Electricity and Magic - Ask Me Anything

0:10Hi, everyone. Welcome to the first ever, the inaugural exponentialist Ask Me Anything. A huge welcome to all of you because this is the first one. You will know this. We've done something really special and opened this Ask Me Anything to everyone across the platform. But of course, it's huge and special hello and welcome to those of you who've already joined the Exponentialist. I just, as I've said countless times now, I am so excited about the journey we're all going to go on from here. But let's think about today. The purpose of these Ask Me anythings is for me to talk a little bit about the monthly essay, the big ideas that I've written about this month.

0:54And then, of course, even more important to talk to you. So I want to use these sessions. Yes, to give a summary of the core argument that I'm making this month in the essay, to give a bit of behind the scenes thinking, to give you a bit of the journey that I went on to get to the ideas, to get to the framework or the thesis that I want to argue for this month, to go deeper in places if that's necessary. But then, of course, primarily to talk to you, to hear from you. Because as I've always said, from the earliest days of putting this thing together with Rao, it's the conversation we can build around the exponentialist that is so hugely exciting to me.

1:37Speaking of Rao, I know that when he doesn't ask me anything, he likes to have a drink, maybe two. I'm steering clear of alcohol for my first Ask Me Anything. I've got tea instead because I am so profoundly British and tea is the British answer to everything. So look, hearing from you is key. As I said in the essay, think of the monthly essays as the campfires we gather around. And these Ask Me Anythings as the conversation we're going to have. and over time we can find lots of new ways to expand that conversation there's going to be a special exponentialist community where we can all chat on the platform that's coming we'll think of other ideas but these ask me anything's will always be the core of the conversation too and there's obviously a huge amount of knowledge and insight and profoundly interesting careers and life paths out there in the exponentialist community i mean i can see the amazing people joining it every day.

2:35And I want us all to be able to access that. I want us all to be able to talk to one another, to learn from one another, to go on this journey together. Like as we're endlessly saying, we're all beginners in this new world we're building. Collectively, we can face a tsunami of change that is coming better than we can individually. So I want us all to be able to access that. And of course, that aligns with Real Vision's broader mission when it comes to community and network. So that's what we have in store, talking about the essay, and then we'll move to questions. Let me talk first a bit about the essay and to give you, like I said, a behind the scenes look at some of my thought process that brought me to the essay.

3:24And then we can dive into the core argument. Because the truth is I reflected for a huge amount of time about what the subject of this first essay should be. And that story just, you know, I can tell it very briefly, but it really helps contextualize everything we're doing here, the reason that Raoul and I built The Exponentialist and this journey we're going to go on. There were so many directions I could take for the first essay, you will know that you've heard it often enough. The exponential age is about multiple technology revolutions all happening at once and all overlapping with one another and interacting with one another.

4:09Raoul calls it the biggest change to his macro framework, which he's obviously been building for decades. The biggest change to his macro framework that he's seen in his lifetime. Just a wave of change, the like of which we've never seen before. That's the fundamental thought at the heart of the exponentialist and the exponential age thesis. And we're on a mission to understand what that all means. Yes, to profit from it as investors, but what it means for us as human beings, as professionals, what it means for our careers, for our futures, for our families, for our children. Yes, profit, but how do we flourish in this new world we're building too?

4:52So amid all that, what do I choose for the first essay? I want the first essay to set the scene, to set the foundations for this journey we're going to go on. And the essay that I came to is an essay that delivers a framework that I call electricity and magic. And some of you will know this because you'll have read the essay already. Electricity and magic is about delivering a big picture, overarching framework for understanding AI and understanding how AI, how machine intelligence will manifest and the role that it will play in our shared future and in the exponential age. Yes, the exponential age is about multiple overlapping technology revolutions.

5:40But the reason I wanted to zero in on AI for this first essay is because there's something special, and we all sense this, there's something special about machine intelligence. It's a kind of meta technology that's woven through everything that helps guide everything else in the exponential age. So if we're going to understand the decades ahead, if we're going to understand the exponential age thesis, we need to build that on a foundation of some understanding of AI and machine intelligence and what it means and how it will show up in the world around us in the exponential age, in the years ahead.

6:22And remember, with this essay, and I'll come back to this point soon, I'm building frameworks and ideas in these first essays, the first few essays of the exponentialist, and building frameworks and ideas that will come back to time and again on our journey, and we'll build on them and build on them over time. But we need to set a foundation. And understanding machine intelligence, understanding AI is so profoundly important to that foundation, to everything else that's going to happen. So that's why I wanted to write about this framework, Electricity and Magic, for the first essay in The Exponentialist.

