The Exponential Age Series: Part I

0:00People are going to lose their minds. This is a moment in history unlike anything humanity has gone through. It's a very different world for humans to come. Take a step back and see the broad picture, which is the way all these technologies are interlinked. Because this is all about exponentiality, and humans can't think in exponential terms. How consequential do you want to say machine intelligence is? It's almost certainly as consequential as writing. How long did writing take to disseminate through the human population? You know, hundreds, thousands of years. And we're dealing with it now on a scale of months.

0:33But in this kind of world, you're compounding 100 % growth every year, and the numbers become astronomical. AI is going to spot patterns in the world that were just completely invisible to us. Even if you think that the AI and the robots are your demise, you might as well bloody invest in them and make some money out of it. If not, you're just going to be angry man shaking your fists at the clouds.

1:10so welcome everybody i'm really excited about the week ahead because everything is going to be on a topic that's dear to my heart which i think is the biggest opportunity set we've ever been given but it comes with a lot of complexity as well and that's something i refer to as the exponential age. And the exponential age, just to remind you where the idea came from, originally, I'd started to look at the Fed balance sheet as a denominator and realized that there was debasement of currency going on. And I started looking at everything versus the Fed balance sheet. And I found two assets that miraculously outperformed it, which was crypto and technology stocks.

1:52And I'm a macro guy. I hated investing in technology because technology was always for the stupid optimists, the people who never believed in mean reversion. And it was always expensive. And it just didn't suit me as a macro person, most macro people in the same boat. And then I started to understand, well, why is it doing this? And as I did the work on crypto, I started understanding Metcalfe's law and network effects and adoption models. And I realized that technology was going through the same thing. So So I then started going through the work on technology to understand why does it trade expensively?

2:24Why does it keep going up? Because you look at the log trend, it just keeps rising over time. The log trends of Amazon, Microsoft, the NASDAQ, Tesla, all of these things are endlessly bottom left, top right. But they're volatile when they're not on the log trend. So I want to understand that. Part of it was obviously the debasement of currency. But the other thing I started to realize was technology had been increasing in speed and also in kind of leap forward. And we were about to enter a period where we were probably going to get the fastest period of technological change humanity's ever lived through.

3:01That was the building blocks of what I then discovered were, now I'm going to read through the list because there's a lot of them. But I broke it down to different layers. One was the infrastructure layer, which was kind of data networks, compute energy, and that's including EV and new energy and nuclear. And then on top, David Matten, who's going to be joining me on this as well, thought that we would shuffle AI into that layer of the infrastructure layer, the base layer, the important part, which everything's built on top of. I had it in the next layer, which is the productivity layer. I think David was very right in that.

3:42Productivity layer is like the application of these things together, plus robotics, and you get robotics, 3D printing, and autonomous vehicles. Autonomous vehicles are nothing more than robots with AI. In fact, most robots are just machines with AI. So AI, I did realize, was a foundational layer. Then there was the digital value layer. That was how in this new world, this kind of world where we're getting increasingly digitized. We need digital currencies, digital payment systems, storage systems, methods of verification, everything else that blockchain does. And that was a very important foundational layer for all of this.

4:16And after that, it was like the human layer. How do we extend and understand humans better, extend their lives? And stuff like that was wearables, biotech. Because when you put massive compute with lower energy costs and AI, you're going to make massive scientific breakthroughs. And we will increase longevity, we'll solve all sorts of diseases. And these kinds of things are all coming together, the internet of things and the metaverse itself. So when I saw all of these things, I realized that we'd gone from the phase that these were all being built over the last 10, 15, 20 years, but now they were at the age and point of adoption.

5:00And when I realized that, I realized that all of these individually had Metcalfe law type models, which means they're very rapid. You have to look at them in logarithmic terms. But if you realize that all of these are actually interlinked, and if they're interlinked, then you get something which is Metcalfe's law squared, which is Reed's law. I actually asked ChatGPT4 if there's any example of Reed's law actually that's happened in technology. And it doesn't come back that there is one. It's happened in biology, but not in technology. So all of these things are about to accelerate. I first wrote the article, I think it was in like April 2021.

5:42And even I've been shocked by the speed of which AI came onto this world. We'd seen crypto, and that was the fastest adoption of any technology we'd ever seen. And AI dwarfed that. The speed of which it happened was extraordinary. So I've been writing about this in Global Macro Investors since April 2021. And I think it was January that I asked David to help me. So David, how are you, my friend? Hello, Raoul. Very well. How are you? Great. Yeah. I mean, you've been down this rabbit hole on this journey with me as like sending postcards from the exponential age. Well, they're actually quite long articles.

