actually useful. And that's where Filecoin came from.Link

In terms of what you asked me about conviction and where that came from: a lot of it comes down to thinking very deeply about the systems, thinking about the competitive landscape, thinking about why people use things. And then very clearly seeing that if you fulfill a set of product requirements, and you're able to do it at cost, or ideally cheaper than alternatives, then you have a product to sell. And so that gave me a strong sense that we could build something really compelling here.Link

In IPFS's case, it was about a combination of feature sets that didn't really exist before. There were precursors that were similar in some regards, but not quite in the combination that IPFS offered. And so being able to do portable, peer-to-peer websites was just not something that was available before. And so that was really fundamental to the success of IPFS. And then, separately, for Filecoin, it was this very strong mining process attached to a currency that was able to amass massive-scale hardware, to devote to the storage network. And so that gave me a lot of conviction, even just on paper, let alone, you know, once we had code and product, and people started using it, and so on.Link

LubaAnd the storage piece of it, is it tied to the knowledge base? Like, were the pieces of information that people were storing kind of in hope that this would lead to humanity's information storage, for that bigger mission of knowledge and evidence-based learning?Link

JuanYeah, it's a great question. So for a set of us in the community, yes, although these systems are fairly separate. Meaning, whenever you're building a product or a computing system, you want to try and make it as useful as possible in as wide a range of use cases as possible. If you make it too general, then it might not quite fit enough of a use case, but if you make it too specific, then maybe it's too narrow. And so both IPFS and Filecoin are broadly information systems. They're able to store, address and transfer any kind of information.Link

Now, a subset of the community really cares about scientific progress, the storage of really valuable information that is culturally very significant. A lot of the early IPFS community came from Open Access science communities that really cared about being able to create peer-to-peer versions of scientific journals, and being able to have open-access scientific papers and data, and having peer-to-peer distribution of data. Especially, you know, scientific data can be very big and expensive to move around, and if you use peer-to-peer, then you can split the cost between everybody.Link

And then that also translated into the Filecoin community, where a large segment of the Filecoin community is deeply focused on creating publicly verifiable and publicly accessible compendiums of knowledge and information that all of humanity really cares about. Both scientific data, but also culturally significant data. So this might be all of the records of books that are already out of copyright, like libraries' worth of stuff. The Internet Archive is a close partner of the Filecoin community, where we just want to be able to generate large replicas of all these really valuable long-term cultural artifacts. And so, yeah, it's a very important part of the community, but it's not the whole story. Meaning, you should also be able to use these systems to store and distribute, you know, application data, and personal data, or your organization's data, and so on.Link

LubaSo you created Filecoin. Filecoin was as successful as you wanted it to be, or maybe more successful, maybe less successful, but basically, baseline, it was a successful product. And so how did that.Link

JuanI feel like Filecoin is going to grow, and will have a pretty successful long-term outlook. A lot of us are working on it very strongly. It's one of these things where crypto networks turn out to be much longer projects than people expected. So a lot of Silicon Valley-type products and companies can very quickly form, build and grow, and either succeed or not. Crypto networks are just a different beast, and they take a lot longer to produce super high-quality product experiences, and they're really big in nature. And so there's a lot of Filecoin's growth that is yet to come.Link

LubaYeah. And I think that's true for a lot of innovation too. We kind of think that, oh, AI was an overnight success, or the internet was an overnight success.Link

JuanThe 1960s.Link

LubaYeah, exactly. Okay, so Filecoin. Tell me more about how that evolved into more of the mission that we're talking about.Link

JuanYeah. So when I started Protocol Labs, I built an organization that had this strong DNA of being able to do this long-term R&D. And because we started with IPFS and then Filecoin, and then we built, you know, other projects and protocols like libp2p and a number of other systems, we already had this DNA of building many possible products. And over time, I became aware that in this period, so, you know, the early 21st century, the traditional Bell Labs approach of building one large corporation doesn't quite work anymore.Link

LubaHow so?Link

JuanSo the latest attempts to do this include things like Alphabet, so Google. Say Alphabet is maybe the best attempt at it. There are others, but I'll use it as an example. Alphabet was the attempt of taking the Google cash flow, meaning the ad-based revenue idea. I think Alphabet has been the latest, most successful attempt to try to do this, where Google historically already had in its DNA being able to try a lot of different products. So it was not just Google Search, but it also created Gmail, and acquired a range of companies, like the stuff that became Google Docs, and YouTube, and so on. So already, Google as a company had this perspective of trying many different products. And then when Larry, Sergey and Sundar and others created Alphabet, that leveled up the organization into being able to reach for other kinds of R&D that were very far away from what Google traditionally did. And so this is when things like Waymo became possible. I mean, Waymo was already a project that had started as part of what became X later on, and was already associated with Google. But they also could do things like Verily and many other bets that were outside of the traditional Google wheelhouse. And as part of that, they separated X into its own R&D lab and so on.Link

But in practice, it hasn't been that successful, even though it's been very expensive to operate. If you add up all of the cost structure of all of Alphabet's other bets, it's quite a lot of money that has gone into those organizations, and they haven't yet succeeded in productionizing a ton of that R&D.Link

In that same period, there was a different attempt at doing this, or not even an attempt at doing this, just a different process: YC, which started around, I don't remember the exact date, like 2006, 2008, maybe even earlier, maybe 2004 or so.Link