7:02This is where it'd be handy for me to share some slides, because I want to get in briefly to the core argument. Just take us all on a tour of the core of the electricity and magic framework. I hope that will be useful. And then we can get into some questions. So let me do a little bit of magic myself and share my screen. And we can continue from there. So yeah, look, this is the mission I wanted to go on with the first essay. How can we understand machine intelligence, how it will manifest, the role it will play in the years ahead? We're not talking here about the realms of technical understanding, technical understanding of how AI works.

7:44We will do plenty of that in essays to come. We can drill down into some of the mathematics essentially behind how AI works and make that accessible. But here I'm talking about conceptual understanding. I'm really talking about model building. How can we build a model, a framework that allows us to understand how AI will manifest in the years ahead? And the generative AI revolution that we're living through makes that even more challenging and confusing because we all see AI doing incredible things across the last couple of years, things we never expected AI to do, you know, creating incredible works of art, incredible images, music, even films.

8:24We're at the beginning of text to film models. On the other hand, you have AI doing very practical things. You have AI driving autonomous vehicles, navigating robots through complex environments and solving Rubik's cubes. So how do we understand a force like that? It's like an alien force. I mean, some people have compared it to the rise of machine intelligence, essentially, to a kind of alien invasion. And perhaps there's something to that. We can think about that more deeply in future installments. What really is AI? what is machine intelligence at the deepest level? How should we best understand it?

9:07That is the challenge I wanted to look at. And to skip to the very end, and I've already given this away, the framework for understanding my answer to that question, is this framework, electricity and magic? So what do I mean by electricity and magic? That's what I want to spend a few minutes on. And of course, if you want the deep answer and the full answer, go and read the essay. It's all in the essay. I can't get into every twist and turn here. I just want to spend a few minutes giving the core ideas. So to understand the electricity and magic framework for AI, first, we need to understand the most simple, the most fundamental model that Raoul and I use to understand the exponential age.

9:53This really is a baseline, extremely simple model. It's a layer model of the exponential age. And you can see it on your screen here. We've built this layer model to help structure our thinking about exponential age technologies and about what's coming. And as you can see, there's four layers. The foundation layer is all about the infrastructural technologies of the exponential age. And we typically call those energy, data, and compute power. Those are forms of infrastructure that will fuel everything that's happening in the world around us in the decades ahead. Then there's the productivity layer.

10:32There are tools. There are the things that allow us to enhance our scientific understanding, to be more productive, to grow the economy, to do more. That's what the productivity layer is about. There's the value layer, which are the exponential age technologies that are going to transform the way we store value and we share value. And ultimately, they'll change the way we think about value. And of course, primarily, we're talking there about the blockchain revolution. And then there's the human layer. There's how this is going to impact on individual human beings, on all of us collectively, our careers, our lives, our health and wellness, our own bodies.

11:13How is this playing out for us as human beings? Now, look, where does AI fall in that simple model? At first glance, we had it originally in the foundation layer. Sorry, in the productivity layer. We had AI in the productivity layer. We were modeling AI as most people model it, as the mainstream models it, as a tool. this thinking is, okay, how should we best understand AI? How's it going to manifest in the world around us in the years to head? Well, it's a tool. AI is a tool, among other tools, that will allow us to do more. It will unlock new forms of scientific understanding. It will allow us to perform new kinds of complex tasks, 3D printing.

12:02It will grow the economy in those ways. It will do knowledge work for us, all of that. That makes a lot of sense. In many ways, AI will manifest as a tool in the years ahead. But we all know that doesn't quite capture the entire picture. Intuitively, we can feel there's something more to AI than just a tool. Other tools, for example, can't build new versions of themselves. You know, this thing, as I say in the essay, is a wonderful tool, but a phone is never going to design and build me a better phone. We also don't think that, you know, any other tools are going to suddenly determine goals of their own or start nuclear wars or enslave us or all of that.

12:53I mean, all of that's extremely far fetched talk in my view at the moment. But there's something different about AI. it's best understood not as a tool, but as something else. I think we can all intuitively sense that. Modelling it as a tool just doesn't quite work. So what is it then? What is AI then? We're still stuck with the original question, how should we best understand AI? How will it manifest in the exponential age? To answer that, you need really to take a step back and ask yourself, Well, if we're trying to answer the question, how will machine intelligence manifest? What really is it?

13:36We need some form of answer to the question, what is intelligence? That is, as you all know, a highly contentious, highly contested question. There's really no consensus answer to what intelligence is, even among cognitive scientists, among biologists, among philosophers. But here's a definition that serves our purposes, that aligns, I think, with most mainstream thinking on this question, and that makes sense. Intelligence is what does useful work on information. Intelligence is the ability to do useful work on information. That's my definition, my working definition of intelligence. And when you understand that definition of intelligence, you take a step closer to understanding how machine intelligence, how AI will manifest, how it will show up in the world around us in the exponential age.