6:18and super interesting. What impact did it have to you when you saw this framework? Because you're in the technology space. You've been around it much longer. Yeah. I think it seems to tie it together, right? Yeah, exactly. I think it's easy to see the individual elements. It's when you take a step back and see the broad picture, which is the way all these technologies are interlinked, you know, robotics, simulations, AI, virtual world, space technology, and you see the interactions they're going to have with one another and how that is going to accelerate what is already an acceleration, if I can put it that way.

7:03That's when you have your old fuck moment, essentially, you know, and you realize, I mean, and you realize we're about to live through a wave of change that is historic. I just can't see how that's not the case. And you can look back to other times in history when essentially you saw the death of an old world and the birth of a new world and how exciting that is, how exhilarating it is, the opportunities that are latent in it. But let's be frank as well, the difficulties, the pain that can be latent in it, all of this is heading our way and it's a tsunami and it's not going to play out across hundreds of years.

7:42Like some of these ships, you know, like that axial age shift around the time of Christ is going to play out in decades. It's very, very difficult to understand, but we have to do the work of trying to understand it. Because this is all about exponentiality and humans can't think in exponential terms. You can't understand when things double every year, what that actually means. It's like the power of compounding. People understand that when Uncle Warren Buffett It talks about the power of compounding. You're thinking, I'm going to compound my 8 % returns. But in this kind of world, you're compounding 100 % growth every year, and the numbers become astronomical.

8:25Like we saw the speed of AI getting to 100 million people in, what, three months? Yeah, right. And we've set up an infrastructure now, essentially a connected world, the internet, where innovations can just smash into humanity in months. We simply have never lived in that world before. When comparable technologies were invented, how long did it take them to disseminate through the human population? How consequential do you want to say machine intelligence is? It's almost certainly as consequential as writing. How long did writing take to disseminate through the human population? Hundreds, thousands of years.

9:07and we're dealing with it now on a scale of months, years, we can't comprehend that. I guess it's because we don't encounter exponentials in nature. We encounter linear change in nature. You do see viruses, stuff like that, the somethings. That's true. That's true. But they take us unawares, don't they? It's a good comparison. They just radically destabilise societies. and now we're building something that's going to do that. So for people like us who love to observe and try to make sense and build the big puzzle, it is a thrilling time to be alive. It's kind of a scary time too because we're heading into a future that just feels so uncertain and so new that the work of understanding it is so crucial.

10:03and I think you know people are going to lose their minds we've got like the election I've got a piece coming out on Wednesday about AI in the election we've got the US election which is always a heightened sensitivity because of social media has now created this beast we're going to throw in AI and then we're throwing in geopolitical stresses and then we're going to throw in the existential threat to jobs and the economy and what this all means and people just are going to grapple with this and it's going to be how i've always thought about this is even if you think that the ai and the robots are your demise you might as well bloody invest in them make some money out of it because you know if not you're just going to be angry man shaking your fists at the clouds and where does that get you yeah i mean i under people of course people all of this triggers a huge range of human emotion.

11:00And that's completely normal. And we can't change that. And people's emotions become confused. And then they're thinking about it becomes confused. They think about whether or not this wave of change is all going to be positive. You know, it's not. They think about whether they personally are going to like all the implications. You know, they're not. And then they allow that to sort of govern in a simple way whether they want to be interested at all or whether they want to participate at all or even pay any attention. And those are very different things, you know, exactly as you say. I mean, any intelligent observer is going to see things happening around them that causes perplexity, dismay, all kinds of feelings.

11:42But you need to understand this. And, you know, you can't fight the weather. you may as well participate on some level, even if your participation is about saying, look, from my perspective, this is about insulating me from the insane wave of change that is coming. I want to build certain guardrails around my modes of existence as they are now, my work as it is now, whatever it is. But to do that, I need degrees of freedom and capacity to do it. That's a perfectly coherent position. But if you love all this stuff and you just want to dive in and you want the chip in the brain, you're signing up for Neuralink, all of that, fantastic, great.

