Intelligence in the World
An excerpt from The Exponentialist
This week I launched my new premium and enterprise-level research service, The Exponentialist.
It’s a partnership between me and the macro-economist and Real Vision CEO Raoul Pal. To mark launch day, we’ve made an excerpt from the first essay free for everyone to read. So, naturally, I wanted to share it with all of you.
This excerpt is from a longer essay called Electricity and Magic. The essay attempts something ambitious: to build an overarching framework via which we can understand the way machine intelligence will manifest in our shared future.
In the short excerpt below, I introduce one of the essay’s core ideas. That is, that via AI intelligence is becoming something infrastructural; more specifically, something akin to a new form of energy that powers much happening in the world around us in the Exponential Age.
That sounds a strange idea. But intelligence as a new form of electricity is a model I’ve been developing for a while now. Below, I make my case for it.
Anyway, enough preamble. If you’re intrigued, read on!
Intelligence in the World
We’ve seen how the Exponential Age (EA) will host multiple technology revolutions all unfolding at once.
But amid all that, we all sense that there’s something special about machine intelligence. So how should we best understand AI? What is it, at heart?
Typically, we model AI as part of the Productivity Layer of the EA. That means we see it, fundamentally, as a tool. One that will allow us to unlock new scientific understanding, enhance our engineering, accelerate creative outputs, and much more.
And that all makes sense. In the decades ahead AI will function this way: as a tool.
But we all know that this conceptualization of AI leaves much unaccounted for. Ordinary tools can’t use themselves, for example. A phone is a wonderful tool, but it’s never going to make my calls for me. Or design me an even better phone. What’s more, we don’t talk about the possibility that we might lose control of other tools; that they might develop goals of their own, start a nuclear war, or enslave us all — all highly unlikely scenarios for AI, too, admittedly.
Intuitively, we can all see that AI is different. That it’s best understood not as a tool, but something else. So what, then?
To answer that, we need to go one step down to a deeper question: What is intelligence? There’s nothing close to a scientific or philosophical consensus answer, and clearly this essay isn’t going to change that.
But here’s a broad definition of intelligence that aligns with much mainstream thinking on the subject: just as energy does useful work on physical systems, intelligence is what does useful work on information. When we clarify intelligence in this way, we take a step closer to understanding the role it will play in the Exponential Age.
That’s because of the fundamental role that information, that data, will play in the Exponential Age. The EA means an interconnected world that runs, above all, on truly staggering amounts of information. Understanding this is crucial if we’re to understand the role that AI plays in the decades ahead.
According to the International Data Corporation, global data creation is going exponential:
And some excellent analysts believe that even these numbers are a vast underestimate (more on all that and its implications in an essay to come).
What’s driving that insane data explosion?
It’s happening because information is breaking out of the digital realm and into the physical world. We’re weaving a layer of information, of data, through every part of the world around us. Every building, vehicle, household object, and person; all of them will be connected and streaming a constant tsunami of data that reports on their status, and the status of the world's physical environment in which they’re situated.
Just look, for example, at autonomous vehicles (AVs); multiple forecasts have numbers going exponential across the coming years. This forecast is from Next Move Consulting:
Each AV will be a rolling supercomputer, constantly streaming vast quantities of data about itself and the world it is moving through.
Meanwhile, the number of connected objects — from kitchen appliances, to wearable devices, to industrial tools — is forecast by research firm IoT Analytics to surpass 27 billion in 2025.
Some consider even that a huge estimate. In 2020 Softbank-backed chip maker ARM forecast 1 trillion connected devices by 2035.
Household and personal robots are going exponential, too:
I could go on (and on), but you get the idea.
The Exponential Age will be characterized by the deep merging of information and physical reality. A merging, that is, of the realm of bits with that of atoms. Via a wave of sensors and microchips, we’re weaving a layer of data through every object, sidewalk, building, factory, vehicle, person, and more. An always-on supernetwork is blinking into life around us, and it’s happening now.
And with that idea in place, we’re empowered to make more sense of the role machine intelligence will play in the years ahead.
Because while that supernetwork is emerging, another monumental shift is underway. For the first time ever, intelligence is becoming abundant.
Throughout history, intelligence in its highest and most usable form has been the preserve of we humans. In other words, it’s been rare. Now, that truth is about to be flipped. There’s going to be super-abundant intelligence, on tap, everywhere.
You can catch an indirect glimpse of this via the insane exponential increase in the parameter size of AI models across the last couple of decades:
Jensen Huang, the CEO of NVIDIA, predicts we’ll have AI models 1 million times more powerful than ChatGPT within the next ten years.
With all this in place, we’re on the verge of an awakening moment; that flash of white light where you finally realize, wait, I see something new…
So we’re seeing the deep merging of information and physical reality — the emergence of a unified digital-physical field. Meanwhile, for the first time in history intelligence is becoming superabundant.
What role does intelligence play in that new world? Here’s my best answer. When information becomes part of the physical world around us via the unified digital-physical field, then intelligence becomes something akin to a new form of energy, which does useful work inside that field.
Under this view, AI isn’t best understood as a tool; as part of the Productivity Layer.
Instead, it belongs in the Foundation Layer, alongside energy, data, and compute. We need to understand AI as part of the foundational infrastructure that drives everything going on above it — the Productivity, Value, and Human Layers — in the Exponential Age.
This phenomenon is so strange and counter-intuitive, it’s hard to find language to fully describe it. We are thrown, instead, into the realms of metaphor; or, to put it another way, the realm of models.
In the Exponential Age, abundant intelligence is best modeled as akin to a strange new form of electricity. That is, as a source of energy that fuels an army of industrial and household robots, billions of AVs, 3D printers, connected factories, smart cities, and more. Real (physical) energy will do useful work on the mass of all those entities. Machine intelligence will do useful work on the information. And a robot is just as much an informational entity as it is a physical one.
This is a non-standard analysis.
We’re used, after all, to thinking of intelligence as something intangible and abstract; the apex characteristic that differentiates we humans from all other beings. When we realise that in the decades ahead intelligence becomes foundational and a form of infrastructure, we’re uncovering a deep and counter-intuitive truth about the manifestations of AI the Exponential Age, and one we’ll return to time and again in future essays.
But before I move on to the magic side of my AI framework I want to go one step deeper, to the bedrock of the electricity side.
Nowhere is the power of this framework more evident than when we look at a specific phenomenon set to unfold across the coming years. That is, the special relationship set to emerge between machine intelligence and (real, physical) energy.
It’s via this relationship that AI really achieves its most fundamental incarnation as infrastructure — as itself akin to a form of fuel.
The relationship I’m talking about is at the very foundation of the entire Exponential Age. It’s the most important technological relationship inside the new world we’re building.
In fact, it might just be the most important technological relationship ever…
From here, the essay goes on to a detailed discussion of the symbiotic relationship that is emerging between AI and physical energy. I call this relationship the AI/energy flywheel.
The essay then moves on to the other side of my overarching framework: AI as magic.
Members of The Exponentialist can read this and all monthly essays in full. The service also includes monthly video presentations and Q&As with me, access to my ongoing deep dive conversations with Raoul Pal, and more.
If you’re interested in joining The Exponentialist, just go here for more details.