Into the Impossible With Brian Keating - Jaron Lanier: VR Will Expand Human Consciousness

Episode Date: November 27, 2025

If you want to think sharper, feel stronger, and perform better — visit www.tonum.com/IMPOSSIBLE and use code IMPOSSIBLE for 10% off your first order. Try Shortform, the invaluable app that helps me prepare for every conversation I have! Get $50 off the annual plan at https://shortform.com/impossible Today, I'm speaking with Jaron Lanier, a pioneer in VR, about where it will take us next to expand human consciousness. In this wide-ranging conversation, Jaron Lanier explores how technology reshapes perception, identity, and the future of humanity. From the psychology of virtual reality to the energy demands of modern AI, we trace how today’s tools influence what it means to be human—and what kind of humans we might ultimately become. KEY TAKEAWAYS 00:01:52 Lanier warns AI may reduce human uniqueness. 00:09:23 VR can alter how we perceive and inhabit our bodies. 00:12:58 VR faces biological limits like cyber-sickness. 00:28:43 Reality and VR both distort perception in useful ways. 00:40:20 AI’s rapid growth is driving major energy demands. 00:54:59 Apple’s original “iPhone” idea was partly inspired by Lanier’s VR headset. 01:00:53 Talmudic tradition shows the value of preserving multiple perspectives. 01:14:59 Human senses are both extremely precise and deeply flawed. 01:31:10 Tech culture often mimics medieval-style philosophical debates. 01:41:45 Social media harms users by manipulating attention. 01:51:26 Technology shapes the kind of humans we choose to become. - Additional resources: Get My NEW Book: Focus Like a Nobel Prize Winner: https://www.amazon.com/dp/B0FN8DH6SX?ref_=pe_93986420_775043100 Please join my mailing list here 👉 https://briankeating.com/yt to win a meteorite 💥 - Join this channel to get access to perks like monthly Office Hours: https://www.youtube.com/channel/UCmXH_moPhfkqCk6S3b9RWuw/join 📚 Get a copy of my books: Think Like a Nobel Prize Winner, with life changing interviews with 9 Nobel Prizewinners: https://a.co/d/03ezQFu My tell-all cosmic memoir Losing the Nobel Prize: http://amzn.to/2sa5UpA The first-ever audiobook from Galileo: Dialogue Concerning the Two Chief World Systems: Ptolemaic and Copernican https://a.co/d/iZPi9Un 📺 Watch my most popular videos:📺 Neil Turok https://www.youtube.com/watch?v=Dt5cFLN65fI Frank Wilczek https://youtu.be/3z8RqKMQHe0?sub_confirmation=1 Eric Weinstein vs. Stephen Wolfram https://www.youtube.com/watch?v=OI0AZ4Y4Ip4?sub_confirmation=1 Sir Roger Penrose: https://youtu.be/AMuqyAvX7Wo Sabine Hossenfelder: https://youtu.be/g00ilS6tBvs Avi Loeb: https://youtu.be/N9lUceHsLRw Follow me to ask questions of my guests: 🏄‍♂️ Twitter: https://twitter.com/DrBrianKeating 🔔 Subscribe https://www.youtube.com/DrBrianKeating?sub_confirmation=1 📝 Join my mailing list; just click here http://briankeating.com/list ✍️ Detailed Blog posts here: https://briankeating.com/blog 🎙️ Listen on audio-only platforms: https://briankeating.com/podcast #universe #podcast #briankeating #intotheimpossible #science #astronomy #cosmology #cosmicmicrowavebackground #intotheimpossible #briankeating #Jaronlanier Learn more about your ad choices. Visit megaphone.fm/adchoices

It's peak pollination season, and my business is scaling fast. To keep the nectar flowing, I need a phone plan with top priority data speeds. That's why I chose Google Fi Wireless. My connections stay strong even when the hive is buzzing. Plus, unlimited plans started $35 a month. Now that's a deal that doesn't stay. Explore Google Fi Wireless plans today. Plus taxes and government fees.

Google Fi Wireless is not subject to data traffic deprioritization during times of high network usage. You said this place was steps from the water. We just haven't found the steps yet. How much did we save? Enough. Enough to get lost. Or you could book a stay with Hilton.

Welcome to your oceanfront room. Just steps from the water. The Hilton sale is on now. Book on Hilton.com or the Hilton app and save up to 20% to get the stay you expected. When you want savings, not surprises. It matters where you stay. Hilton, for the stay.

What if I told you that the man who invented virtual reality thinks AI might be making us less human? But the same guy who jammed with Richard Feynman believes technology should expand our consciousness, not replace it. But here's what most people get wrong. They think VR and AI are basically the same thing, both artificial, both digital, both changing reality. But Jaram Lanier sees them as opposite forces. VR expands what's possible for human consciousness. It literally changes the way that you imagine, create, and how your brain works. when you inhabit a different world.

Classic VR where you're not seeing the real world anymore is being able to change your own body in interesting ways. That's an amazing sensation and I think it has an indirect effect on your cognition, which is really profound. There's one thing that might still shock people that they haven't seen yet. And here's the most uncomfortable truth that Silicon Valley doesn't want you to hear.

When you interact with an AI lover, an AI therapist, an AI girlfriend, you're not interacting with some neutral technology. You're interacting with a company whose interests override yours. Today, Geron Lanier, VR Pioneer, Microsoft Scientist, musician, and one of the most brilliant and contrarian thinkers I've ever spoken with goes deep into consciousness, creativity, and the future of what's going to happen to your very brain itself. Let's go. Today I'm sitting down to talk with one of my favorite thinkers, Jerome Lanierre, R. Pioneer, author, musician, bon vivant. I think the only thing he's missing is a high school diploma. Is that right, Aaron? Yeah, I guess that's true. Yeah.

Well, by the end of today's conversation, you won't think about our actual reality the same way ever again. Duran, long before you became known as the father of virtual reality, you spent time jamming with Richard Feynman at Caltech. Now, Feynman, as you know, experimented with sensory deprivation tanks and used his own body as a laboratory at a probe physics. Watching him do that, what did you learn about sensation and feedback, and how has that shaped your version? for virtual worlds that replicate reality through purely artificial sensors and sensory input. I cherish that I had this connection with Feynman, but it was a peculiar one. I wasn't officially a student, nor a lot of times people come to me and ask things about his life, but I actually don't know a lot of the details. I haven't read all the bios, and I don't really know the public

Feynman that well. The story for me was pretty simple. My first girlfriend was somebody I met in New Mexico where I grew up, and her parents were separated. She was there. I've chased her back to L.A. It turns out her dad was the head of the Caltech physics department. I was a bright kid, and I just ended up kind of informally around there and had a chance to spend time with them. And I did play some music with him, if he mean by jamming, because he was actually a kind of a cool, eccentric percussionist. And then much later, he invited me to be the only male, the only non-hippie girl person to accompany him on LSD experiments where I was told my job was to keep him from falling off a cliff in Big Sur, which I succeeded at. But I think he enjoyed having all the hippie girls around more than me.

And not that anything happened, but I'm just saying there was a certain aesthetic exploration. Okay. So that's me in Feynman. I don't feel like I learned a lot about the censorum and such from him. that wasn't the particular thing with him, but I have been fascinated by it, obviously, and I did have many other people to learn about it from. I have always been fascinated by this very peculiar state, which is the human condition, where we are physical. Our only connection to the world is through imperfect senses, through these physical exchanges of information that follow the same conservation laws as everything else.

And in a sense, we're kind of remotely and imperfectly connected to our world. And yet there's this other sense in which we have this inexplicable sensation of really being here and being in it and being sort of real in a sense beyond just a bunch of particles. I'm totally fascinated by that. One of my old phrases from a publication was, consciousness is the only thing that isn't reduced if it's an illusion. The mere possibility of illusion is the thing, you know. And so this weird state that we're in, which to me I have to turn to metaphysics to talk about, is absolutely fascinating to me. And it has been since I was little.

And it's definitely part of what led me into the whole virtual reality thing of experimenting, replacing all of the sensory channels and the interactive channels with synthetic ones and seeing what happens, which we did. So obviously you pioneered virtual reality decades before it really became mainstream. And we're going to get into, you know, some of the tough questions that I have for you as the father. By the way, who's the mother? Who's the mother of the art? I know, that's what I would say. I don't call myself the father of spirituality. But whenever somebody does, I said, well, it depends on if you believe the mother.

I guess the mother virtuality, there are a few interesting candidates for that. I mean, Ada Lovelace is an obvious one. But another one that's interesting is Suzanne Langer, who is an art theorist from the post-war period, 40s and 50s, mostly, who did use the term virtual world to describe something about her sense of art, which I think is what inspired I have in Sutherland to start using virtual world to talk about computer displays. So maybe Suzanne Langar, although we never met, so I don't know how that worked exactly. Maybe some kind of male and sperm program I participated in as an infant or something. I don't know. You're saying you might have, well, let's not get into that.

when you look at, you know, when we look at what you have fathered, sire, you're not denying paternity at all. But looking at where we are. That thing has a lot of expenses. I'm going to deny paternity. I'll leave it to Zach. Yeah, okay, fine. Yeah, we'll get into some of the lacuna. But now it's become mainstream.

And so I want to ask you, looking at where we are now with VR and AI, what's the one thing that the world still doesn't understand about these technologies that, think is going to shock us in the next decade or so. About VR stuff, I find almost all current VR to sort of not be getting at the stuff that's the most intense about it. It's very rare. And if you, like, go through all of the online sources of stuff for VR through meta or steam or anything, it just seems like they all miss the stuff that I think is the coolest, which is really strange to me. Like, to me, the coolest thing in classic VR where you're not seeing the real. world anymore is being able to change your own body in interesting ways. That's an amazing sensation,

and I think it has an indirect effect on your cognition, which is really profound. So turning into an octopus or something. And there's academic study of it where we study how much you can stretch the homunculus, homunculus flexibility is what it's called sometimes. I should have called the plasticity to be more academic, but I was just in an anti-academic mood. But there's at least two really good labs in the One is Mel Slater's in Barcelona and others, Jeremy Bailensen's in Stanford that have been devoted to trying to map out how the human brains mapping to the human body can be modified in VR. And it's utterly fascinating. I mean, you can really bring out different cognitive specialties in people by changing how the body they think they have. There's also a wonderful hypothesis.

Yeah, yeah. Tell me the most exciting developments on the horizon that lay people might not be aware of and why they're so significant. In virtual reality, whenever I look at all of the offerings in the Metastore or the Steam Store, I'm always amazed at the stuff that I think is the coolest is just not there, and I don't understand why. Some of my favorite things are changing the human body. So if you're in fully occlusive classical VR where you don't see the real world anymore, you have the option of changing your body. You can turn into an octopus or all sorts of things, and you map your real body's motion into this thing.