12:26Knock yourself out. That's also a totally coherent position. There are all kinds of ways to face this. But if your position is, I'm going to turn away and not face it, guess what? It's not going to work. It's going to smash into you whether you're facing it or you're turning away. I just think also, this is a moment in history unlike anything humanity's gone through. And I know every day is a moment in history and there's always something big going on, but this is a fundamental shift to humanity. I mean, like one of the things you and I have talked about, I mean, we've not talked on camera since June, but one of the things we talked about was that AI was the infinite scaling of knowledge i mean we can't get our heads around what these things mean and where this is going to lead to but we are living history and that's why i i like a writing about it and documenting it in real vision to put our markers down as this thing unfolds because you kind of forget it was like the pandemic oh what happened so fast and then you forgot to keep a diary of it and then you've kind of got these blurred memories.

13:35But just to anchor yourself in what the hell is going on, I think is just really important for people. So as you said, whether they like it or not, it doesn't matter, but you can't ignore it. Yeah, I feel the same. There's just a huge value and huge pleasure simply in documenting this. This is what it was like to live through it. This is the sound of someone trying to make sense of this insane wave of change. And I just don't think people understand how fast this is all coming. Here's one example that I saw just a couple of days ago, and this is totally mind-blowing. And it's an example that covers, look, robotics, simulated worlds, and AI, and brings them all together.

14:17NVIDIA have this simulation platform called Omniverse that allows the user to simulate complex, real-time physical environments. It simulates the physics in those environments. You can bring the physical world into virtual space in this simulation platform that NVIDIA are developing called Omniverse. They've just launched, a couple of days ago this is, an AI agent in Omniverse called Eureka. And what that AI agent does is autonomously writes code that teaches simulated robots how to move. And it's just written a ton of code that's taught simulated robot hands to perform incredibly complex manipulations.

15:07And it does this completely autonomously. It generates code that starts a cycle of kind of trial and error learning with the simulated robot hand. And then it iterates. So it gets feedback from the robot hand to improve the code, which improves the learning of the hand. And what you're seeing is, you know, in a matter of days, weeks, hours, the training of robots to perform complex movements and operations in physical space that we'll then be able to take and translate into real robots. It's mind-blowing. We have AI just autonomously training our robots for us now in simulations. people don't understand how fast it's coming i wrote about a piece in gmi this month about the fact that they think they can use large language models to understand the language of whales and dolphins yeah you know scientists are producing these papers saying yeah we think this is applicable and it should work and we can finally crack the code of what they're talking about and how and what their language is.

16:14I mean, that again is bonkers where we're starting to understand the language of animals. Hey everyone, we're going to take a quick pause and hear a word from our partners. We'll be right back.

16:41See a trading opportunity? You'll be able to trade it in just two clicks. Feel ready? You can move to real money with as little as$100 once your account is approved. And the great thing is that in addition to crypto, Plus500 gives you access to a wide range of instruments. S &P 500, NASDAQ, gas, and much more. Explore equity indices, energy, metals, Forex, and beyond. With a simple and intuitive platform, you can trade anytime, anywhere. Experience the fast, accessible futures trading you've been waiting for with Plus500. With over 20 years of experience, Plus500 is your gateway to the markets.

17:15Visit us.plus500.com to learn more. Trading in futures involves the risk of loss and is not suitable for everyone. Not all applicants will qualify. Plus500, it's trading with a plus.

17:30We've both been doing some pretty interesting interviews around this exponential age one. The Robert Scober one, I loved because he's kind of all over this. Talk me through a little bit about that. Yeah, I mean, he's an incredible, he's an incredible figure, you know, a Silicon Valley legend for someone like me, a real legend too, because he's been writing about technology online his whole life. You know, in the early 2000s, when I emerged blinking into the world, you know, out of university, Robert Scoble was online writing all about technology. So he's a real OG of the kind of technology blogger, to use an old fashioned word, space.

18:06The interview, I mean, go and check out the interview. It's incredible. He talked a lot about humanoid robots and how fast they're coming in this piece. He's got a kind of framework around how humanoid robots kind of converge with autonomous vehicles. So you'll have autonomous vehicles delivering humanoid robots into people's homes on demand to fold the laundry, do the washing up, load the dishwasher, all of that, then hop back into the autonomous vehicle and go to the next house. everything as a service and how that completely rewrites the economy you know it's the it's the next technology that utterly transforms the texture of everyday life in the way that these things did you know in the way the phone did he's also massive into spatial compute and vr and ar and all of that nexus i don't know did you ever see that i think it was a bbc or at least a uk series called humans yes yes because you know that was brilliant because anybody who didn't It was on Netflix, maybe still on Netflix, but it's a series about humans living with the AI robots and how they treat them.