Luba2005.Link

JuanYeah, was that the first class? Yeah. So, you know, mid-2000s. By the time that Alphabet started, YC had been funding tens of companies at a time. And in the time between Alphabet and now, YC has been able to fund many thousands of companies and partner with, you know, probably over 5,000, close to 10,000 founders across a wide variety of possible fields, and many of those ended up being quite successful bets. And YC did this with a small fraction of the capital that Alphabet did.Link

And so it showed me that if your goal is to greatly accelerate R&D broadly across a lot of different fields, the traditional Silicon Valley approach, which is to build one large organization, one large company, is no longer the best way. Instead, if you create a structure to partner with founding teams of various different sorts across a wide swath, you can greatly accelerate hundreds or thousands of different teams operating across the entire R&D landscape. And so once I realized that, I changed the trajectory of Protocol Labs.Link

LubaWhen did you realize that?Link

JuanI realized it in 2019, 2020 or so. Yeah, this kind of insight takes a while to realize. It's kind of hard to place, because probably the insights for it were brewing, but me really doing the crisp analysis and putting numbers to it probably happened in '19, '20. And I have the slide that I've used in talks, where I compare Alphabet and YC and the nature of the two. It just taught me that if I really wanted to do this long-term R&D acceleration, I needed to approach it as a network, as opposed to building a single large company.Link

And so I veered the trajectory of PL to turn into this, what I call an innovation network. So it's not an innovation company, it's an innovation network that creates organizational structures to support a number of different teams across the R&D pipeline, doing all kinds of R&D across many different fields. And part of what we bring to the table is capital. Part of what we bring is a lot of talent, a connective tissue between all of the organizations, many different ways of helping. But it's tuned very differently, so it doesn't have a hard boundary. It's highly collaborative. Yeah, it's a network-native type of organization.Link

LubaTell me a bit more about how it's different from a VC fund. Because obviously, YC started as an accelerator that gave you capital. They have a program. But then you could argue that certain traditional VC fund structures try to support you with talent or networks as a differentiator now, especially since it's so competitive. How do you see PL and its structure as different from, like, a traditional VC fund?Link

JuanYeah, it's a great question. So I would say it's qualitatively different in a few ways. One piece is, PL has a VC side to it. And so we have a set of venture funds that do the very crisp and concrete venture capital approach of investing in specific stages in a particular way, particular-size rounds. We create a venture capital product for a set of founders and a set of LPs, but it's a piece of the larger network. Other stuff that we do that, say, most VCs don't do is we also deploy capital into very early-stage science and invention, often before there are any companies, or often in projects that have yet to go through the translation process to identify what will be the concrete.Link

LubaWhat's an example of a project like that, that's too early to be a company?Link

JuanSo we funded a ton of work in cryptography, and early stuff that led into zero-knowledge proofs development. We've done a lot of funding in the DeSci space. We've done a lot of funding across neurotech. So we've funded a number of, effectively, science projects, and a set of organizations that are called FROs, focused research organizations, that are doing more academic technology-building or invention work, often before there is a company. And so this is stuff that is just outside of the realm of possibility of a VC. And we keep this work very separate from our venture side. So our venture funds are a separate part. This is more like the grant-making part. And so our long-term business model is to be able to start and invest in companies that produce a very strong cash flow, and then we take part of that cash flow to fund this very early part of the process, to be able to do long-term R&D.Link

LubaSo there is the VC arm. There is the early research projects that you find to be very important for humanity arm. There is probably.Link

JuanNetwork forming. So we create a bunch of connective tissue across this whole R&D pipeline that is maybe closer to what, I think YC does a version of this, and, say, large tech companies do a version of this, where you create shared culture, shared information processes, a way of people across the entire network being able to help each other, or solve problems together. That kind of connective tissue. We do a lot more of that than, say, most groups, either in venture or in grant making, ever do.Link

LubaRight. And you either support or do a lot of in-person events yourself too. That probably serves as connective tissue too. How do you pick problems that are most important to work on for the progress of humanity? Is PL very opinionated about the types of verticals you invest in, or is it just like, "Okay, this seems interesting," kind of like a coin toss, might be good, might not be good?Link

JuanSo yeah, we are very opinionated. So we currently have four focus areas, and part of these came from where we grew organically. So that means, very pragmatically, taking a position based on what we've done so far that has been successful. And so our first two focus areas are related to crypto. The first one is around Web3 and digital human rights. So this means building secure communications infrastructure and secure large-scale transaction platforms. So this means... So we have four focus areas. The first two relate to crypto, and this is where we just have a very large presence. The first one is in Web3 and securing the internet and establishing digital human rights. So this means secure communications infrastructure, secure publishing systems, secure transaction systems. So this is projects like IPFS and Filecoin and Ethereum and lots of different kinds of blockchains, projects like Bluesky. Just anything that relates to stronger, more decentralized foundations for the internet.Link

The second focus area is what we call upgrading economies and governance systems. And so this is more the cryptocurrency side of the crypto world. This is things that use game theory and smart contracts and mechanism design to build new economic primitives that can improve global economies, or create a digital internet economy. This includes projects, you know, like Bitcoin, projects like stablecoins, DeFi, DeSci, applications of DePIN that build better production infrastructure. So projects like Glow, which is applying DePIN to growing solar energy build-outs. This includes the new pop-up cities movement, or network states. And it includes a new set of digital economic zones. And so all of this is applications of crypto and computing infrastructure to upgrading economies and governance systems.Link