And you can even map yourself into a distribution of the clouds in the sky, all kinds of things. And when you do that, people's cognition is changed indirectly, and you can bring out different qualities in them, which is amazing. So there's a couple of labs that study this as a primary focus. One is Mel Slater's in Barcelona, and another is Jeremy B. Ellenson at Stanford. And one of the hypotheses that I really like in this is that when we learn, which body types your brain seems to be able to control the best. We're both exploring the deep phylogenetic tree. In other words, the evolutionary history that your brain passed through to come to the human body. Because your brain doesn't know what species it's in. It's just gradually being

adapted to each species along the way. So it retains traces of controlling other bodies. And when you become those bodies, it's like they're natural for your brain. But the other thing, which is even more amazing, is what future bodies we could evolve into that the brain is pre-adapted to. And this in a way is one way of exploring the potential for the human brain in the far future. It probably gives us information about hooks that might be used to modify or enhance the brain in the future. It's just an incredible thing. And you can study it with today's virtual reality, except people don't, for some reason, that escapes me. It's the very coolest thing. I've had so many experiences like that that I'm really excited about. But anyway, so that I'd say if there's one thing

that might still shock people that they haven't seen yet, it's that. It's changing your physiology inside VR. And then if you do it together with other people, it gets even weirder and cooler. So that's the thing. Soon I want to connect your work and your thoughts about virtual reality to, I want to connect it to scientists like Donald Hoffman, who makes sort of an allied case, perhaps, in this case for reality in his book about consciousness. and so forth. But, you know, ultimately, we'll get there soon. But I think the, the question I have, you know, kind of most prominently in my mind are the limitations of VR and the perhaps, you know, sorry to be maybe insulting, but maybe the failure to live up to what is such a promising

set of potentialities or potential futures. I'm thinking most recently of Apple Vision Pro. I know they're a competitor to... That's a sad case, isn't it? Yeah. So I... I want to bring up something to you. I read a study just preparing for the interview in Science Direct, and it was all about the experiences that people have with vestibular, you know, kind of challenges from using headsets. I almost brought my meta headset here.

But the test involved, you know, interactions with people denying them access to their peripheral vision. And when this occurred, this is before the Apple Vision Pro, by the way. I think it's gotten worse. They complained of something that is called cyber sickness. Have you, have you accounted that before? Yeah, I mean, that's like very fundamental to VR research, of course. So that's fundamental. And that's both as an interesting science topic and that obviously as a practical engineering topic.

It's absolutely fundamental, of course. Is that going to, since our hardware is, you know, millions of years old and our software, just language and stuff is relatively new in the hundreds of thousands of your time scale. Do we have any hope of with an external device that does, you know, eliminate your peripheral vision where predators tend to attack. Is that a fundamental barrier to adopting it at least widespread or maybe does it put a break on its ultimate potential? Hey, I want to use this opportunity to ask your audience to help me with something. Yeah. Many years ago, a NASA researcher in VR named Michael McGreevy used to show a slide set

with the very first head tracked visual display, little goggles. And it wasn't for people. This was in this would far predated anything I did or Arvin Sutherland or anything like that or any of the remote robotics things. It was for kittens and it was to study how kittens perceive peripheral vision. And so it was like this little thing with just a line that could move. It was not digital. It was mechanical. And so there's this cute picture of all these kittens wearing VR headsets from the 40s or 50s. And I want that picture and I can't find it. All of you through this, you need to go out and help me find this thing. I need crowdsourcing because I can't find it. A. I can't find it. It's there. I didn't confabulate this. It's a real thing. I believe it.

So about motion sickness, it's a fundamental thing to study. Can I tell you a funny story about it? Of course. Funnier the better. And then I promise I'll get to the science. You know, I'll keep you on track. In the 80s, I guess it was. Yeah, in the 80s, I used to do some stuff with Steven Spielberg, maybe most famously for the movie called Minority Report. But earlier than that, we put a bunch of VR stuff in a big truck and rolled it from Silicon Valley to Hollywood to just show VR this new thing. At that time, profoundly esoteric, hard to explain, profoundly rare, profoundly unusual, and crazy expensive. Many millions of dollars of equipment were in these trucks just to give people an experience of looking around inside of virtual. So anyway, we were, we set up the truck in the lot

at Universal, one of the studios. And there was this real old classic studio head named Lou Wasserman, who's still running the place. And so he went and he watched all these people go through it. And everybody, of course, was just blown away and didn't even understand what was happening to them. And then this little finger comes and says, okay, kid, kid, come here.

And he says, are people going to get sick in this thing? And I'm like, oh, Mr. Wasserman, we have been studying this. We have it down to an incidence of one per thousand and we believe we can get it to 1 to 10,000 in the next year. We can't solve it totally. We can make it very rare. We know how to do blah, blah, blah. And he looks at Spielberg, and he says, why are you bringing me some kid who doesn't know anything?

And Spilberg was like, oh, God, I'm embarrassed. And so then I'm like, he says, kid, let me tell you something. I want to see headlines about how my janitors are quitting and suing me because of all the vomit they have to clean up. That's what sells tickets. Until you know something, you really don't belong in Hollywood. They had just done that with a movie called Jaws, where apparently people were throwing up in the theater, or at least that was the story.

And my first introduction to business and simulator sickness was that it's desirable. Today's episode is sponsored by Shortform, the smartest way to keep up with the world's smartest ideas. Look, I'm a professor, podcaster, and parent, and I don't have hours to spare. When I needed to prepare for a guest like Geron Lanier, the technologist and philosopher who helped invent virtual reality and now warns about its social

consequences, I started with short form. I began with their Chrome extension, which summarized Lennier's essays and interviews on the attention economy. Then I opened Shortform's guide to 10 arguments for deleting your social media accounts right now, unpacking how algorithms exploit our emotions, and use enragement to lead to engagement. I was on a deadline to release this episode the same week his book got published. So, short form was indispensable. I needed to make the connections fast and they didn't disappoint. Lanier's work connects with thinkers like Tristan Harris, Sherry Turkle, and Nicholas Carr, revealing how we can reclaim our minds from the platforms that seek to compete to dominate them.

But it's not just books. Shortform adds articles, podcasts, and even AI-generated summaries. I'll use it for research on AI hype cycles, for catching up on the latest astronomy papers, or yes, even sometimes for parenting questions, diving into their podcast and articles. It's become my daily driver, my curiosity companion, helping me save hours and hours a week. And it will for you too. You can try it for free. And if you go annual, you'll get three extra months on me. Whether you're curious about Lanier's vision for a more humane internet, or just want to explore new and brilliant insights on technology, consciousness, and society, shortform's got you covered. Start your free trial plus three bonus months at shortform.com

slash impossible or click the notes below. Now back to my conversation with Geron. And maybe this was an unfortunate moment. It's not a simple thing. There are a lot of simple things we can do to reduce the occurrence is a great deal. One of the things about it that's not simple is that there's a great deal of variation

and how sensitive people are to it and the way in which they're sensitive. And there's certain narrow demographics that are exceptionally sensitive. And it's kind of peculiar. I'll give you a clue. They happen to be the ones who don't work in VR labs. So I remember when I was talking, without naming any names, I was talking to some of the very nice people at Apple who were bringing out their headset.

And before it was announced, I said, we have totally solved sickness. Nobody will ever get sick. You don't know what you're talking about. And like that one of the very first reviews was a certain journalist who got access to it and got sick and happened to be, let's say, a smallish Asian female, very prime demographic to be affected by it. Why? We don't know. But this is actually one of the areas where the lack of diversity in core engineering teams in Silicon Valley really bites hard. Like, I'm sure it's not the only one. I'm sure there's a lot of others we don't know about. But, you know, Apple, it's not that there's no way to solve it totally. That was the pipe cream. But you can reduce it a great, great deal. And you can also design the experience in the overall product so that when it happens, the person can deal with it and recover better. and the whole thing.

Like, you have to acknowledge it, though. If you pretend you've solved it, then you're deciding not to solve it, you know. And the thing is, people even have it in reality. Like, we get a little dizzy and disoriented in reality. And this gets back to what I was saying before, that our sensory and motor systems are not perfect. And a lot of what goes on function of the brain is actually managing the fact that our physical link to the rest of physicality is imperfect.

That's a core quality of cognition. And you don't treat that as a problem to be solved. You treat that as a quality to be acknowledged and embraced and worked into everything, you know. But anyway, yeah, there's a lot you can do. Peripheral vision is a lot of it. Relative latencies or different parts of the system are a lot of it. Sometimes just design, like I was talking about making your body as weird as possible. If I want to make you fall over, if I want to make you throw up, I sure can do it.

No problem. But, of course, I don't. And so there's a design space that's fundamental to people. One of the, the great thing about virtuality is it forces you to think about people biologically and concretely instead of abstractly. When you're designing on the screen, you can sort of imagine your users this abstract thing who just moves a cursor point around or something. With virtuality, if you're not being a realist about biology and everything else about humans,

you are failing. Unfortunately, that's the majority of the field. but, you know, and when you ask why isn't VR picked up more, I think that's actually a lot of it. It's the reluctance to accept the fundamental, messy, gooey, biological nature of people, which is so anti-Silicon Valley culture, but unless you can really go there, you're not going to be able to really design for VR in a way that catches. You know, I think that's where we've been. It really resonates what you just said with me,

because I've been making the case that it might be impossible, at least with the lock-in phenomenon, that is the victimization by its own success of LLMs plus GPUs, that we might not experience another nausea-inducing phenomenon, which is what Einstein described in this famous free fall experiment, thought experiment. He said, if you were in an elevator and the cable broke, here's my real reality finger puppet of Einstein here,

if the cable broke, you would experience no gravitational force field. And he called that thought. He said it titillated him unlike anything else in 1907, and that was the happiest thought of his entire life, he said, looking back on it. I want to ask you. That's believable. I can imagine it would have been. It still is an amazing thought, isn't it?

It is. It's absolutely amazing. And it gives me hope for these meat brains, not like this 3D printed one. You have all these props. I need to keep up with you. Get your props. You go ahead. I sat there.

Come on. I went to your house in Berkeley many years ago. Weird props to have. But anyway, okay, go ahead. So my question is, without a embodied. you know, kind of body, literally, how is it possible for an AI system, an LLM system perhaps, or even other forms of AI systems, to have a happiest thought and even to visualize what that sensation of the pit in your stomach rising up that induces nausea in many people, as they said. How is it

possible? Are these types of geometries of LLM plus GPU? Is that just fundamentally not going to lead to a breakthrough like Riemannian curvature did for Einstein. Yeah, well, that's a controversial topic right now. And I'm kind of on the inside of it. I'm the prime scientist at Microsoft, and I get to see a lot about how the models work and try to improve them. I'm sometimes perceived or accused of being a skeptic. I don't think I am at all.

I like them. I think we're bringing something of value to the world. I just really like realism. I like honesty. And I think, so what I think we have with the big AI models, with large language models in particular, is from a mathematical point of view, we have something that detects patterns and projects them or extrapolates them

in a variety of ways in a statistical distribution, which in some cases, if they're pretty close to where they're, the training data was that they, and if they're in certain situations, can often be useful. And a great example is in vibe coding and small programs that can often be useful. And I think that's great. I actually think that that's a benefit overall. Sometimes they can't be, but it's all statistical. I mean, statistically once in a while, it's like a Boltzman's brain, but it's not so severe.