14:42The reason this definition is so important, intelligence does useful work on information, can be summed up essentially with this chart. Look at what's happening to information in the exponential age. This is a classic exponential age curve of the kind. We'll look at lots of them in the months ahead. Data in the years ahead is set to go exponential. Data creation is going exponential. This is the mainstream forecast, but there are great analysts, Tim Davies, who's been on the Real Vision platform a few times, who think this estimate is a vast underestimate. Even these numbers are hugely underestimated.

15:32Data creation is going to explode. And why is that? The reason is that we're weaving a layer of data, a layer of information through everything in the world around us. So the data is essentially leaving the confines of the digital realm of our computers and going out into the world around us. We're weaving through sensors, through microchips. We're weaving a layer of data, a layer of information through every building, every vehicle, every sidewalk and lamppost, every object, you know, hundreds of billions of connected objects coming, every person, because so many people have a wearable device now.

16:19We're weaving a layer of information and data through everything. And the result is a super network that's blinking into life around us, a connected world, a world that's far more connected even than the world we live in now. What is this really? This is about the deep merging of information and physical reality, the deep merging of the realm of bits with that of atoms. And this is a profound phenomenon of its own that underpins so much of the nature of the exponential age. It's the emergence of what I call the unified digital physical field. And I'm going to be writing a lot more about this and talking a lot more about it to all of you in the months ahead, because this deep merging of information and physical reality, I think is a profoundly important dimension of the exponential age.

17:14it's kind of the final innovation platform we're going to see a huge wave of innovation built on this platform itself anyway you have the emergence of the unified digital physical field you have information now as part of the physical world around us because of connected buildings connected cars connected sidewalks hundreds of billions of connected objects connected people all in this super network Now we're another step closer to understanding the role that machine intelligence plays. Because what role does machine intelligence play in that environment, in that super network? Well, when information becomes, this is the answer I came to.

18:01This is the fundaments of the conceptual framework that I'm putting forward in the essay, is this idea. When information becomes a part of the physical world around us, when you have this deep merging of the realm of information and the physical world, then AI, machine intelligence, becomes something akin to a new form of fuel that helps power everything. AI becomes something akin to a new form of electricity that helps power everything happening in that unified digital physical field. Just as real physical power, physical energy does work on physical mass in the unified digital physical field, so machine intelligence does useful work on the information that is part of that field.

18:54And together, they drive everything, you know, hundreds of millions of autonomous vehicles, billions of robots, every connected city, every wearable device, every connected object. They're being fueled by, yes, of course, energy, which has to fuel everything, but also by machine intelligence, which fuels the machine, the informational aspect of all of this. Now, that is really a new, admittedly a non-standard and something of a head-spinning idea, because we're used to thinking of intelligence very differently. We're used to thinking of intelligence as something sort of rare and abstract and the preserve of we human beings, you know, something that makes we human beings special.

19:40Only we have intelligence in its highest form. Um, what I'm saying is in the exponential age, intelligence becomes abundant and it becomes infrastructural. It becomes a form of power that helps fuel everything going on above it. And so it belongs to take us back to the original model I showed you, not in the productivity layer as a tool. And we talked about why modeling it as a tool doesn't really make sense. It belongs in the foundation layer alongside energy, data, and compute. Intelligence, I think, is part of the foundation layer of the exponential age. Intelligence in the exponential age becomes a form of infrastructure that helps power everything.

20:26Getting your head around the implications of that is really difficult. I haven't done it. I mean, no one's done it. I think conceptually that model is deeply powerful in helping us to make sense of where we're heading and of the role machine intelligence plays in that. But getting your head truly around the implications of it is really difficult. You see this come to life when, for example, you look at the emergence of the kind of useful humanoid robots that we're going to see all around us pretty soon. I mean, this robot is not particularly humanoid. But this is a robot from Alphabet's Everyday Robots division.

21:08And essentially, what they've done is they've unified robotics technology with a large language model, which means the robot can understand natural language and respond to natural language instructions like, oh, I've spilled my Coke. Will you tidy it up for me? Just something that natural, that intuitive to say this robot can understand it and it can take a series of actions that will tidy your coke up, clear up the mess for you. Now, look, obviously physical energy in the form of a, you know, coming from a battery is powering this robot. Machine intelligence is in some sense powering it too. It's powering the informational aspect of what's happening here.

21:53And this is a great example of the unified digital physical field. You have the deep merging of information, of bits and of atoms, all coming together in this machine, this robot. Just one tiny example of what I'm talking about. There is, as I wrote about in the essay, another even deeper manifestation of the infrastructural role that intelligence will play in the exponential age. And that's the role it will play in the energy network itself. This is the part of the essay that spurred some of the most interesting comments on the platform. And we'll come to some of those comments and some of those questions that you've sent in very soon.