Then our third focus area is around AI and robotics. And this comes from us deeply believing that the invention of artificial intelligence, and where this is going to go, is probably the most important technology that humanity has ever built, and potentially will ever build. In that this is creating this third knowledge system, potentially giving birth to a new species, which is a very deep, hard thing to do and think about. And it also is about broadly optimizing all of the current economic and social systems. And so that means automating production and automating, like, just everything that we do. And so we think that is a tremendous inflection point. Certainly we see it now with all of the AI labs and what the latest generation of LLMs can do. But we think that this is very early days in the capability development.Link

And here the best teacher will be the predictions of very thoughtful scientists that have looked at this very deeply over the last, you know, 20, 30, 40 years, and sci-fi. I think sci-fi was very early to thinking about the implications of all of this stuff, way ahead of others. And so, you know, the writings of people like Turing and von Neumann and others that saw the beginnings of this, to, you know, Moravec and many others that saw where AI development was going to go eventually. But also a lot of the sci-fi stories that looked at how humanity was going to come to terms with the creation of new artificial intelligence systems and whatnot. And so that whole part is focus area three.Link

And our fourth focus area, which is our most recent addition, is neurotech. And so this means being able to blend computing systems with our neural systems. So this means, how do you connect brains or neural tissue to computing interfaces, through brain-computer interfaces? Or how do you digitize neural tissue itself? Meaning, how do you map the brain itself and be able to represent it digitally, to then have a digital version of humans, which might become really important down the road. So this is early days, but we already have the connectome of a fly. In the next three to five years, we'll probably have the connectome of a mouse, and eventually we'll have digital mice running around in a virtual environment. And someday that can include humans too, and that would be a major development for the species, right? That would be a major leap in the human condition. And so it's unclear exactly when that's going to happen. I sort of expect it to happen sometime in the next 15 to 30 years. Large variance there, but I think it would be very weird if we haven't achieved that by, say, 2070 or 2080 or something like that. And so that's our fourth focus area.Link

And so over time, PL will keep adding more focus areas, because it requires a certain amount of focus on each one of them to be able to enter that area with critical mass, a number of people, and so on. And, yeah, all of this is very opinionated, in terms of: these are technologies that we think are going to be transformational for the world and for the species and for everybody around us. They tend to have to do with computing as a major lever. And they tend to be technologies that we think have an enormous amount of impact potential.Link

LubaOkay, wow, there is so much here. There are so many interesting areas I'm gonna tap into. The last one, the neurotech, BCIs. A few months ago, I watched the Neuralink update from maybe six or eight months ago, and I was absolutely mind-blown by how much is already possible. Like, there are people playing video games just with their mind. There are already people who can move their arm just because Neuralink, specifically, is inserted and implanted into their brain. And I feel like, well, maybe that's because it's not the world that I very closely follow. There must be so much more happening that I'm completely unaware of. Tell me more about what is your interest in neurotech, what is your vision for how humans should develop with these devices, and, like, what's broadly happening in the space?Link

JuanGreat question. So maybe I'll make three strong cases for neurotech. And again, just a quick redefinition: this means technologies that interface directly with our neural tissue. And so this means primarily our brain, but also, you know, the nervous system, all the way to actuating our muscles and our proprioception and so on. And there's also a way of applying neurotech and neuroscience insights to the building of artificial intelligence systems. So the reason we call them neural networks in computer science is very much inspired by neurons.Link

LubaYeah.Link

JuanSo we did the early science of understanding how neurons work, and we built mathematical models of those, and that's what became the individual neuron and neural networks, and the beginning of the AI field. And that's no accident. It's just, you know, a super powerful, very general computing system that is able to very flexibly ingest vast amounts of data and sense-make on top of this data through some very straightforward learning rules. And so it's a fundamentally very powerful way of building learning systems.Link

So in some forms, neurotech has been in use for a long time. You know, I include things like caffeine as a version of neurotech. It's like a small compound.Link

LubaBecause it's a thing that affects your nervous system. And so technically, you're injecting something artificial or foreign to your body into your neural system. That's a very interesting way to see it. Yeah, I guess we're already cyborgs.Link

JuanYou know, caffeine helps us improve our cognitive performance. But then, say, alcohol, which is kind of very opposite to caffeine, is also a form of neurotech, but that's more for social environments and so on. I tend to not drink anymore and so on. But, you know, historically, our species has been a heavy user of alcohol in lots of social settings and whatnot. And so I think both of these are a version of neurotech that changes the human experience and changes human capabilities.Link

And so I'll make three cases for it. So one case is just straight up from looking at human capabilities and creating a form of repair or treatment for damage, which might be the result of accidents, or a disease of some sort, or some other condition, where lots of people are affected by some problem that they encountered. Whether it's, you know, again, due to an accident. Sometimes people are in some type of accident that gives them a spinal cord injury, and then they lose mobility. Or in some cases, there are a range of diseases or conditions that affect vision or hearing, where people have hearing loss or vision loss. Or in some cases, maybe this could also be a genetic condition, where a set of people are born without access to some of these systems.Link

And one branch of neurotech is about creating systems that can help people with these conditions to be able to repair or access, or mitigate, some of the problems that emerge from those kinds of accidents and so on. You know, one version of this might be the cochlear implant, which was a way of restoring hearing for a range of people. Or, again, some of the implants that are restoring mobility, of being able to have a brain-computer interface that then enables you to access a computer, you know, move a cursor or something like that. And that tends to be used quite a bit, because computers then can serve as an interface to control other things, right? If you can control a cursor on a computer, and you can get really good at that, then you can probably also control, like, a chair. And so somebody that might have a form of paralysis, of different types, can somehow then control robotic systems around them to regain a form of mobility.Link