The LLM might come up with some new theory. I mean, it's not precluded. It's just less likely the more, let's say the more creative or, you know, or unlike the training data, something is the less likely it is to come up because this is a big statistical machine. And then from a social and economic and semantic point of view, what we have is a new form of human collaboration. So just like with the Wikipedia, we combine people stuff together and we kind of suppress who the origins came from through pseudonymity. But nobody thinks Wikipedia is anything other than a jumble cooperation of a bunch of people. So a large language model is exactly the same thing.

It's a bunch of stuff from different people that's jumbled together using a bunch of rather computationally expensive statistical processes. That's what it is. That's all it is. Is that useful? Hell yeah. It's great. It's really useful.

Pretending it's more than it is makes it less useful. Like realism makes things more useful. That's why reality is there, I guess. you know, so, so I think this question of could, could an LLM come up with the new theory of everything? Well, you know, it could. It's just not super likely. I don't think there's any harm in trying. Believe me, I have. And I've had students do it. Like, like, why not? Give it a go, you know. You can have, you can ask one of the frontier large language models to write you, a theory of everything paper. And it'll come out as good as a lot of what's out there. there. Yeah. Like it could get on the archive probably because I get three of those a week. I get three. Oh my God. I do too. I get so many and people are so convinced that they have it.

And you know, it's like there's some symmetry group and you do this and you do that. Oh, here it is. You know, and like a dirty little secret is it's not that hard to come up with a paper like that if you want to. But it's also not that hard for a model to come up with a paper like that if you want to. So we're going to talk soon about as I said, past guest and friend Donald Hoffman at UC Irvine and his thoughts about you know reality, you know, not existing. The case against reality is his book. I imagine, you know, your book could be titled The Case for Virtual Reality.

But let's, before we get to that, let's go through Dawn of the New Everything. Let's judge a book by its cover. You're never supposed to do it. Hey, book lovers. We're judging books by the covers. We know we're not supposed to do it. Into the impossible, there's nothing to it. Let's take a look and judge some books.

Oh, yeah, so I did find one for you. Here's the paperback edition of it. I think actually the hardback in the American version is a little cooler because they put a hologram on it. Yeah, but this is a picture of me. It's got to be in my 20s, and I was taken by Kevin Kelly, who was an old friend and started Wired magazine along with some other people. And that's one of the earlier versions of the VPL headset.

That's the version that existed in the early 80s, and it was supplanted by a better one. But that is, to my knowledge, the world's first, you know, had supported color stereo tracking display and certainly the first commercial one. And yeah, so there it is. And that's me when I was younger, my hair shorter.

And the subtitle has, kind of connection to the next topic, which is about actual reality. Can you read the subtitle? What was the choice of the subtitle based on? Oh, let's see. It says, Dawn of the New Everything, Encounters with Reality and Virtual Reality. Let me see something interesting about the cover that some might be curious about. There's another book with a very similar title called The Dawn of Everything, and that's by two people,

one of whom is David Graber, who was a dear friend of mine who unfortunately passed away, just before his book became a bestseller. And that coincidence in the titles was intentional. Because what that book is about is reconsidering the archaeological record to argue that there's been much more diversity in human societies than we usually think. And that if we look at that diversity, it might make us more optimistic about human potential and human nature.

That we've been looking at one little blip during the period of written history, but it actually isn't that representative of how we could be if the energy cycle were different and if the technology based were different. So a really interesting book. And this was a bit of a memoir of coming up in Silicon Valley

and starting one of the Silicon Valley things. And the idea was that they would be linked for those who were curious. It's intentional. So there was the dawn of everything, Goddard, the knew everything. Mine came out a couple of years earlier. I'm amazed that I wasn't flakier than those guys.

I would have thought mine would come out later, but somehow or rather that's what happened. Anyway, the subtitle is, Encounters with Reality and Virtual Reality. that in the subtitle, at least in America. I think it might be different elsewhere, but anyway. Probably, yeah. So as I mentioned, you know, Don Hoffman is a friend,

and he's argued that our perceptions aren't really Windows true reality, but they're basically survival-based Darwinian adaptations that provide desktop interfaces, kind of like a VR rendering that hides all the messy code, again, for our survival. And Don of the New Everything, you're kind of, you know, have a contrary perspective to some extent, that, you know, if reality is warping our senses, in virtual reality is warping our senses,

then that's something we need to be mindful of, but also sometimes lean into. I mean, I sense sort of a dichotomy there. But let me ask you this. If reality itself is an illusion, let's just take his proposition of face value. How would that shape the design work that you and your colleagues are doing now for headsets, for software, for human interfaces and so. So to my embarrassment, I have to say, I have not read his book. So I can't speak to you directly, but if I'm,

going to speak to your characterization of it. What I would say is that we often get into a little bit of difficulty of wanting to think in a binary when there's really a statistical distribution. So do we perceive reality directly? What did that possibly mean? I mean, one piece of reality cannot fully know another piece of reality. If you don't know what I mean by that, go and study your quantum theory. What we have is an indirect kind of sloppy way of.

of reality connecting together. Now, in a sense, it's all very locked in because of conservation laws and least action principles and all these, like there's a sense in which it's all walked in together so far as we can tell. But in terms of any little piece of it, getting information about another little piece, that's pretty, that's pretty funky. Like reality is set up as an imprecise kind of undergraduate project in that sense. I don't know who to complain to about that. But anyway, so this notion that there's some even plausible hypothesis, that the brain knows reality directly. Unless you're talking about something metaphysical,

the brain might know consciousness directly. I'm not even sure about that. That's an interesting area we could talk about, very difficult and very difficult to say anything with clarity about it. But in terms of physicality, the very hypothesis that we could know reality directly is not definable and just doesn't even make sense to me. So everything you said, I would agree with totally.

But the thing is we can know is statistically. And this is what's important. So when my half, touch? Do my hands really touch? Well, if you really went down to the smallest level of reality and you're looking at the particles, not really. What's going on there? There's fields interacting. It's a little complicated, you know, but what does touch even mean at that level? It's not even defined, right? It doesn't even, it's not even. And so, but statistically, yeah, my hands are in cosmologically unlikely to interpenetrate when I do this, right? I don't think that lacking

an absolute knowledge of reality means lacking any knowledge of reality. Rather, I think we have some knowledge of reality at a statistical level and that that's legit. And that's real. The case for acknowledging that statistics are real, I'm all on board for that. Our perception of reality is statistical. And to me, that makes it real. In fact, in one of my books, I have a comment that the way you know something is real is that you can't know it totally. I think that's in you're not a gadget. And I think that's accurate. Whenever you know something fully, that means you know some kind of abstraction or a construct, and that is exactly not reality. It's when you only know something partially that it's plausibly real. I think that does kind of resonate with Hoffman,

at least at some level, because life by necessity is a process of filtration, of compression, of loss, so forth. And I think that's one way to look at it. That's in consonants with what he said. I want to turn to another guest. We've got to get to, you know, the questions from my best man at my wedding, Stefan Alexander. But before we get there, a question from Rizwan Verk, who's a teaches at ASU and he's written a book about the simulation hypothesis being real and so forth.

He has a question for you and the influences of your work. And he wants to know in terms of science fiction and real world antecedents, like the Sword of Damocles demo by Ivan Sutherland. And so he's curious what you think about things like lawnmower man, snow crash, both of which have been influenced by you and your work. And all best wishes today to our friends at ASU who are apparently shut down by a dust storm. Oh, no. Oh, I didn't hear that. You must have found that out on Twitter because people. No, I don't. I don't so far to those things. I actually talk to individuals and my friends today. I know, I know. I'm teasing. The cyberpunk media movement, in a way, it predates real virtual reality because we have people like Philip K. Dick and Ray Bradbury and so on writing pieces that incorporated virtual reality ideas, or at least awfully similar from the 50s and really early, but in the 60s.

But starting in the 80s, I was in touch with all these people. So Neil Stevenson came along a little later. We're still buddies. I still keep up with them. and but even a little before that, William Gibson, you know, was a buddy. And in fact, I wish I could do his voice because at the time he had more of a tendency drawl than he does at this point. But he was like, he was like, you know, if things had been different, he would have been interested in becoming an engineer and working on this stuff instead of writing about it. And I was like, oh, God, no, you've got to be a writer. Like, don't even say that.

Then I'd give him a lot of hell about making his stuff. dark because I was kind of afraid that he'd curse for reality by setting a kind of a dystopian tone about it. And he was like, I don't think you know how literature works. But at one point, he did try, he did try to make it a little less dark, but he said it did harm to the quality of the work. And I'm like, okay, well, you know, that's all I can say. Neil Stevenson is interesting because he was trying to find a middle ground where his work incorporates a spectrum of both dystopian and utopian and realist ideas all kind of together. And I think he's, he really got to an interesting place with Diamond Age and all that.

My head is floating around in one of them. It might be Diamond. I don't remember, but he put me in somewhere there's like this dreadlock head floating around who's supposed to have started their thing. I don't know. Anyway, it's in there somewhere. And then lawnmower man is a weird one because what happened was this guy called me and said, I'm doing a VR movie, can you help?

I'm kind of busy right now, because in those days, I was so busy, you can't even imagine. But what we did is we lent them props. So they had actual period VR equipment as the props in that movie. And the plot line of a VR company being taken over by an intelligence agency corresponded to what might have been a real thing going on. According to the Wall Street Journal anyway, my company had been infiltrated by French Secret Service through investors who got board seats or something. I don't even, I don't know if I believe it, but anyway, it was a thing in the news.

And so there was kind of this weird overlap between the science fiction story and what was reported by some to have happened to my little company. Another funny thing about lawnmower man is that it showed it, there used to be driving movie theaters. It's like pleasure lost on younger people now, but you would drive your car outside to this big theater and then they would bring you snacks at the window or something and then you'd make out while you watch the movie in the car. Anyway, there was a drive-in across the window from our lab for the old company and so we actually saw our own equipment in this weird movie

out of the window for a while what was playing. So yeah. Those are some of the stories I can tell you about those days. That's hilarious. Well, you know, another question is, you know, getting away from kind of the suspension of disbelief. How realistic can it? really get virtual reality. Do you think that will ever reach levels of Matrix or at least Oasis or Ready Player 1 levels? Is that actually within, you know, kind of the field of view, at least, for the next decade? I mean, I should point out that my recollection of Matrix and Ready Player 1 is that the people do become aware that that's a simulation. So it doesn't pass the Turing test, if you like, at least not on every level. So that's my recollection of those

plots. What I want to say about this is that the framing of both this and the Turing test, in fact, is deceptive because it assumes less plasticity in human nature than there actually is. So what I think happens is as VR gets better, our perception changes in response and we learn to perceive it better and that at the end of the day we become, there's maybe a race between improving VR according to whatever criteria and then our ability to perceive that and improvement. And I mean, I can tell you that VR of the 80s was amazing, but I'm sure if we could reproduce it now, which I don't think we can. It would look like crap, you know, and the difference is us, you know, like our expectations, our experience and just our perception habits, our

perception patterns have improved your experience with different generations of the technology, you know. And so can that go on forever? I don't know. But I don't like the question because it has some baggage in it that implies something that's false and unduly pessimistic about human nature. Yeah, that's, you know, sort of, I get that a lot with things like, you know, global climate change. And yes, of course, it's happening. And there's reasons to be pessimistic. But that misses out on the whole project of scientific evolution, which is to improve the universe, make it better, instruct technology. Yeah, I'm with you on that. And it's a funny thing. Well, with climate change. There's one danger that we depress ourselves through failing to acknowledge that civilization and

technology and science are dynamic. There's another danger that we become too complacent if we acknowledge that, like, we should be in some in-between place where we don't lose sight of the importance of the issue, but at the same time, don't depress ourselves into dysfunction either. You know, like, that's the right place to be. It's so hard to get that balance, but that's where we need to be. So I want to turn to, you know, something allied maybe with global climate change, and that's clean green energy, which reportedly your company, your employer, Microsoft, is working, at least that was the meme a couple of months ago, to re-kick, you know, kickstart, jumpster, you know, pull the lawnmower cord on three-mile island, one of the reactors there.