22:43But the fundamental thought there is, without getting too detailed, and again, go and read the essay for all the details and the entire argument. The fundamental thought there is, look, we are moving directionally towards a global energy system that is more founded in renewables, particularly solar and wind. And what you have there are all kinds of challenges. As some people said in the comments, you have intermittency, the sun doesn't always shine, the wind doesn't always blow. So you have the need for storage. You have all kinds of new challenges. You end up with a highly connected, smart energy grid that is based on energy plants, but also energy resources at the edge, solar panels on roofs and all of that stuff, streaming a tsunami of data just every second into the cloud.

23:41the only way that we can manage that is, well, essentially the answer is we can't manage it. The only way that can be managed is via machine intelligence. Only machine intelligence can handle, can make possible the kind of highly interconnected, highly variable, intermittent, prone global energy system that we are moving towards. And so what you end up with is a situation where energy and AI fall into this deeply symbiotic relationship. Because of course, AI is very energy intensive. But the answer to that, I think, is going to be more AI. More AI will allow us to manage an interconnected energy system based on renewables.

24:28More energy will power more AI. More AI will allow us to power more energy and so on. And what you end up with is this relationship. I call it the AI energy flywheel. And this is a hugely important symbiotic relationship in the exponential age. This technology relationship, in my view, this deeply symbiotic relationship between real physical energy and AI, which becomes akin to a form of energy in the unified digital physical field, this deeply symbiotic relationship is the technology relationship at the heart of the exponential age. It might just be, and this is a bold thing to say, and there's lots of time to talk about it in the months and years ahead, and we're going on a long journey where everything I say can be challenged.

25:20It might just be the most important technology relationship ever, perhaps, to be discussed. But it's profoundly important. And look, lots of people, well, not lots of people. Some people have made extremely good points around nuclear. Why isn't nuclear a part of this picture? I mean, the short answer is it is. It is, in my head, a part of this picture. And we've got some questions about that. So I'll leave that, actually, for now. We can talk a bit about that during the questions. That is the electricity half of the framework. Just super quickly, so AI becomes a form of infrastructure that powers everything.

26:01It becomes itself akin to a form of energy, which is a head-spinning idea. Just super quickly, the other half of the framework is, of course, the magic half. What is the magic half about? Well, the magic half of this framework is about saying, look, we've set up the electricity half. But actually, in our daily lives, day to day, as we go about our lives, we go about our work and so on. AI won't particularly manifest that way to us. It will perform that role, but that role will fade into the background. Hundreds of millions of autonomous vehicles and billions of robots and connected factories and all of that will hum and buzz and whir about their work.

26:49We won't really pay any attention unless something goes wrong. It will just operate and we will let it operate. The things we will focus on, the things that will draw our attention when it comes to AI will be entirely different. They will be AI manifesting not as something akin to a form of energy, as something infrastructural, but as something akin to magic. And you see that happening now with this incredible generative AI revolution. Like I said at the top of the call, you see AI now doing things that even two, three years ago, hardly anyone was dreaming about. Incredible text-to-image models producing beautiful images that can stand on their own terms as works of art in some cases or are incredibly photorealistic.

27:43We're now on the verge of simple but improving text-to-world models. So you're going to be able to describe an immersive 3D world and AI will just generate that world around you. And imagine that when augmented reality and virtual reality technologies are really compelling and really usable. You know, we're taking a step closer with the Apple Vision Pro headset coming soon, but we're still not there yet. We will get there. Imagine when VR and AR is super compelling, really photorealistic, and you can just create an entire 3D reality around yourself simply by describing it. That is magic. That will be AI manifesting as a form of magic in your life.

28:30We'll see AI reincarnate superstars and media stars and actors and singers who've long since passed away. We get a taste of that now via deepfake technology. I mean, in the essay, I talk a little bit about this deepfake Elvis TikTok account. I'm a big Elvis fan, so it's the one I noticed. And Raoul has talked about, of course, you know, the flip side of this, which is some of the deep problems that deepfake technology is set to cause. We're going to see, get a taste of that, as Raoul said, in the US election year next year. Hey, everyone. We're going to take a quick pause and hear a word from our partners.

29:10We'll be right back.

29:14Your favorite neighborhood spot grows with Square. Indeed, my favorite neighborhood spot has quickly become Todd Snyder in Williamsburg. Todd Snyder is one of my favorite menswear shops and has supplied me with all the clothes I have needed this quite hot summer. Every business has different goals, but Square is the business platform that supports them all. From opening a new location, selling something new, or just expanding their reach. Indeed, I've seen it with Todd Snyder. In Square, also, you can get real-time insights, so don't wait for end-of-day reports. Go to square.com forward slash go forward slash real vision to learn more about how your business can grow with Square.

29:57That's S-Q-U-A-R-E dot com slash G-O slash R-E-A-L-B-I-S-I-O-N.