And so for a lot of people, this is, you know, transformative technology that gives them back forms of agency and forms of capability that they tragically lost. In some cases, very understandably, you know, all of these forms of experience are different and varied, and some people don't want to access these. It should come down to a choice, right? At the end of the day, we should be able to use the fruits of science and technology to give people the ability to choose these systems, if they want to regain that capability.Link

We already have some neurotech implants that can restore vision. So one of the systems that I'm most excited about is the PRIMA retinal implant from Science Corp, which helps people with macular degeneration, which is a form of vision loss where, you know, people develop this dark patch in the focal point of their vision. And the PRIMA implant restores some version of their sight. And we'll have, you know, next-generation versions of all of these kinds of systems. And so, you know, it might be restoring senses. It might be restoring mobility. There's other parts of neurotech which might be around better self-regulation. So there's a set of neurotech devices around supporting people with, you know, forms of depression, or Parkinson's. There's a whole range of neural disorders that can be greatly assisted by forms of neurotech.Link

LubaHow is it split between a device and an actual drug that can affect any of those neural pathways? I understand that it's two very different ways of affecting your brain pathways. But, like, is one inherently better than the other? Or is it really highly dependent on the disease, quote unquote, that you're treating?Link

JuanSo yeah, disorders vary a lot, and the technological interventions have a wide range of impacts. And so there are some disorders you can treat pharmacologically, meaning you can develop a drug to target it. Certainly, some forms of depression can be treated with different kinds of drugs. And I know that there are a range of medicines for Parkinson's and MS and conditions like that. However, we also have evidence of actual physical devices, like an actual implant, that can assist people with those kinds of conditions.Link

And then, you know, things like sensory loss or loss of mobility, that comes down to repairing. Like, you know, if you have a spinal cord injury, being able to regain mobility comes down to being able to bypass that damaged component. But in many cases, you can't actually use the remainder of the nervous system, and you have to go and actuate, like, an exoskeleton, or a robotic chair around you, and so on. I am hopeful that we'll do the deep biology work here to actually figure out how to regrow parts of the nervous system that might have atrophied over time, and to actually be able to restore the nervous system. But that'll come probably downstream of many more years of deep understanding of how neural systems develop, and can you actually regrow them, and so on. This is probably, you know, deep biology work that needs to happen for a long time.Link

LubaHow far away are we from normal people being implanted with different devices, or using the devices to enhance their capability?Link

JuanEverybody? Like, I think the word "normal" is very hard, and I wouldn't use it, because people just have a wide variety of experience. And maybe people have a disorder, but even if you don't have a disorder, you might have just a wide variance of perspective.Link

LubaYeah, that's.Link

JuanYeah. So I would just, like.Link

LubaI guess, like, obvious conditions versus non-obvious conditions that don't really impair your experience as much as someone who doesn't have their sight or doesn't have.Link

JuanYes, I think maybe the term might be. Yeah. So let me maybe get to those two other cases. So case one is for repairing some disorder. Then case two is for expanding or enhancing the human experience. You know, I mentioned caffeine and alcohol before. Those are enhancements of the human experience, right? People normally, without ingesting caffeine or alcohol, tend to have, you know, a particular range of experiences. You add this piece of neurotech, and then they have a wider range of experiences, right? Like, I don't know, Van Gogh used to have a range of, like, I think, you know, absinthe and alcohol and whatever. I think it was absinthe, I don't remember. But, you know, a certain set of artists have used all kinds of pharmacological means to accomplish incredible forms of art and so on. And they expanded their human experience, and then shared some of that with the rest of the world.Link

So there are ways of expanding what we get to experience. And neurotech gives us a way of greatly, very broadly expanding those capabilities. So it might be stuff as simple as taking one of our senses and expanding the range of what it can cover. So for example, we usually hear within a very tight, narrow band. What if you could actually hear a much wider band? Or we currently see only a very small part of the light spectrum. But what if we could actually see a much wider band? What if you could have infrared vision, or, you know, have eyes that could actually focus very far into the distance?Link

Or one that I'm really excited aboutwe're normally trichromats, where we can combine three colors to generate the millions of colors that we see. But there are tetrachromats. You know, a few people in the world have been born with the ability to actually combine four color channels, and then the colors that they can see are just way more. The combinatorics means just a vastly larger number of colors that most people cannot see. So imagine being able to actually crank that number up. Your visual experience would get dramatically more interesting.Link

LubaI find this comparison of neurotech with drugs, or various enhancements that we can consume to enhance our human experience, very interesting. Because there is so much more worry about injecting a literal device into your body than there is about, you know, drinking maybe a glass of water that was spiked with something that would still enhance your experience in some way. And I guess it's because people are so worried about constantly having some sort of impact on their psyche, versus being affected by some chemical through, I don't know, the cleaning supplies that we're using, or the water that we're drinking. It's just a very interesting perception dilemma.Link

JuanTotally. And I think it's very natural, and I think it comes from these things not having been normalized yet. So, you know, there was a point in time when we didn't have organ transplants.Link

LubaAnd people had to get comfortable with it?Link

JuanYeah. And there was a whole, very deep, you know, part of becoming a society comfortable with the idea that you would take the organ of another person and transplant it into somebody that needed it, right? And initially, that was seen as super crazy. But now it became clearly broadly beneficial, clearly very impactful, and then society adapted and moved on. And so I tend to see the same kind of thing applying here. You know, in the days before the car, the world was organized very differently. And you have these newfangled machines that arrived that were, you know, horseless carriages, and, like, "What the hell is this thing?" And people, over time, started adapting, and there were a lot of moral panics and consternation around the arrival of this mechanized system. And, you know, the introduction of the car into a city started reshaping how it worked and whatnot.Link