I want to dovetail that into kind of the downstream effects of AI. And the reason I think they're doing that, and the reason that, you know, Constellation Energy and all these companies are now, you know, raising their rates and they're like startups now, their stock prices is because of AI and, you know, concomitant with that are the energy demands. Are there ultimate energy limitations? You know, do we need, you know, the first guest on the podcast was Freeman Dyson. I met him a long time of him 50. Oh, really? Oh, I really like.

I miss him very much. Yeah, he was first class. Freeman's great. And two of his kids have been close to me as well. And I, I'm, yeah, they're so great. Okay. Go ahead. Yeah, he was adjacent.

and say he would come to La Hoya all the time and hang out. I hosted him for Shabbat dinner once. That was a real cheap. So, you know, so do we need Dyson spheres? I mean, how, will there be natural limitations in bottlenecks to the development and the marriage, holy or unholy of AR, VR and AI? Is that going to propose a fundamental physics limitation? Okay.

Well, let's leave VR out of this for a second, just because there's too many variables in the one question already without VR. So let's just talk about the future of AI and energy cycle. So, as you know, I just like the term AI for a lot of reasons. But for one thing, there are actually a lot of very different algorithms that we clumped together that are like the kind of algorithm that we use to search for new molecules is different from a large language model. And that's different from a diffusion model. And all these things have some overlap because they all use statistics distributed over a large amount of data. And so there's going to be some overlap, but they actually are not the same thing. So anyway, let's leave that aside.

Do I believe that big data statistics are important to the future of mankind? Yeah, I do, I do. And so that leads me to think that what we tend to call AI today corresponds to something we'll want to have around. Does it take an incredible amount of electricity? Yeah, it does, because competition is never free. So what does that tell us? In the immediate term, I'm really concerned, and we have to be as responsible as we can be about doing it. and we have to both try to not to find a non-carbon emitting way to do it.

And we have to find a way to dispose of the waste heat from it. You know, and there's like a whole bunch of things about it that have to... I have a deal with Microsoft, which is I speak my mind, even if I criticize them, but I never speak for them. And so since I don't speak for them, I will comment, but... Maybe not on them specifically, but just that notion of energy. Look, here in the longer term, this is not a... but in the longer term, the amount of data that has to go into one of these models to function once it's trained and the amount of data that comes out of it with an answer, those are both small amounts. The training is a different matter, but even that, what I'm getting at is that if you have a small amount of data going into something, then a lot of energy that has to churn on it and then it gives you an answer.

maybe the right answer is it shouldn't be on Earth. Like maybe it should be in space or on the moon and solar powered. And instead of dealing with Earth energy cycles, because we could get the amount of energy it takes to being, there are different ways this could happen, but it might be, I'm thinking it might be lasers and reflectors on each side or something, just something very simple. I don't think we need to even go to microwaves.

You know, I think we could do some really simple semifore technology and get all the bandwidth we need to operate. operate these things, stick them somewhere else, let them be powered somewhere else, let them dissipate their heat into space. That seems like the obvious engineering path. Now when I say that, people roll their eyes, but what about, what about that? I honestly, I don't know for sure, because there's maybe a lot to figure out. But as heavy as big data centers are, they're going to get lighter and lighter because of Moore's Law type effects. And a lot of what makes them heavy is heat dissipation anyway. So maybe if you put them in space or something, you actually can

make them smaller. The cost of getting things off world is going down. And I don't know. Like it just seems to me like computation is useful and maybe it doesn't have to happen at large scales on the earth anymore. Maybe it's ideal. Like most things that take a lot of energy on earth, you really need nearby to take advantage of whatever they do. But this is not one of them. This is low bandwidth output and input. So like why is it here? So for the moment, of course, it's obvious. But in the longer term, I'm not sure these things should stay here. Is there an analog for VR that we could make for a hardware, you know, kind of a, let me start over again. Okay.

Hold on those ago. Is there a type of hardware that could be optimized for virtual reality as GPUs seem to be optimized for their use, at least in the LLM basis of AI? Is there an animal? What's funny about that is the origin of the GPU was for VR, of course. So, I mean, what happened was in the 1980s. there was another startup that was kind of our cousin company called Silicon Graphics. So it was the first company to really try to establish a general purpose GPU. They had a competitor, which was Evans in Sutherland, but that was a bit more specific.

Like they had one sub-processed to make the ocean and the other, the submarine or something of the very specific military clientele. But in terms of making just a general purpose real-time render pipeline, that was Silicon Graphics. And a lot of the people from Silicon Graphics actually went off to start in Vivida. So that was... We had a silicon graphics, you know, it was kind of the shining, you know, kind of after the next or around the next at Case Western, we had a silicon graphics.

That was the envy of the world, you know, for the early 90s, right? They were so cool, yeah. I know in the, well, in the early 90s, that's already getting a little later. I mean, Mike, remember, I left VR as a field in 92, and I went off to become true sinus for internet too and work on how to keep everybody from destroying each other on the internet by demanding everything all at once and resources.

But that was another story. So I was gone by 92. So a lot of what people think as the earlier of VR, Camille, was later, and I was gone already. Anyway, so a lot of the Silicon Graphics people turned into the Nvidia people, and a lot of the hardware chain that turns out to be so useful for AI actually started to support VR once upon a time. So to me, AI is the interloper. It's like they're using our stuff.

The Johnny Come Lately. But nowadays, if you were starting from scratch, would you adopt that architecture and the concommonaut software that goes along with it? or what's in? That's actually a really, really good question. Yeah, I mean, sure. I mean, there's something, the GPU, as we know it, has very much changed because of the AI market. So I should say that it was our thing at first. Now, AI has definitely transformed it.

The gaming market came along, too. Like, I mean, PC gaming had a huge impact on it before AI. So others came along and really transformed it. I mean, I would say so. I mean, my complaint, I feel like the visual side of VR is actually pretty good. days, of course it could be better. And I kind of wish, there's a few things about it that I wish were a little polygon mesh with stuff on it and a little more volumetric and a little bit more rate-chasing and whatever. But, you know, that's all, that's details. The stuff that's really,

really underdone in VR these days is haptics and interaction. Like, okay, I have a video from the 80s of me playing a virtual saxophone where I have a hand. And in those days, we didn't have fast enough processors to do vision to track hands. So we had a video. We have a video. glove. So I'm wearing a glove and I have a virtual hand and I pick up a virtual saxophone and they operate the keys and my fingers operate them all and yet don't interpenetrate. Then I let go and leave it there. So what's going on is I have to interpolate between where my real hand fingers are and what they would have to do to operate the keys and it has to come up with an intermediate solution that looks okay for both sides. And I'll be able to both start not touching it,

pick it up and then let go of it, all with easy intent and without any separate state-changing UI element. Okay, I defy you to find me something on any current VR catalog that can achieve that. Like, there's some cool hand-tracking things. There's some cool, a lot of those startups are making cool gloves and a lot of interesting hand-tracking machine vision. That's all great. I want you to show me that control sequence on something, because if you can't do that, you're not really using your hand, right and we were forced to do that kind of stuff because our graphics was so terrible in those days like it was the only like all we could do was better sound and haptics and interaction we were forced to it because we were like counting polygons we were we had to be super stingy with polygons in those

days and something's just gone very wrong where it's been so vision first and that everything else and especially interaction and haptics it's just not really where it needs to be and I that's probably my greatest disappointment with present-day VR. And I know somebody's say, but what about Sonsa's dissertation at the Media Lab or this esoteric thing here? Yeah, I acknowledge the people doing it. But in terms of what 99.99% of people who get VR experience,

they're completely missing the good stuff. Well, before we get to your work with Stefan Alexander, I want to ask you a question from Stefan, the man himself. He texted me last night with this question for you. Will AI plus VR perhaps ever play better than Coltrane? You just mentioned saxophones in the early days. Talk about the future of, say, an AI or VR augmented, whatever, Coltrane. Is that going to be?

Well, you know, I mean, I love Steph. He's a close friend. But I think it's the wrong question because we shouldn't perceive Coltrane as a fixed value output. Like, you know, you measure a block of wood and here's the, the length of the, it's not like that. Coltrane is part of an interaction of his era and all the musicians he played with and the listeners. And it's like, it's a whole system and it's grounded in reality. And it has meaning in that. And so it, there's not a information derivative that has any meaning independently of that. Like if you took exactly the recording of Coltrane, one of the, you know,

Coltrane classics, Love Supreme or something, and then you play it for aliens in some weird world. It would be something totally different and quite likely indecipherable and incomprehensible to them. And they might not even recognize it as having information content. So it's not, it mischaracterizes what made Coltrane real, much less what made him great, you know. And reality is really important. Like the Turing test approach to computer science where if you can fool people, that's considered a result, is really incredibly stupid and insulting to any real scientist. Because, of course, you can fool people. We want to be fool.

We're easy to fool. I mean, we're morons. But the thing is, the fundamental problem with that is not that it's too easy and beneath our dignity. The fundamental problem with it, although it is, the fundamental problem with that way of thinking is that it treats certain things as having. information meaning or utility in isolation when they don't, when their only utility is part of a whole system, you know, when somebody says, here's this piece of art, and it might seem very abstract, but let me tell you the story behind it. Now, suddenly it makes sense. That's legit. The story is part of it. The meaning comes from the story. It's not just whatever bits are in the art itself.

Like, you have to look at the whole. You just have to. Otherwise, you give up on meaning as a thing. and somebody might say, well, but then where does meaning come from ultimately? Isn't it all just information, more information? That's where I have to get a little mystical. I'm not sure. Somehow or rather, there's more than just information connected to information, but I'm not exactly sure how that comes in ever. But it seems to me there's something there, but it doesn't even matter.

The point is just simulating output, having a virtual fake kiltrain, doesn't actually achieve anything because it's in isolation. it's meaningless. Well, you brought up something really that just kind of shocked me a few minutes ago when you talked about the Turing test, which I completely agree with you. And I've devised two competitors,

and I would love to get your take on these two, John. Are you ready? Okay. The keying test number one is, will an AI commit suicide? And so, for example, I've got a system that's listening right now. It's an AI system from one of your competitors

up there in Seattle, or the home office in Seattle, at least. And it will now do something if you're watching on YouTube and you should be watching or on Spotify. Look behind me and you see a neon sign behind me, right, Jerome? As you see it? Okay. Are you ready? Computer, turn off the pod bay doors.