19:15And it kind of addresses it from every angle, from racism to fear to empathy for the robots, hatred for the robots, the whole thing, and what it does to society. I thought it was brilliant, and it feels like it's closer than we imagine as well. It is. It's a lot closer than it was in, like, 2015. 2016 when that show was playing and yeah i mean we can get deeply philosophical about it because you get you know what does it do to the economy but then you get to places around what does it do for society and for politics because at what at what point do we start to have a genuine conversation about whether machine intelligence has some form of personhood i'd bet in our lifetime it's it sounds crazy now and i argue about our lifetime we get to that conversation You see, I think humans are anchoring themselves on humans and what we perceive as intelligence.

20:11You know, I always keep going back to watch the documentary on Alpha, Go, the Deep Mind, playing Go and it wins the first four games. It loses game number five, I think it was, against the human. There's commentators who are experts on the game. They're like, well, it's playing this move, it's playing that move. And that's like the Garry Kasparov versus the machine, you know, Deep Blue, you know, which was a machine learning model. It lost the game and then never lost the game ever again. And the next game, these humans are going, well, this is stupid. Why is it playing that? Nobody play this move.

20:48And it just didn't play human moves. And it was at that point, I'm like, I don't believe the bullshit narrative, stochastic power because it cannot do that it cannot suddenly learn how to play a game and play it like no human has ever done it before when he's never been taught that because that model of kind of predicting the next thing doesn't work that way i think that's really interesting the guy from anthropic there was an excellent interview with him uh on some podcast and he's like we've had to build an MRI machine for our AI, you know, to try like we do with MRI machines, try and figure out brain activity.

21:29We're trying to figure out how this bloody AI thinks. We can't, we can't figure it out. So I just, you know, when people talk about, you know, when, when is it a person, when does it have intelligence? I'm like, I don't think we have any clue whatsoever. There was another thing that really struck me was Lex Fridman interviewing Sam Altman. And Sam stopped him at one point, stopped Lex and said, do you think we've invented AGI? We just don't know it yet. And Lex stopped because he's an expert AI as well. He stopped, thought about it for a while and said, I don't know. Yeah, that is an amazing conversation.

22:13I mean, what we're going to be forced to do is disentangle all kinds of difficult concepts that at the moment are extremely entangled. I mean, one of the big ones that people tend to entangle, and this goes somewhat towards your point, is consciousness and intelligence. And they think that when we say these AIs are going to become superhumanly intelligent, like extremely, extremely intelligent, that we are already saying that they have some form of consciousness or subjective experience. we're used to being in a world where the only highly intelligent creatures around us also have subject consciousness subjective experience we act as though we believe they they do you know and that's other people it's very hard even just to decouple intelligence and consciousness and then you start to talk about what is it a person what does that mean you know what is the concept of personhood i mean you you will start to talk about should it have certain rights you know things like that i mean this gets deep deeply philosophical and it's going to it's going to shred people's heads basically and shred their conception of what human beings are and how we relate to one another and how we relate to the technologies around us because i think we're going to get to the point where we feel even more acutely than we're starting to now that we're amid another way of seeing the world around us and this is like the whale thing ai is going to spot patterns in the world and see things in the world that were just completely invisible to us and feel like alien insight but it's going to then sort of be able to translate it into something we can understand and use um and that's going to be head spinning you know because it just it will see the world in a completely different way and we've never had another creature or entity among us that sees the world in a different way and can sort of tell us about it.

24:16And, you know, we can always, we keep going down the rabbit hole of AI, but that's just part of the exponential age because one of the other ones is compute. And you wrote an article recently about quantum computing. And, you know, we've got an interview, I think on Monday as well on ProCrypto about quantum computing and what it does to cryptography with Daniel Darjean, who was also part of the narrative around for distributive computing power, which is all part of the same thesis. What did you find out about quantum? How far away are we from that? And what would it mean if it comes? Because I feel like quantum computing plus AI plus cheap energy equals unfathomable outcomes yeah well yes that's the too long didn't read it is unfathomable i've always found quantum to be a difficult space to assess but i went because it is early and it and the potential implications are so head spinning and it's obviously deeply deeply technical but i went down the rabbit hole exactly for that for that essay for gmi and my conclusions essentially were that we are at something of an inflection point moment.