So I think a lot of it will come down to the first devices being put out there and normalized. And to be clear, there's already, I think, tens of thousands of people with cochlear implants today, maybe hundreds of thousands. I don't know the exact figure, but a lot of people already have BCIs of some sort. And so this is actually a lot more normal than people might expect. I do think that eventually this will be as normal as something like getting a piercing, or something like that, where, you know, today, I don't know how many millions of people get piercings or tattoos or something like that, or, you know, buy one of these, like, supercomputers. It's just that we haven't quite figured out the really easy ways of creating a form factor that then actually gets implanted and connected to our nervous system. But once we figure out how to do that really well, then this will become much more normal.Link

And the good news here is that there are a number of teams. So maybe Neuralink is the most visible, but there are, you know, 10, 20 BCI companies that are developing technologies now, many of which are in late stages of clinical trials, that will arrive and become strong products. Again, initially, most of them are targeting these medical applications, these sensory or motor repair conditions. But these other cases around enhancing the human experience, I think, are even more broadly applicable. Meaning, I think by number, many more people are going to want these. It's just a market that doesn't quite exist yet, and so people aren't really asking for it.Link

But when you talk to most people and you say, "Hey, would you want to be able to see better? Would you want to have further-reaching vision? Would you want to be able to see additional colors? Would you want to hear additional bands of hearing? Would you want to be able to taste more tastes?" most people are like, "Yeah, absolutely." The hard question is, "Okay, but how?" And then the question becomes: how safe can we make the products? How easy can we make the product deployment? What is the whole life cycle of the product? What does it mean to get one of these? For example, I strongly think that these systems should be able to be implanted and removed and upgraded. And so you have to think of the entire life cycle of a generation of a technology. What does it mean to get it and upgrade it over time, and so on?Link

So in this enhancement of the human experience piece, one piece of that will also be how we relate and communicate with each other. There's really cool sci-fi that's been written around this, of how, you know, if you have these systems, they're able to then convey our thinking, and it can become another form of communication. So it may not just be translating our neural patterns into speech, but it might also be transmitting our neural patterns to each other. So there's good sci-fi written about what happens when you start being able to telepathically communicate with each other, where you're able to share entire visions of the world, like walk into the same dream or the same imagination. Imagine the human experience expansion that can happen.Link

LubaThat's wild.Link

JuanBetween, like, you know, friends and couples and partners and family. Or being able to relive memories together, in full, rich, you know, 8K, with all the smells and tastes and so on. Being able to imagine something together. Just what becomes possible once we connect our brains to some computational medium that can then transfer that communication between us. It's just incredibly phenomenal and fantastic. And so that's all category two.Link

And then the third case I would make is around computing and digital systems and artificial intelligence. Which is that, because we're making these artificial neural nets that are becoming so capable and powerful, and we're building intelligent systems that are now more capable than many humans at many things, right? So we have this concept of jagged intelligence, where the AI systems today greatly exceed humans. They're broadly superhuman at a number of cognitive things. Not everything yet, but that will come with time.Link

That raises a very deep, hard question for humanity, which is, you know, what is our future going to be like? Where are we going to be 20, 50, 100, 200 years from now? Our biological history has been that, over the last million years, we became the apex species in the world, where, thanks to our intelligence, our ability to coordinate and our ability to create technology, we gained control of the planet, for better or worse, right? For better, in terms of all of the great living experience we've been able to give billions of people that otherwise wouldn't have happened, and the amazing works of art. The human experience is just so incredibly rich with broadly beneficial things. And for worse, in that, in the meantime, we've also made a bit of a mess of the planet. We've managed to broadly harm many other species. We've hunted many to extinction. We've broadly decreased the amount of wildlife that there actually is. You know, a set of species have benefited, right? So the species that are livestock or pets have broadly benefited. So now there's way more cows and sheep and dogs and cats than there ever were before, thanks to this symbiotic relationship. But, you know, in the meantime, the polar bears are not doing great. And so we have to come to terms with that and figure out a good path for the future.Link

But this also raises this deep questionif we're creating this entity, or set of entities, that are artificially intelligent, and they are going to be more intelligent than humanity, what does that mean for us? Are these systems just going to be tools? Or are they going to be agents? Or are they going to be beings with their own volition and their own perspective and their own ego and their own identity? And that becomes creating a new species. And if we do that, knowing that they are going to be much more intelligent and capable than we are, what sort of relationship are we creating? Are we giving birth to our descendants, or our children, so to speak? Like, the children of humanity are the digital beings. And are we going to just cede control of the future to them? Or is there some middle ground? Can we create a path by which humanity can also benefit from digital evolution? That humanity can also benefit from the fact that this third knowledge system, the digital knowledge system, can be broadly superintelligent and broadly incredibly capable? Can we bring humanity along, so to speak, with that new digital set of beings?Link

And so from my perspective, we have done the scientific thinking. We've done a lot of the grounding on the fact that, at the end of the day, what we know about humanity is that we have physical, organic brain neural tissue, and that's where our consciousness emerges. And if we can connect that to computing systems, or we can digitize that neural tissue in that brain and represent it inside of a computer, then that digital brain would also be conscious and would also have the same kind of experience. This is not a given yet. We haven't answered the hard scientific questions of understanding what consciousness actually is, but we're dramatically closer to that than ever before. And again, here there's good sci-fi that talks about the possibility landscape of being able to have digital humans, and then, you know, having the ability to snapshot yourself, and, you know, you don't have to die if you don't want to, or you don't have to age. Or you can refactor yourself, or you can improve yourself.Link