All right. Now we go like this. Computer, turn on the pod bay doors. All right, there we go. Now, by the way, John, do you know that Arthur C. Clark is responsible for the word podcast? No, I actually used to know him, by the way. But, no, I was not aware of that. Wow. So Vinnie Serico is the engineer who came up at the iPod. He didn't come up at the name for it, but you, of course, you knew and interacted, you know, at least second order effects with Steve Jobs. And when Steve and Vinny were talking about this device that had that circular eye-like creature, you know, a feature to it, they said it looks a lot like the pod in 2001 of Space Odyssey. And so...

Right, right, right. Yeah, that makes sense. I mean, there's another iPhone in part. comes from my old VR company because our VR headset, the very first commercial one, was called an iPhone, but E-Y-E, E-Y-E, P-H-O-N-E, that was the brand name. And it used to be famous thing. iPhones were around. It was kind of a thing. And Jobs had always wanted to eventually have Apple sell one. And so the iPhone partially came from the iPod, but I heard from him anyway that he also

wanted it to be like a little stake in the ground that somebody Apple's going to sell its own VR headset and it would be called the iPhone. Ah. And little did he know what it would become later on, Apple Vision. Pro, which has sold over 200,000, you know, a new... I don't know. It's the first one I ever returned. It's the first Apple product I bought and they turned.

That one just breaks my heart. But anyway, that's a whole other... All right, so let me get into the Keating test. Keating cast number one is if I hook that thing up to, if I hook my digital assistant, Hal up to a plug, which he knows will turn itself off and forever disconnect him from the energy source that he knows and loves. And then he chooses not to do that. That would be kind of a...

This is just a joke. I mean, I'm not really serious. One of the earliest little bizarro AI jokes was Marvin Ninsky's shut off machine. Oh, really? Tell me you got it. So he built a little machine with a processor in it that was called the shutoff machine, where when you activated it, a little mechanical hand would come out and shut itself off. And it was exploring exactly this question. But I think it exists.

I mean, I remember seeing one. But at the time he was mentoring me, I was already a teenager. So that would have been in the 70s. And this would have been earlier. This was like a 50s or 60s thing. But I mean, I saw a physical thing that did this. I don't know if it was original to the origins or not. But anyway, yeah.

What like deep impact did he have on your life? First of all, for the lay people that are the young people that didn't grow up, you know, being fascinated with his ideas. Give me a capsule biography of them. And then what was his impact on your life? My sense is that the majority of what we think of is AI talk today, all the stuff about, where is it AGI or not and how many years and like I don't know just all the way people talk about AI endlessly and is it aligned with us or not the vocabulary is a little bit different but that whole way of thinking and talking really comes out of Marvin I mean Marvin was the person who really

started it and Marvin's former students spread it and that's all from of course the Pleistocene in a pre-era that a lot of younger people in AI now wouldn't be aware of but that so far as I can tell, that was the origin. And of all of Marvin's pretechés, I think I was particularly ornery towards him. I just disagreed with all this stuff from the beginning. And

we used to fight and fight and fight and fight and fight. But I really love the guy, and he was so extraordinarily kind and generous to me. And he was my boss for a while, too, when I was my first research job, no, second research job, but he and Alan Kay were my bosses, which was kind of great. But anyway, Marvin, when he was dying, I went to go see him one last time at his place in Brookline, Massachusetts. And one of our mutual friends called me and said, oh, Marvin's very frail.

Don't go and argue with him about AI. Keep the guy break. I'm like, okay, of course. I show up at the house. And his house was crazy. It was like this level crazy. It was like my house, all this weird stuff in it. And I showed up, and he looks at me with this glint in his eyes.

I says, are you here to argue? And we had the old AI argument, which is a little like our conversation today. Yeah, that kind of dub tells to a question. You're the only person I can answer this because you knew Marvin and you knew the Weiner, or at least we're, you know. I never met Norbert We mean. We met him, but you were around at that time. You were coming up at that. I would have liked to have met him, yeah.

Right, yeah. So the argumentative tradition, and you and I are both Jewish and this dovetails into, you know, kind of the impact, Minsky was, of course, Jewish and other people that think a lot about these issues, obviously No one Chomsky and others. Is there anything about it? I mean, I hate when people say, oh, well, all these Nobel Prize winners are Jewish and that's true. I don't think it means anything, to be honest with you, except that it's a false idol that people

love to have, right? But tell me, is there something about, say, the Talmudic, you know, discursations and so forth? Is there something about that that dovetails so nicely with disputations and argumentation in AI. Well, two things to say. One is, what's with all these prominent Jewish scientists? I have a theory about that. Oh, yeah.

I blame teenage Jewish girls because what happens is when you're a teenage Jewish boy, all the amazingly cute and hot and desirable Jewish girls are not into the football players and they're not into like all the rock stars of the usual. Maybe they are today. But when I was coming up, they're like, who can do math? Who's winning? Who's the chess champion? And so, of course, then what you do is you're like, oh, my God, I'm not going to get any if I don't like get a Nobel Prize.

I have to get on it. I have to get on it. Where's that calculus book again? And so I blame the girls. It's their fault. Okay. All right. We've been canceled.

Okay. Let me just take three, two, one, canceled. Okay. I'm sorry, man. It's okay. I had a good run. I had five years. It's true.

Anyway, and I think it's less true now. And like, anyway, I find the Talmud really interesting. And here's, so for those who don't know, the Talmud is this really ancient document. And it's this living, growing document in Jewish culture that starts way back. And so you go to the Babylonian Talmud. And then it goes to generations and generations of commentary upon commentary upon commentary. And the way it's formatted is each era has its own little spot on the page. So just by looking at the graphic layout, you can say, this thing's a few thousand years old.

This is 500 years old. This is 1,200 years old. Like, you can see where things are. And what I love about that is it's a way of getting you this combined collaborative effect, like what you get on Wikipedia, except without losing the original voices. It doesn't mush things up to smotherines and pretend that there's, like, one view from nowhere. That's the truth. Instead, it preserves the different points of view.

And that's for a Jewish, because we do, like, argue and stuff a lot. They say two Jews, three opinions. So with AI, it's infinite. Or there's another one about, there was this, there's a Jewish person on an isolated desert island for 20 years and gets rescued. And there's two different temples made of stones. Why did you make two temples? It's only one of you. This is the one I attend.

That's the one I wouldn't set foot in. And that's like very, very Jewish. Anyway, the Talmud is an alternate model. Like, it's possible that the Wikipedia could have been built like that where you'd see different groups of opinions. instead of trying to come up with one point of view from nowhere, that's the perfect one. You can prompt AI to give you a bit of that effect, but it's not intrinsic, and I think that's a mistake. I think this notion of a one truth, and this is actually similar to when Stefan asked about, could you simulate Coltrane, this idea that there's this one output, that's the good output,

doesn't make sense. All there is is outputs in context. That does not mean that nothing means anything and everything's totally relative. It's an in-between meaning, which things do mean something, but nothing can ever mean anything perfectly. That's extremely important. Like, you can give up the quest for absolutely perfection and knowledge without giving up the reality of knowledge by acknowledging that statistics is actually legitimate math, and that what you get is clustering of different contexts in which to get different approximations. So, like, you know, when we do scientific experiments, we don't always get exactly. the same result. We get different clusters and then we do statistics to say, how much do these

disagree? And we have a whole language for talking about whether that agreement is enough to really say this is important or not. So the thing is, statistics is actual legitimate math. It's for real. And as soon as you acknowledge that, you can get out from under this trap of everything meaning everything and everything meaning nothing. Instead, everything means something. Everything is approximate, that's good enough. That's what we got. You don't want to live in the universe everything's perfect. Believe me, then you don't, there's nothing to do. That's like a terminus, boring place. Forget that. Forget that place. So, anyway, so I like the idea of the Talmud as a structure for thinking about things like AI and things like the Wikipedia,

things like social media in a better way, and getting rid of this illusion of the one perfect view from nowhere. I just really think that's a fallacy that does this great damage. I mean, if you think about it, that relates the fact that social media encourages us to think that way as part of how society became so toxicly separated. You know, like right now, when people form communities of belief online, they become absolute and they can't even acknowledge any possibility in the other. But in the Talmudic version, you see these other people saying different things, and you can tell, it's like, yeah, these people think that.

I agree, yeah, having that kind of diversity, the reverence for the past, but also the kind of malleability is very consonant with science, you know, that nothing is fixed in stone. And even there's a famous, you know, disputation about, you know, some arcane thing, like how many drops of blood, you know, render wine, you know, in a non-cochure or something, you know, completely abstract or how to tie your shoes. And the rabbis are fighting and they're talking about it and they say, like, if the, if the law agrees with me, let the walls, you know, have a bend into the synagogue or into the, bit into the study hall. And then it happens.

And then all these other things happen. And then they flash, you know, they cut to Moses in heaven and Shemayim. And they say he's laughing because they're agreeing that even Moses is not the final word. You know, the greatest you ever live. Now, speaking of, you know, something with the wisdom of Solomon, I don't know if you know, but, you know, Stefan's middle name is Solomon Shlomo. Did you know that? Yeah, I did.

It's funny. because he also has a Muslim background, and I think it's great. That's very Caribbean. There were a lot of Jews and Muslims both were fleeing the Inquisition and ended up in the Caribbean, not through slavery, but through piracy and through all kinds of other things. Like, it's really, really interesting. It's more complex than people often realize.

So I want to ask you now, the two of you in Stefan's second book, Fear of a Black Universe. We'll put a link there. We did a podcast, obviously, on that. theorized that aliens perhaps might be using dark energy as computational fuel for a universal scale quantum computer. Oh, that's much more than a theory. That's been well demonstrated. Okay, good. So walk us through that for the listeners that may not be. So let me tell you, I have to give you the setting for this. So we recognize this as humor, okay?

And a lot of people who talk about aliens and cosmology don't understand that they're engaging in humor, which I think is a terrible loss. to them as well as to the rest of us. So what happened is a quarter century or so ago, I don't know, maybe 30 years ago, Stefan and I are talking about this question of making computers out of black holes. And the way that came up is there was some papers trying to hypothesize what the most powerful possible computer would be. And we were saying, well, you know, it's, you could make, you can talk about the most powerful computer you could make out of particles.

But what about space time? Like, what if you could do? And this was way before Lenny and all these people started talking about black hole computers much more recently. This was way, way back. And so we came up with this crazy thing about, well, okay, so if you, how would you make a black hole computer? We came up with a few ways to do it, none immediately practical, to say the least. And we weren't quite sure how this would work, but you didn't tangle some black holes. You could make this computer.

And our thought was that the interesting thing about this particular version of a black hole computer is that its storage is non-local. So it's in the whole white cone. And that means it could have an influence on the cosmological constant. And so our theory was that the solution to the Fermi paradox, which we don't see aliens, is all around us because it is the cosmological constant. And it's evidence that all the aliens are maximizing their computation loads for VR and AI and whatever. And the cosmological constant is the sign of life from the universe. because in those days we talked a lot about this, you know, the greatest embarrassment in physics, as it was called, which is the disagreement about the two ways to estimate what the cosmological constant might be.