25:33And it's really interesting the companies making progress in this space. I mean, IBM are, how often do we talk about IBM? Not too often. But they really are leading the charge on quantum computing. And right now, we're still in a place where it's about coaxing quantum bits, so-called qubits together, and massaging them in the right way. so that they remain kind of coherent, they remain in a quantum state. And IBM are working with a quantum computer that is, you know, say a thousand, I think a thousand and twenty four qubits. And they're saying, you know, next year that they've got a four thousand qubit computer that they think they'll be able to kind of operate usefully.

26:16We need to get to tens, hundreds of thousands, millions of qubits operating in unison before we get to the kind of results we're talking about or we dream of. It's probably, what, 20 years away? Yeah, I think we're talking decades away here. But if we get there, like I say in the essay, we're not talking about operations that might have taken a few months, but now you can do in a few days. We're talking about operations that would have taken 80 million years for the world's fastest current supercomputer and now take a day you know and you know so number one yeah it just blows apart all cryptography the security of all the world's data the encryption of all the world's data because that's built around the idea the fact that currently you know to break that to break that encryption would would take millions of years of compute power but quantum will do it in seconds.

27:19But then there's just incredible implications across the board. If we can cultivate that kind of compute power, the things we can do, for example, in the life sciences and the material sciences are just amazing. And it's the life sciences and the material sciences, particularly, that are such exciting spaces. Because obviously, at a subatomic level, reality is operating in quantum ways and quantum computing because it runs on quantum bits bits that can be in more than one state at once is a very good at simulating that kind of behavior so we'll be able to you know cultivate new kinds of like small molecules assess new kinds of proteins develop new kinds of materials using that kind of compute power just exponentially faster and more more powerful than we can now quantum physics has been quite interesting because of these quantum computers and kind of proving that quantum physics kind of happens where two things are in the same time in different spaces um d smith who i've just interviewed on the platform got me down the rabbit hole he said read a book called american cosmic and it was about ufos yeah yeah and i started to think i've been i've been going down some weird rabbit holes about both ghosts or those kind of other experiences and the ufos and i'm kind of thinking i don't know if you've ever seen um his dark materials the the northern lights trilogy it feels that maybe a reason for that is because these are kind of the ability to see into a different quantum field whether it's time so people have a different time things with different time i don't know I think there's a lot.

29:07I mean, this is absolutely, it's true. It's absolutely fascinating. And this is one for a late night one with a cocktail or a whiskey or something like this. Because, I mean, yeah, quantum, the quantum account of how reality works at the deepest level is, of course, we all know, utterly, utterly strange. And you have to ask yourself if the current picture we have of ultimate reality out there is adequate. It would appear there are things going on that are just realms and realms and realms beyond our current understanding. And yeah, when you start to think about sort of unexplained phenomena, you know, and UFOs can be part of that, you have to be open minded to the idea that essentially what I'm saying, the current scientific mainstream account of reality is not right.

30:01This is what we could get. People are going to watch this and think they've lost their fucking minds. But honestly, these are from serious people. These are not kind of UFO hacks. These are serious people, serious scientists about all of this and past lives and all of this kind of stuff. And it just feels like it's connected within this quantum physics. I have no idea. But anyway, that's just an aside that something big is going on. Another thing that I'm really interested in is the nexus of, we talked a bit about robotics and AI and EV and new energy. And that to me gets me to Tesla, which is that stock everybody loves to hate.

30:41But I don't think people understand still that they've essentially trained a large language model to be like a video language model or a large video model. So they're taking, so how all autonomous vehicles are being used was basically machine learning. So somebody was typing in stop sign equals stop. You do this, you do that. And kind of that's how it was coded. and then elon flipped on his head i'm sure others have done it before him but basically flipped on his head and scale and said we're just going to take the eight cameras in each tesla there's five million teslas we're just going to ingest all of that and train ai how to drive based on how humans drive and that's what they're doing with the world's largest supercomputer dojo which is still ramping up.

31:30The amount of compute power and the amount of video processing that trains now cars to drive without essentially the machine learning element is a massive breakthrough because that's what trains robots to operate in a human environment or a non-human environment like Mars. These are really, really big things. So a Tesla becomes a robot with intelligence drived on non-fossil fuels, an alternative fuel source. And you take the same technology and put it in a humanoid robot, and you've got the optimus. And then the other idea is the Neuralink, which is the connector between humans and all of that technology as well.