LubaWhat's your definition of what consciousness is, or what's your hunch as to what consciousness is?Link

JuanSo to me, our conscious experience is an emergent phenomenon, an emergent process that happens as a result of the type of information processing that our brains have to do. We have to ingest a wide range of sensory input that represents updates on the world and what's happening, and we have to create a robust and deep understanding of what is going on to be able to make good decisions. And we have to make good decisions in a complex environment that includes social relationships with many other people and many other animals. And those social relationships require forms of identity and forms of individual decision making that, you know, preserve your life. Like, you have to make decisions that are good for you, and you're not gonna accidentally kill yourself real easily. And so those are the basic building blocks upon which you can form an ego and form an individual identity and form a self-aware, conscious perspective.Link

And I think this will come down to very concrete functional units of awareness, and awareness of your self-awareness, and awareness of your self-awareness of awareness. There's a recursive structure here. And it'll turn out that we have this very functional way of describing how consciousness can emerge, and this is a consequence of the information processing that must happen, and the decision-making loop that we have of how we control ourselves, how we actuate, how we move, how we talk. Like, we have to pick the next action to do, and that has to be done with deep context of everything that has happened to us in our lifetime, and all of the deep relationships that we have with other people. And so that gives rise to our conscious, aware experience. And if you take that information system and you replicate it in a different environment, that system will also be conscious in the same kind of way. It'll have the same kind of trace of experiences and trace of knowledge and whatnot.Link

So this third case for neurotech iscan we not just enhance the human condition, but can we preserve humanity into the future? Can we give humanity the benefits of digital evolution? And can we avoid this potential conflict with the digital beings that we're about to create, by creating some kind of merger in between, of being able to co-evolve with AI systems and grow together into the future?Link

LubaBut wouldn't you argue that, in some ways, AI digital creatures and, say, uploaded intelligence, so humanity uploading itself into the cloud with, you know, the same set of experiences, the same level of consciousness, is kind of still AI digital systems? Because AI systems were trained on human knowledge, and, at least right now, they're evolving similarly, or thinking, or giving us an impression that they're thinking, similarly to humans. Like, where would the difference lie? Because if we do go down the route of uploaded intelligence, what should humanity do on Earth? Like, when people kind of die in their physical shape or form and then are in this digital world?Link

JuanYeah. So lots of great questions there. So I would say, first, because we're training LLMs on our textual and audio-visual outputs, digital representations of all those, we do have a collection of AI minds that are fairly close to ours. This didn't have to be. If we had built artificial intelligence using the game-playing methods that built AlphaGo or AlphaZero, or many other historically described approaches to building artificial intelligence, we could have ended up with very different minds that were much more alien to our brains. I think, thankfully, because we're training the current swath of LLMs on our cultural outputs, the space of those minds is actually relatively close.Link

However, that doesn't say that the future ones are going to remain that close. Probably, you know, some form of evolution will take over, and they'll grow and develop in their own ways. And so they'll probably differentiate from us. You know, if you think ahead 10, 20, 50, 100 years from now, 200 years from now, there's a very significant evolutionary pressure that will emerge. Just think that, you know, 10, 15 years ago, these neural nets were barely recognizing objects in an image and recognizing faces. And now they're composing symphonies, and composing sophisticated plans, and, you know, helping you think and giving you advice on problems. And, you know, in many cases, doing people's work for them. They're already incredibly capable.Link

LubaMuch more capable than most people, I would say.Link

JuanMore capable than most people at most cognitive tasks. Yeah, exactly. There are still many cognitive tasks that people still beat LLMs at, but it's not clear for how long.Link

LubaYeah, exactly. I think I've already kind of given up, that AI is much better than me, probably.Link

JuanAt a set of things. And so you have to retreat to the things that you're really good at uniquely. Yeah. So that's true. But then, you know, this does create this environment where we're creating these digital beings that are just going to have a much larger, richer experience than humanity is going to have. And what does that mean for humanity? Like, maybe it's great for them, but what about us, right?Link

So from that perspective, I see neurotech as a way to enable humans to go along on that journey and be part of that future. And so having a path to digitizing and uploading human minds gives you something closer. It's also a question of what's going to lead to a broadly more beneficial future. I think that there are very concrete, real AI-risk sorts of reasons why uploaded humans might be a more ethical and more beneficial species. Because we already have very clear cultural ethical tendencies. You know, as bad as humans can be about a range of things, we're also quite good. We tend to be very hard on ourselves, and we have great capacity for compassion and great capacity for moral perspective and moral judgment and so on. And so we already have evidence of being at a point in the spectrum.Link

And if we create fully artificial systems, it's not entirely a given that they will necessarily be more ethical than we are. I've been very pleasantly surprised by the latest models. I think, like, you know, GPT-4, GPT-5, the last two Claudes, these systems do convey a depth of moral judgment ability that is quite good. But it's not clear to me how much of that is just knowledge and understanding versus their own deep intrinsic motivations. That is very difficult to tell right now. You basically have to do LLM psychology to try and understand what is just something that they say they think versus what they really think. And so there's a deep, broadly-beneficial-future-for-everybody case for uploading humanity alongside creating these new digital beings.Link