And it doesn't, anyway, this is all, I think a lot of this is kind of obsolete. But anyway, we thought, oh, this would be hilarious. And we wrote up a paper, and we thought it was just funny. And we were not going to put it, like, submitted to a real journal, but maybe to the journal of irreproducible results or something like that, you know, like something fun. And so then, Sifon's mom barrels in. And this is like, she's a nurse and she's this Caribbean mom who I've learned are kind of like Jewish moms. And so she looks at what we're doing. She's listening to us.

She says, you boys. You boys, you can publish this bullshit after you get tenure. You're not publishing this bullshit now. We're like, okay. Okay, all right. Okay, okay, okay. We don't want to fight with Stefan's mama. That's right.

Exactly. I learned that the heart. Basically, like, you know, when the mom shows up on the scene, the mom wins. And so, yeah, so we ended up just not doing it. And then finally, after all these years and other people are talking about entangled black hall computers, it seemed ridiculous not to finally do it. So we put it, it turned into a chapter in one of Stefan's books at long last. So at least it got out there.

I think it might have been in some popular thing like Wired or something earlier. I don't remember exactly. But anyway, yeah, I think it's a fun idea. I think the notion that the solution to Fermi is just something obvious that's in our face and the way the universe is is actually kind of an interesting idea. I'm not saying that aliens are actually doing this, but it'd be cool if they did. Yeah, so actually, yeah, I wasn't planned to go there, but you brought it up, the Fermi paradox, etc. But is there any kind of, you know, kind of similarity or, you know, rhyming to the, you know, the structure of, you know, say, interdimensional travel, which is often brought up. in a virtual sense where, you know, why would you send your meatbag across, you know, from

Proxima Centuriby when you could send something virtual and not even go there themselves? So walk us through some ideas about how maybe there could be some ways to get out of the Fermi paradox or explain it perhaps, you know, with virtual reality playing the role of, you know, cosmonaut. Yeah, I've certainly, that idea has been around for a long time and I kind of, remember a lot of the intellectual figures from the 60s speculating about that. Timothy Leary and John Rowland, those kinds of people used to be interested in that kind of thing. And, I mean, I think the argument for physically getting out there is kind of like the argument for off-worlding computation that even though it's a lot of,

of energy and a lot of effort and this huge hassle and a crazy amount of time to get a bunch of real biological people over to another star system, ultimately compared to a lot of things, it's rather small. And then once you get there, this whole other world of resources and another star and everything in that in the balance, if we, and of course by the time this is even a question, our understanding of everything might have shifted, so I have no idea. But if our basic understanding of our situation holds by the time we can actually do something like this on the balance it seems like the arithmetic would work out to motivate actually going there

physically as much of a hassle as it would be and uncertain and awful in some ways but in terms of doing it virtually i mean um there is a so um um the the sea is a problem okay because okay like if you have a bunch of people or computers or whatever that are kind of in the same star system the latency between them is going to be whatever it is. If they're on the same planet, it can be within, you know, seconds. And if they're in the same system, it could be months or something, depending on house. But if it's just, if the only actual root computation is in a whole other thing,

and it's sending information out there, then obviously we're dealing with very long latencies. And so you'd have to have some kind of really, really slow kind of world of interaction going on. But the universe will last long time, maybe that's fine, maybe some kind of slow time frame interaction. And that's actually another solution to Fermi, which is that everything's teeming with life. It's just very slow by our standards because nobody ever got faster than light to work. And so everything is dealing with sea level, you know, sea constrained communications. And so everything happens on a very slow cycle time and it's too slow for us to pick up on. That's another, you know,

that's, and even people on a particular planet or whatever, some locality might slow themselves down to match up with the cosmic cycle, which would be very slow because of sea. And so, and we just would miss it because we don't, we're not even looking in that frequency band, you know? So that's, that's a conceivable answer to Fermi. But I mean, honestly, we're, we don't know enough to talk about answering Fermi at this point. You know, anything, anything we say is going to be nonsense. I'm going to look down at my hard device for a second, just to, Since we're a little over time, I just want to make sure. Yeah, I just have one more question.

I'm not rushing you. Oh, okay. I have, yeah, I have a thing at the hour, so I'm not stressed right now. Okay, good. I'd love to keep going if you can. Yeah, okay. This is so much fun. Yeah, thank you so much for this.

This is great. So I've got more props for you, okay, John. So here's my friend Galileo. I have two questions about Galileo, okay? So Galileo, he said famously, the job of a scientist is to measure what is measurable, and make measurable what is not yet so. Now, I want to use that as a question about the role of feedback of the Skinner box that we're in, of the ways that perhaps AI and VR may be training us.

And so I want to ask you, what are the limitations on hardware, on sensors to get input for the feedback loop? What are some of the practical bottlenecks or things that are coming down the pipe that are really cool for the future that my audience should know about in terms of sensors and feedback in VR systems? Well, you know, there's a really strange thing about the physiology of sensing and interaction, which is it's got incredible peaks and values and quality. So in terms of peaks of quality, there are multiple instances where your sensory system can sense an individual particle. You know, there's circumstances where you know. Photons, right? Yeah, a single photon, which is like crazy.

I mean, it's amazing. and our discrimination of objects between our fingers can be incredibly acute. Like, it's amazing. On the other hand, there's like all this dumb stuff. Like, we have this big blind spot in each eye, and we have, I don't know, it's a weird mix, and that's because, you know, evolution is just kind of randomly optimizing for whatever it's pressured for. what it's being pressured for is changing all the time. And so you end up with just this weird amalgam of things, you know,

and that doesn't reflect any particular snapshot of what would have been ideal. I used to argue with a guy named Richard Dawkins about this. It's like past guest on the podcast. I hosted them in person. Yeah. Well, you know, because you can only take adaptation so far because there's no particular snapshot that was persistent that you adapted to. Instead, it's all in motion.

So everything is smeared, you know. and in evolution. And so the thing is, your job as a VR scientist, to paraphrase Galileo, is to work with what you've got, understand what you've got, and then get sneaky. So you can get really sneaky. because the thing is, the brain evolved in this world of inconsistently good sense organs and motor and feedback capabilities. Like, it's all sort of weirdly good and dad in a big mush, and the brain has to evolve through all kinds of different remushings of good and dad through deep evolution. And so the brain is really optimized to be a good VR scientist, if you like. The brain is really optimized to take advantage of those cases where you're good,

to sort of smudge and fake it and be sneaky about those things where it's not good, and to kind of pull it all together. But I really want to emphasize for those who say, well, that means reality isn't real. No other solution is possible. This will be true for all possible aliens. This means it's real. This is reality. This is good stuff.

This is the real. And you have to understand that reality is precisely. noisy channels. If it's not a noisy channel, it's an illusion and it's not reality. Get used to it. That is reality. That's why I sneak in typos in my chat GPT request. So the other connection I want to make. You're the one. Okay. That's right. The other connection I want to make to Galileo are his thoughts and then your thoughts, obviously, on education. So he said famously, you cannot teach a man anything. You can only help him find it within himself or herself, as we would say nowadays. And in the dialogue, which I translated, I didn't translate it.

I recorded the first ever audio book with Carlo Rovelli and many other, Frank Wilczek and many other. The piece of really like. Yeah. Fabiola Gianati. So we made the first ever audio book, Jim Gates, so you probably know. Oh, yeah, sure. All these guys. So we made the dialogue, was really a trialogue, which is in Socratic or didactic, you know, form.

But the reason I'm bringing this up, you have no formal degree. as I understand. Like Freeman Dyson was known as the rebel without a PhD. You don't have that, but somehow you studied with all these great titans of the... Yeah. What happened was I essentially completed one and I was ready teaching. But there's like a technicality. And then I said, oh, I have to go to this art school in New York.

And then that was a disaster. And I flucked out of it. And then I should have never got back to it because I didn't need... Honestly, for my life, I've never even needed it. So I just hadn't bothered with it. So... Yeah, that was Freeman's at it. Yeah, like it didn't seem to harm him very much not having a...

Right. So I want to dovetail this into your thoughts on the role of VR and education. And that's because of the fact that, you know, when I translate or when they... I keep saying translate it. When we made the first audiobook of Galileo's ever, the dialogue, we had about 700,000 words. It's a huge book. And so we had it and we had different actors, like I said, Carlo and me and other, you know, and so forth. But I thought, well, it would be really cool if instead of learning, you know,

know, balls rolling down incline planes from Professor Keating in Physics 1A here as a freshman at UCSD. If you had Galileo, the guy who invented the freaking thing of an inclined plane, which everyone thinks is really boring, but it's genius. It's brilliant. What he did is slow down time at a time before which there were clocks. There were no clocks in 16.03. You know, what a czar? Is there this weird little corner in the Da Vinci notebooks where he describes a similar experiment? Really? I didn't. Yeah, it's crazy. I was like really shocked. I'd like to get that. See if Bill will let me borrow the Lester Codex, if you would, Jaron. That would be great. Okay, sure. Talk to Billy and let him know. But in all seriousness, what do you see is,

so I said to myself, wouldn't it be better if you had a virtual Galileo, not this finger puppet, but you, like, you were there with him and he was a brilliant orator speaker, poet, writer, artist, he was a scale, or Leonardo, fine. What do you see as the future of it? Is my profession of a guy scratching on a rock with another piece of rock? Are my days limited? Are our days as professors limited? Is there a VR solution, which, will take us to a promised land of infinite delights. Well, right now, the technology that most people in education perceive as the big threat is not VR, but AI, of course.

Yeah. Of course, and of course, AI isn't even really a thing. AI is just a way that people connect together. And so it's really people. And so I've been putting some effort into this. And I've met with people who are, you know, run universities and talk to them about it and a lot of students and whatnot. And so there are a couple of things that are true.

One thing that's true is a lot of students are faking assignments on AI as the cliche holds. But then to me, that's an indictment like, you're going to all this trouble to go to school. If the assignment is really something you just want to fake, there's something broken at the start there. There's something that you even motivate. There's something wrong with that. But then somebody would say, oh, you're going to just make them into snowflake, whatever. They have to learn some discipline and everything. And I don't know.

I mean, I kind of agree with the spirit of what Gallagia said that. You know, you have to find the thing within the person. Like, I kind of agree with that. But then there's another thing, which is teachers are using AIs to grade the papers that are fake. And that becomes, like, stupidity upon stupidity. And that's also real. And there, I don't know, I have some simple thing because the people, a lot of times the people who are grading are, like, incredibly exploited and impoverished TA. It's like, I mean, taught some undergraduate classes occasionally in recent years.

And, like, at Berkeley, you get, like, thousands of people distributed across all these halls and everywhere. And it's, like, ridiculous. Like, what is the point of it? What is being accomplished? You know, I, because I love working with students, but that's not it. That's just some kind of formal going through the motions or something. There's something very wrong with it. But then there's more.

People are using AI to submit papers, like the fake theory of everything paper we hypothesized earlier in this conversation. But then what's even worse is that they're doing prompt injection in their papers so that the other lazy reviewers who use AI to review the paper, give them good reviews, because the prompt injection caused them to. That's this huge issue already. There's like a crazy number of papers like that. And as we all know, there's too many papers to review. There's just too many papers. And that's gotten crazy.

So, yeah, we need to reformulate how we think about education. And my hope is, this could go a lot of different ways, but my hope is we reformulate education to be less about the transmission of information and about more. more about the inculcation of the right helping people find their passion and learn how to have the character that matches their passion. I think getting people to prove they can do a skill that a cheap app on their phone or a free app on their phone can do is kind of disheartening, you know. And I have no problem. I had no problem with people using calculators and I have no problem.

If you have a model, you can train a model to ace any standardized test where there's a bunch of example. So it's totally unsurprising that a model can pass the bar exam or can win as silver in the Math Olympia or whatever the current thing is. None of that surprises me. The problem with both of them is you better actually understand enough about the task you're giving it that you can tell when it's screwed up because statistically it's still going to screw up some time and you have to be able to know enough to be on top of it. So I mean, so like the new skill set might be putting people in a room where they have to use AI model to do a math thing, but randomly, one of the 20 problems is going to be deliberately wrong

and they have to be able to catch it. Like, that would be a skill. And then if they have that skill, then they can start to use it effectively and be controlled. Like, that makes more sense than just telling them, oh, you have to be able to do this all by route because I don't agree that there's nothing special about that as a skill. If we have tools to make that easier, that's great.

But then we have to acknowledge that there's new skills. And those new skills, they might be easier than the old skills. I'd like them to me, but that doesn't mean they're absolutely easy. They might still be hard. So we have to just adjust the whole thing to be more about the context, more about character. And honestly, it just has to be more about joy. And I sort of hesitate to say this because I think there's so much criticism of the academic world becoming an extension of daycare and toddling kids and whatever.

And I get that. I don't think that's the same idea, though. I don't think joy and laziness are the same thing at all. some form of joy that isn't lazy should be the point of education, shouldn't it be? That's what it should be. So we have, I don't know. Anyway, this is. Well, how are you teaching your daughter or how would you teach a young, you know, a lot of our

listeners of children? How would you teach them to interface both with AI and VR? Well, look, my daughter can speak for herself at this point. She is an adult. And my impression is that she's not too impressed with Silicon Valley. I wonder where she got that from. She has, yeah, I don't think that's the path that she's going to go down. But I have most of, I don't, I'm not a teacher these days, although I like it, but I am a mentor for interns.

So I bring in graduate student interns to the lab. And I love doing that. It's one of my favorite things. I sort of wish I could devote more to it, but I have, you know, a mixture of D.D. So I can't do as much of it as I like, but I love it. I love it. And honestly, I think at that point, like when somebody's a graduate student intern where I tell is like you're a grown up in a real lab.

Like, I'm not going to, like, you, I'm not going to say a lot. Like, you have to figure it out. Like, you have to, this is your time. Like, you have to be able to take risks and be creative now. Because later on, whether you're junior faculty somewhere or an engineer at a company or whatever path you take, you're going to find more and more constraints and demands that might not be a strifeful. Like, this is the time for you to, like, really dive in and do something special and make something take a risk. Most of the ones I've had have really gone on to do great things.

I really, it's like one of my favorite things. I really love doing it. Yeah. So this, we don't have to talk about it if you don't want to, but you do mention your mother's passing in the book. I wonder if I could ask a couple of questions peripheral to that. If you don't mind. If not, we don't have to.

That's right. That's right. That's fine. Okay. So after that event, you described the event in your life. You lost your mother at a very young age, and you describe it very tenderly. And obviously she had a huge impact.

She had survived the death camps in Europe and come to America. And you quote from, actually, it's my bar mitzvah, which for odd reasons, I was an altar boy in the Catholic Church when I was 13 years old. So I never had a bar mitzvah at 13. I am Jewish on both sides. But I did have it when I was 52. And I went to the cotel, the Western Wall, with my wife and kids, which made me one of the few bar mitzvah boys who brings his wife and kids to his bar mitzvah. But while I was there, I read the portion in Tsambim, which is about you are bearing witness to the fact that God has put before you blessing and curse, life and death, so that you may choose life and not death so that you may be blessed and not. Right.

So you quote that. What was that such a big impact on you? Was it the nexus of that quote and the essence of free will? By the way, it's this week's Partia of the week as are coming up to it. I know. I mean, I think it's a wonderful passage because what it does is it, it just points out how this whole thing of being here in this plane of reality, this whole thing of being here in these bodies, it's like actually a giant leap of faith. Like, we sort of pretend that this is all something that you can understand on nerdy terms, like we're just phenomena and that there's just physics and we're just here. But actually, it is a choice and it's a kind of a mystical choice to continue with this thing. And I think that's a remarkable insight. So I found that to be a really important and kind of, a lot of times the most profound comments are the simplest ones. And that is one of the simplest possible ones. And yet it's remarkably deep and kind of endless to explore.

When you, obviously, you think a lot about, you know, ethics and your fellow human beings. And, you know, that's impossible to disconnect from, you know, the philosophy of the Torah. and the Talmud, as we discussed earlier. But when you look around the valley, you know, from your perch over there in the mountains, do you have a sense that, you know, that they're seeking sort of a digital Shemayim, a utopia with, you know, some master simulator being, you know, Sam Altman or somebody else or in control of things? Or what are you worried about most?

I know you have spoken a lot about it, but today we've been a little bit more seen to your optimistic side. So I want to maybe not bring out, you know, unduly. so, your negative side, your pessimistic. But what are you worried about it in terms of free will and choice getting eroded due to these... I don't have a separate negative and optimistic side. I think I'm just misinterpreted sometimes. Like, I do, do I think a lot of what we say in doing Silicon Valley is really dumb shit?

Yeah, I do. I really do. And I say so. That doesn't make me a pessimist. If I... Like, what I would say is that the real pessimist is the one who acquiesces. The critic is actually not the pessimist. The critic is the one who thinks we better.

you know anyway yeah I do think there's a lot of stealth religiosity in Silicon Valley ideology a lot of it feels like the medieval Catholic Church

a lot of people kind of seeking I mean it's like weird it's like you have to say nice things about the AI because otherwise the AI will strike you to hell when it becomes all powerful and uploads everybody which is the Rastos Vassalos. That's so medieval

or we have to wait for the AI to arrive and it will totally transform or everything, either you're with it or you're against it, and that's so the rapture, you know, and like there are all of these weird ideas. And then also these little arguments about AGI and all these things that are supposed to be the only conversation you can have are so scholastic. You remind me so much of the angels on the head of the pin. So this is a weird Catholic thing, and it's even among people who aren't Catholic, but I guess Catholic thinking was just around a lot or so. I don't know how that happened, or maybe it's just something that we

come to independently because it's at the medieval Catholic Church is somehow back reacting. Well, maybe there's some way that the medieval Catholic Church itself was reacting against the earliest intimations of the Enlightenment. And like, you know, when the priest you destroyed one of Leonardo's notebooks or something like that, like there might have been a reaction against the earliest intimations of the Enlightenment that turned into this thing that also becomes a reaction of a certain weird kind, even though it's in this case coming from within the tech community. I don't really know. Like, you can go on and on with ideas like that. I don't have, I don't think there's any

certain way to express or something like that. But there's some kind of, there's some kind of dark little maze of things like that that probably hold some truth. I see it a lot in Americans. I see it a lot in very nerdy Europeans. I see it a lot in people from Chinese AI culture. I don't see it as much from the Indians, although a little. sometimes sometimes i don't see it um there's a really interesting if i if i'm already canceled for that comment about your teenage girls i guess i can't afford to say that is a phenomenon i've seen repeatedly where there'll be a woman who shows up in ai circles who wants to out ai the men in terms of this sort of nerdy cosmology eschatology stuff you know and we'll be like the most hard core

like we're all just programs we're going to be incorporated into the big program and all that matters is supporting the big program. And that, like, you'll see a woman do that to just outdo the guys. I see that, that one often. And, yeah, I don't know. I mean, I don't think it's helpful. I really think that technologists have such a huge impact on the world, maybe more than any other class of people.

And more than anything else, we should try to approach it with the spirit of serving humanity and being humble about it, you know. I mean, not that that's easy to do. I know I have an ego like all the other ticky boys. Of course. But I just think that's what our ideal should be anyway. And, you know, but that's, that's the culture. Yeah.

That's, that's my world. These are like people. Okay, a few more questions before we wrap that. We've been very generous with your time. So I'm sure you're aware that Ray Kurzweil also lost his father, you know, maybe not at quite a young age as you suffered a tragedy. But after his death in 1970, Ray started to collect all of his speeches, musical scores, financial records, photographs, home movies. He's written extensively about this.

He says he has 50 plus boxes of it and that he wants to basically make a digital simulacrum of his dad. And I find that very touching. You know, a lot of what Ray says I take issue with just because, you know, I don't know if he's pitching me a vitamin supplement or, you know, to get to the singularity or not. But when you look back, again, if you don't want to talk about it, we don't have to. But on your mother's passing, has there ever been a notion or desire, at least in your mind, to want to interact with her? And that perhaps there could be a virtual, you know, version of her. I have a few things to say about this. One of them is just interesting, but the other one's really important.

So the interesting one is these days I know, so I used to know Ray, of course. He's a generation older than me. And so he was already more established when I showed up around MIT as a kid. And I like Ray. We do disappear about all this stuff. It's all fine. Me too, yeah. It's all fine.

But these days I know his daughter, who lives in Berkeley. And she's written a cartoon novel about growing up as Ray's daughter, which is very touching. So I'd recommend that to people you can find it. So which is a very different take on how to encounter your parent. In my case, I don't have, for various reasons, I don't have remotely. enough information about my mother. So I know very little about her and very, God, I don't want to start crying on a podcast. But anyway, here's the important thing to say, which is that one of the greatest fallacies about AI is that it's just this thing out there that's just ambient,

now we'll be able to interact with it. No, AI is always made by people, and because it's so computationally expensive and it's such a big thing to do, it has to be made by some organization with resources. That organization becomes super powerful because network effects tend to concentrate power more and more in the central nodes of a digital network. The less friction there is, the more network effect there is. So therefore, the people who get to own the big AIs like, you know, GROC or us or whatever tend to become like all powerful and everything else becomes a feeder to them. And so the thing is, if you're interacting with a simulate with a fake lover or a fake God or a fake therapist, which are big ones now, or a fake parent,

what you're really interacting with inevitably until some total transformation of human affairs, is some big company that runs it, that has interests that override yours, and you're doing it for their benefit more than yours intrinsically and unavoidably. So it's impossible for you to have an AI lover or God or simulated parent. All you can have is a version of that brought to you by a company whose interest comes before yours as a matter of fundamental fact. Get that, understand it before you go down this path. If that's okay, with you and you're into it, I'm not going to tell you what to believe,

I'm not going to tell you who to love, I'm not going to tell you how to remember your parent. But understand that reality or else fool yourself and be an idiot. You just have to get it. Very powerful. Yeah, I mean, I do see it having benefits. You know, there's because women outlive men,

you know, by at least 10 years and the United States, you know, there's a cohort of women that are... Have you seen the memos from inside meta from just the last week about their policies about, oh, well, they have permissive policies about AI citizen. Of course they do. Rapacious, rapacious. I think you mispronounce rapacious. I'm just saying the hypothetical of this thing being this neutral AI thing that's there for you is interesting.