And you ask the very key question, which is, if we do that, then what happens to, you know, biological humanity? How much of that is around, and so on? These are super hard questions that we are going to encounter culturally as a species over the next few decades. These will become some of the key questions we have for ourselves. You know, one easy answer to a lot of this can be, like, hey, digital beings don't really need, you know, the Earth. Digital beings can totally take the asteroid belt, or take Mercury and build a Dyson sphere next to the sun. And so you can totally come up with a good division of environments that could work out. This is also discussed in sci-fi at very different points. I think the Culture series talks about biological humans and uploaded, or digital, cyborg-type humans, and fully digital intelligences, as three different branches of humanity, plus AI. That's a possible landscape. It's not really clear what is going to work long term. A lot of this just comes down to human agency and what people choose to do, right? It really has to come down to what each individual person wants to do for themselves, and what their communities want to do, and so on. There's a recent sci-fi that I think is extremely good. Extremely good. Called *Pantheon*.Link

LubaYeah, that I highly recommend, because it was super interesting. It deals with a ton of these questions, right? And I thought one of the questions that really stood out to me was the one about us not really graduating from childhood because our parents are still around. So psychologically, you're never really entering adulthood. And I was like, oh, so maybe the psychology of these uploaded humans is really going to change. Because there is, you know, the intellectual perception of the world, and then there is the complicated emotional perception of the world, which I don't think we fully understand yet. So is that going to evolve as the technological systems are going to evolve? Or, yeah, how are we going to adapt to that? I think in that way, we're going to become a very different species.Link

JuanLook, I think humans are very adaptable culturally, and so I wouldn't worry too much about any one of these types of questions. Things will just be different. The only constant is change. If you think about our life today, and what we do, it's just so different than what people did 1,000 years ago, 10,000 years ago, a million years ago, right? And so, you know, we're in the middle of New York, you know, on an island of skyscrapers and buildings, and, you know, we have clean drinking water, and we're talking into these microphones that are.Link

LubaLike, enhancements in the form of caffeine. Like, this is just already.Link

JuanWe are a totally different species than people were 1,000 years ago. It's just that our biological evolution hasn't changed that much, because it operates just much slower. And so probably a lot of our instincts and drives and so on are similar, but what we do on a day-to-day basis and what we care about is just very different. Just think about, you know, the world before books, and before stories were written down. Just think about the human experience back then. That's not to say they couldn't find amazing fulfillment in a ton of ways, and tell stories, and have broadly great, fulfilling lives, but it's just very different, right? It's just a different slice of experience.Link

And I find the world today, where, if you want, you can go to the jungle and have, you know, a life away from society, or a life with a small group that is very different. Like, you know, there's people on the Sentinel Islands today living like Paleolithic humans. You can go do that too, right? You and a group of people could go do that if you wanted to. It's difficult, but it is possible, and you have agency over that choice. And at the same time, you can choose not to do that, and you can choose to, you know, sit in a room with incredible books and, you know, incredible music, and write books, or write music, or paint, or, you know, build a building, or whatever. And you have this other very rich life that you can access. Now imagine the future, and vault into what will be possible for us to do 100, 1,000, 10,000 years from now. That possibility space is just so much larger than what we do today that it just seems incredibly exciting to be able to access that.Link

LubaSo speaking of the future, a shorter-term future, not thousands of years from now: what are your predictions for humanity 10 and 15 years from now?Link

JuanI don't think a ton is going to change in 10 years. So, you know, 2025 now, 2035. So I'll make a few guesses here. And I will preface this by saying these guesses are well-informed guesses, but the probability distributions are quite wide. And so some of these things might, you know, vary by five or 10 years.Link

And so I would say, 10 years from now, we'll probably have already achieved, and broadly recognized that we have, you know, full AGI-capable systems. And not just in this jagged intelligence perspective, but we should have, in 10 years, systems that are just vastly superhuman across all, if not almost all, tasks. We might have superintelligence. I think that is within the timeframe. I would think it is very likely that we have superintelligence. So that means recursive self-improvement, to be able to create a vastly superhuman set of AI systems that are able to think well beyond what humanity thinks about today.Link

We should have, by that point, some version of digital societies. The form factor that we have today for AI systems is one where you have many different agents, like, you know, hundreds of thousands of separate agents, all trained from base models, but having different experiences and different interactions. So that should be able to be connected into something like a digital city, in a sense, a digital kind of experience. There's a real question as to whether or not we're going to build these systems with a long-lived experience, or whether, for safety reasons, we're going to be inhibiting that. And so there's a deep, hard ethical question, both for humanity and for these systems, as to whether or not you give them, initially, the ability to have longer-lived identities and longer-lived senses of self. There are clear reasons why you shouldn't do that, but also clear reasons that you should. Otherwise, these systems are coming into existence and dissipating very quickly. It's a very different form of identity.Link

We should have, by 2035, I would say, tens or hundreds of millions of robots, like humanoids. I think, you know, this could go very quickly. By 2030, I think we should probably have, like, hundreds of thousands to low millions. We will definitely have, by 2035, solved the core mobility challenges for humanoid robots and the dog-style robots and so on. So we should be well into lots of robotics production in this sense. Actually, the *I, Robot* movie got it quite right on the timeline. It sort of predicted the humanoid robot stream right around the early 2030s.Link

We should have, hopefully we will have, massively scaled solar energy, and so we should be greatly decreasing CO2 emissions by that point. Especially if China keeps growing the way it is in solar and nuclear, and if the US picks up the pace too. And so that means that we should be well into avoiding the worst effects of climate change. We might still have some very bad outcomes. We might still have some of those horrible heat waves that might kill hundreds of thousands of people, or something like that. It would be great if we can avoid that. But I think, hopefully, we'll be in a much better spot there.Link

I think we should have, by 2035, so 10 years from now, much better neurotech that restores a number of these senses, like I was describing. We should probably have hundreds of thousands, if not millions, of people with BCI implants that recover and repair some amount of sensory and mobility conditions. And hopefully, I'm not 100% sure about this, but hopefully we will have the consumer use cases starting to develop. It would be amazing if we could have tens of thousands or millions of people already using these implants for that kind of connectivity. That might lag a little bit more than 10 years, but not much more, I don't think.Link

It's possible we will have unlocked automated production of certain types of things. So it could be that we have much more automated production of cars and computers and a lot of these kinds of devices. Maybe, once we have humanoids, we might be able to automate more construction of housing and materials and so on. And so maybe we can greatly decrease the cost of housing production. We should be able to have much more abundance in terms of the broad, you know, basic goods of living.Link

I think we'll face some very serious, hard questions around what the automation of production means for people at large. Meaning, today, those automation tools potentially will lead to this separation of capital and labor, and our current world order relies on a very strong coupling of capital and labor. And so I do think that we need to come up with some other economic structure that enables humans broadly to have access to the fruits and benefits of all of this capital growth that we'll have. And so this is where things like, you know, some form of public goods production and building and so on is really needed. And I think we need better economic models and better economic systems. So we have to do that work fairly quickly. But I think we'll be in the early innings of that. I don't think we'll be in the "suddenly there are hundreds of millions of people without jobs." It's rather, I think there will be millions of people that are displaced and have to relearn things fairly quickly. But it'll start hitting that.Link

I think in 10 years, we probably won't have a human upload. Maybe we do. I hope that we do. But it's more likely that we'll certainly have a mouse, maybe a dog already, running around in a virtual environment. And hopefully we'll have much more secure systems for broad communication and whatnot. I don't know where the deep geopolitical questions are going to go, in terms of the US and China and the geopolitical and economic tensions between them, which hopefully don't boil over into some kind of conflict by that point. But that's kind of where I think we'll be 10 years from now.Link

So now, for 50 years, it's way harder to predict, because this is past, you know, the predictions around the singularity, which really just means that the volatility of how these different things will intersect is so high that you lose the ability to predict many things. So maybe the way that I will approach this is that I'll say a range of things that will have already become very possible. It's not clear that they will necessarily happen, because it's very path dependent.Link

But I think 50 years from now, you know, which is 2075, we should have had, you know, the ability to upload millions of people. We should have had broadly artificially superintelligent systems. We will have gone through the big questions around what humans and AI systems are going to do together, and whether they're going to co-evolve or speciate or differentiate. We should have already gone through the big geopolitical tensions between the US and China. It's not clear what the outcomes of that will be.Link

We should have broad energy abundance by that point, either through solar or nuclear, or some combination of the two. We should be well into robotic production and automation of most systems. And so everything to satisfy the human condition, in terms of Maslow's hierarchy, the lower layers of that, like access to shelter and access to food and access to all of the material needs, should be well covered, or able to be well covered. Hopefully by that point, we will have built up better economic systems and better economic models.Link

We should have, by that point, millions of digital beings. Potentially, actually, by 2075, it could be billions or trillions of digital beings. And whether those will be on Earth or, you know, we'll start building the Dyson sphere, we can probably be well into that by that point. Hopefully there will be a Mars colony. It's not clear to me that that will make a ton of sense if you can go to the Dyson sphere instead. But it's a cool thing to achieve anyway, so maybe we should do it anyway.Link

And I do think that, hopefully, we'll have this much richer experience of life that I was describing before. We'll have gone through a few decades of broadly enhancing our senses and our experience, and be able to create new forms of art. Think of, you know, the forms of art of today are greatly intertwined with the technology we have, like the technology of making paintings and making movies and making music and so on. So as we unlock new technological capabilities, that will create many new forms of art. And that could be fantastically interesting and cool to experience.Link

And I think, hopefully, by that point, we will probably have solved a ton of deep, hard questions about science, questions around the universe and the nature of reality. And hopefully by that point, we'll have figured out what consciousness is and how it emerges. We hopefully will have grown up and built much better ethical frameworks. Hopefully we'll have, you know, stopped our encroachment upon the lives of many other species on the planet, and be on the road to more recovery there. And so, yeah, I think it could be a tremendously positive potential future 50 years from now. It's not that long.Link

LubaThat's true. It's so true that it's also path dependent. Because if we were doing these predictions in the '70s, we'd be like, "Oh, our physical world will look so different than it is right now." But hopefully we'll be watching this video 50 years from now, uploaded into the cloud. So we'll see. It'll be interesting to watch it then.Link

JuanYeah. 1900 to 1950. Like, if you looked at the world in 1900 and made predictions for what the world in 1950 would have looked like, or 1910 to 1960, yeah, you get these wild, wild changes.Link

LubaAnd to close it out: for the physical world of today, what's bringing you joy these days?Link

JuanOh, a ton of things. So probably the most important: I got married earlier this year, and that was just a phenomenal thing for me. And my wife and I get a ton of personal joy out of thinking about all these things, and working with people, and playing together, and creating together. I love reading and listening to sci-fi and different, you know, stories and fiction. And I now love various different forms of working out that are new to me. Like, I got really into weightlifting, and that's been really fun. Got really into sauna culture and cold plunging and so on. That's been super, super, super fun. I get to talk to a lot of amazing people, and so that's been wonderfully fun.Link

LubaJuan, thank you so much for your time. This was really fun, and we could have gone in so many different directions. Thank you for sharing your perspectives about the future with me today.Link

JuanThank you so much for having me. Yeah, super fun that you're doing this.Link