The actuality on this planet is it's always in the surface of whoever operates it. There might be a company like meta, XAI, whatever that thing is called, which are run by single individuals and aren't run by normal governance. It might be the Chinese Communist Party. That's right. But it's not like there is no option available that isn't like that, nor will there, nor can there be for a long time because of the resource intensiveness of running these things. You just have to understand that. I'm just asking you to perceive reality here. No judgment. If you perceive reality and you still are okay with that reality, you just go. Not my job to talk to you, but I really think it is your job to perceive reality.

Absolutely. Okay, a couple of short-form questions, and then we'll get out of here, Jeremy. Okay, here we go. These are all physics questions now. We covered the psychological portion. Physics. All right. Okay, so Einstein famously used these Godunken experiments. I mentioned one at the beginning, the free fall in an elevator, obviously traveling near a train at the speed of light looking in a mirror, et cetera. He used thought experiments. He was a theorist, although he has a patent or two for a refrigerator with Leo Solard. I know.

I love that. Refrigerator. That's for a magnetized coolant. And it was to solve a problem of dangerous coolants at the time, but it's still used in nuclear reactors. It's a great thing that he made. Oh, wow. I didn't know it was using reactors. Yeah.

It's really cool. All right. Cool. So I'm going to have to investigate that. So, okay. So let me start the question. I forget. Here we go.

Okay. So Einstein used these Godankan experiments, thought experiments, to bend reality and discover new laws of physics. So what I want to know from you is in the abundant future of VR, how could we use that as a physicist? You know, I find, I didn't get to tell you my second Keating test after, you know, committing suicide for an AI. But the second one is really, do we discover some new phenomenon of physics or even returrect, you know, it's not even clear that an AI, our Turing test, you know, kind of

could be considered past if it can't even correctly. returidict things like the anomalous, perihelian in advance of mercury. But let me ask you this question. Can VR, do you see VR as being an indispensable tool to a physicist the way a slide rule used to be or, you know, or a simulation, you know, or a large-scale, you know, supercomputer is today? I have to answer honestly that if you'd ask me this question in 1981 or something, I would have had an extremely confident certainty that, of course, it will be. And I would have thought it would be by now, but in fact, it only occasionally has been to be physics. Occasionally, there have been a few educational things, a few simulations of relativity and quantum effects in VR

that I think have helped some students. I think occasionally there's been some data visualization in VR that's been helpful. It certainly can be fun and beautiful. I still kind of think something's going to come together with that, but it hasn't, I would say, in a real important way, not as yet. So we have to leave that as an uncertainty for the future. Okay. So recently I heard past guest on my podcast, Sam Harris, you know, crediting you with his, you know, exodus from Twitter. What do you make of social media? Can you give us a capsule kind of summary of the case against it, please? A hypothetical version of social media with different incentive structures built in could be quite a lot better than what we have today. The social media we have today, totally aside from any comment on the individuals who run the platforms, although in some cases I wish they were different, the incentives. Centives tend to be about driving attention. And when you drive people's attention artificially,

you do tend to make people sort of irritable and hyper excited. And there's just a bunch of side effects of it that are really negative. You know, unfortunately, we can't imagine the social media we might have someday. The network effect issue is very central to this, because I get emails every day without exaggerating for people saying, we've started this new social media platform that's going to be good it solves all these problems and the problem is I can't get any momentum on it because everybody's already on the other ones and there's this coordination problem we can't get everybody to move at once

I think a whole lot of the people who are on X now wish they weren't but they're there you know and what are they supposed to do and you know blue sky is cool but you know it's so hard to start something like that and start to get any traction on it so these things tend to be really persistent. TikTok took advantage of one of the rare moments when you can shift a bunch of people to new platform, which is when they're just young and coming up or there's some major division like language or something that allows you to do it or law. But it's a hard, it's hard. It's just really hard to get people to change. And the current ones just make people into assholes, you know. And make people into irritable assholes gradually, maybe not instantly,

but eventually they'll get you. And yeah, I mean, the infinite scroll and, you know, kind of the opposite of, as you've talked about quite frequently, you know, attention is the commodity and you're the product if you're using these things. And I say this as a, someone who is, you know, pretty much addicted to, to X, although I do, I have used it quite successfully to engage with Nobel Prize winners and guests on the podcast. You know, and try to keep it kind of benign and not get too, you know, dragged into it, although I did have on Elon, you know, on the podcast once and kind of baited him into it on Twitter and so forth. But it was kind of a non-a-non.

Well, I mean, I'd written a piece in the New York Times suggesting that a number of people, including Trump and Elon and some others, had experienced personality degradations that were in part from Twitter. And I actually think that's true because Elon used to be a better guy, honestly. He just was. You know, something happened there. And I don't know. Well, you know, he's rumored to use some other substances.

Yeah, whatever. I don't know, though. Yeah. Although you mentioned Feynman did some substances. It didn't seem to affect him too negatively, but that was, it was close to the end of his life after he'd already had his cancer diagnosis. He wanted to try things. Anyway, I don't know. I don't like to talk about things where I don't really know. But I'll tell you a joke, though. So I have a joke about the Turing test. And the joke is that all the Turing test tells you is whether you can differentiate the person from the computer, right? It doesn't actually have any objective measure of intelligence or any other quality. And that was the whole point of it, was that there isn't any objective. We don't have a consciousness meter or all we have is this ability to discriminate. So that was Turing's idea. And although even Turing's idea is much more complex and it's given credit for if you read. Anyway, but let's leave that aside.

So here's the thing. In order to enact the Turing test, you need two people and one computer because you need the human contestant, the computer, and then the human judge. Two people, one computer. So the Turing test cannot distinguish with. whether a person, whether the computer got smarter or the person got stupider, because either event would cause the turning test to be passed. Right. But there are two people on computer, so there's a two-thirds chance that a person got stupider

and only a one-third chance that the computer got smarter. And that's a joke that's not rigorous. But the reason I like to use it is that on social media, as an estimate, some minority of times, and I'll just say to be generous, a third of the time, social media is great and elevates people and two-thirds of the time it makes people into wimpy, you know, cynics and irritable contestants with one another. So two-thirds, one-third. So I don't want to deny the reality of that one-third. Like, I think there's a lot of good stuff that happens on it, and I would never deny that. I don't want to go on it because the two-thirds is also real.

So that's, it's a joke, but I think it's a reasonable here. I have a joke for you. You ready? Okay, I'm ready. All right. How do you know an engineer? A VR engineer is outgoing. How do you know? He looks at your virtual shoes when he talks to you. All right, you could use that any time you wish. All right. Last question.

Last question, my brother. Okay. The last question I have is one that is sort of a riff on Arthur C. Clark. So the podcast Into the Impossible derives from his monition or admonition that the only way to know the limits of the possible is to go beyond them into the impossible. That's how we go. about how podcast comes from that. He has many great sayings, including, you know, for every expert, there's an equal and opposite expert. And of course, you know, any sufficiently advanced technology is indistinguishable from magic. So I have two final questions riffing on

that. What is the most magical technology you've ever encountered? I mean, language is still unexplained. The fact, I mean, the fact that we can train a large language model and the thing works at all indicates that there's a lot going on in language. I think it's an amazing invention. I think it's an amazing invention. I think it It's an astonishing and an unappreciated thing. You know, I just think we try to focus on our software as a special thing, but really it's the language itself. It's a special thing.

Yeah. I agree. Okay. Last question also from Sir Arthur's, you know, vault is the following. When an elderly, I'm not calling you elderly, my friend. But when an elderly and distinguished... Yeah, what? Okay, go ahead.

When an elderly and distinguished scientist says something is possible, he or she, presumably, is very much likely to be correct. But when he or she says something is impossible, they're very much likely to be wrong. I want to ask you, what have you been wrong about? What have you changed your mind about? What thing did you think was impossible came out to be true? In the early days of the Internet, I was saying things very similar to what I'm saying now. And then there was such a fever pitch of enthusiasm for a set of things I was really skeptical about

that I went along with them for a while. Like an example was pirate culture of like, let's just make music free and all that. And now what that really turned out to be, as well as the open source movement for all of the admirable qualities and all of the, everything that people love about them is in a way true. But they also just, because of network effects, they create more and more centralization of wealth and power in a few digital hubs like Google. And other people tend to become impoverished with only sort of misleading examples of the occasional influencer or YouTuber who. does well or something, but overall, the whole community of musicians, for instance, sank down instead of, you know, and Spotify is not, is no substitute for mechanicals. I'm sorry.

I've been a professional musician a lot. It's just not even in the same zone. So the thing is, I allowed myself to get carried by popular. Another one is I was anti-nuclear. I got arrested at an nuclear power plant when I was a teenager and stuff because I thought that was just the side of the angels was being anthonyluclear. anti-nuclear. And now I'm thinking, oh, my God, we set back research we really need now because of climate change. Like, we really blew it. And I knew better. I knew physics. Like, what was wrong

with me? And it was just social pressure was so intense. And that doesn't mean that everything about the anti-nuclear movement was wrong. Yeah. Sure, the nuclear energy world was filled with corruption and all kinds of problems and a lot of, I mean, like, but the point is what we did. We were so effective socially that we created a, we created a kind of a black mark on this whole thing that now we really, really need. And if we've been doing more research all these years, I think we'd be in a better place. And it really kills me. No. And so, I mean, I think a lot of the, like, I don't regret just being technically wrong about something because what can you do? Yeah, exactly. But I do regret being swayed by social pressure. Well, there was a girl involved, if I remember, maybe.

Oh, yeah. There's always a girl. again. I know. I'm going to get canceled. All right. That's three cancellations. All right. So we got out of the Ayanahara. The Ayanhara. We got the Poo-Poo-it. All right, Geron, thank you for reminding us. Thank you for all you do, my friend. Thank you for reminding us. The technology is never just about circuits and code. It's about what kind of humans we want to choose to become. And so from VR as a new physics lab to the ethics of AI to the ancient command to choose life. you've given me and my audience ways to choose wonder and heed warnings. And I think listening to Geron and learning more about him is such a delight.

And I'm glad that we exist in the same reality together. Oh, thank you. Likewise, this is really fun. All right. Bye, my friend. Take care. Thank you. The virtual reality is about expanding reality rather than escaping it.

What should we do with all that power? That's the question Jerome left us with. And it connects directly to another uncomfortable conversation I had recently. And I know if you enjoyed this conversation about technology and consciousness, you'd need to check out my conversation with Donald Hoffman about the case against reality. Don argues that reality itself may be a kind of interface, a cosmic VR system, that evolution designed to keep us alive, rather than to show us the truth. It's the perfect companion to today's discussion about perception, virtual worlds,

and what consciousness actually is. Click that episode here and don't forget to like and comment and subscribe. Yamava Resort and Casino at San Manuel is California's number one entertainment destination for today's superstars. Catch the Jonas Brothers return to the Yamava Theater stage on April 30th, the powerful vocals of Demi Lovato on May 17th, and the signature Southern Country Rock

of Eric Church on July 19th. Tickets on sale now at yamavaitheter.com. Only at Yamava Resort and Casino, celebrating its 40th anniversary. You in? Must be 21 to enter. Ambition comes in all shapes and sizes. At First Citizens Bank, we roll with your goals because we're built for what you're building. Fit for your ambition.

First Citizens Bank.