Christopher: Your smartphone is about five million times faster than a single neuron in your brain. Yet, you can catch a flying baseball without thinking, and your phone can't. Today, we explore why the secret to intelligence isn't speed, but memory. Lucas: Hold on, five million times faster? That can't be right. My brain feels pretty quick. How is it possible that the little computer in my pocket is that much more powerful than the supercomputer in my skull? Christopher: It's a paradox that sits at the heart of the book we're diving into today, On Intelligence by Jeff Hawkins. And what's wild is that Hawkins isn't a traditional neuroscientist. He's the engineer who invented the PalmPilot. He came at the brain with a problem-solver's mindset, completely frustrated that the world of Artificial Intelligence was getting it all wrong. Lucas: Oh, I like that. An outsider's perspective. So he’s not just looking at brain scans; he’s trying to reverse-engineer the thing. That already feels more practical. Christopher: Exactly. He argues that for decades, we've been obsessed with the wrong analogy for the brain. And getting that analogy right changes everything.

Lucas: Okay, so if the brain isn't a computer, what is it? Because "brain as a computer" is pretty much the only metaphor I've ever heard. Christopher: Hawkins starts by blowing that metaphor up. He points to something called the "one hundred-step rule." Think about it: a neuron can fire, at most, about 200 times a second. When you see a ball flying towards you and instantly move your hand to catch it, that entire process, from light hitting your retina to your muscles moving, takes less than a second. This means the brain can't be performing more than about a hundred sequential "steps" or calculations. Lucas: A hundred steps? A modern computer could do billions of calculations in that same time. So how does the brain pull off these miracles? Christopher: Here's the core idea: the brain doesn't compute the answer; it retrieves it from memory. Think about catching that ball. A robot would need to calculate the ball's trajectory, velocity, the arc, wind resistance—it's a massive physics problem. It would be crunching numbers like crazy. Lucas: Right, a constant stream of calculations. Christopher: But your brain doesn't do that. Through practice, it has stored a vast library of memories of catching things. When you see the ball, your brain doesn't calculate its path. It instantly recognizes the pattern—"oh, this looks like that throw from last week"—and retrieves the appropriate motor program, the sequence of muscle commands for "catch ball." It's a memory recall, not a calculation. Lucas: That makes so much sense. But when you say memory, I think of a computer's hard drive, just storing files. Is that what he means? Christopher: Not at all. And this is a crucial distinction. Computer memory is like a library with a perfect card catalog. You give it an address, like "file C-47," and it gives you the exact data stored there. It's precise, but dumb. It doesn't know what's in the file. Lucas: Okay, so how is brain memory different? Christopher: Hawkins describes cortical memory as having three key properties that are completely different. First, it stores sequences. You don't just remember a face; you remember the sequence of seeing the face, hearing the voice, feeling a handshake. Memories are stories, not snapshots. Lucas: It’s like a song, then. You can't just jump to the middle of the chorus in your head; you usually have to play the verse first to get there. The sequence matters. Christopher: That's a perfect analogy. The second property is that it's "auto-associative." This means you can recall the whole memory from just a tiny, partial clue. You hear the first three notes of that song, and your brain automatically completes the rest of the melody. You smell a certain perfume, and a whole memory of a person or a place floods back. Lucas: Right, you don't need the whole "file" to find it. A fragment is enough. What's the third one? Christopher: This one is the most important: it stores patterns in an "invariant form." This is a bit of a technical term, but the idea is simple. Think about your best friend's face. You can recognize it whether they're close up or far away, in bright light or shadow, smiling or frowning. The actual pattern of light hitting your retina is completely different in every single one of those cases. Lucas: Huh. I've never thought about that. The raw data is always different, but my perception is the same: "that's my friend." Christopher: Exactly. Your brain isn't storing thousands of different photos of your friend. It has extracted the essential relationships—the proportions of their features, the relative structure—and created a single, stable, invariant representation. It's a memory of "friend-ness," not a specific image. This is something computers are terrible at. Lucas: Wow. So the brain isn't a calculator at all. It's a massive, interconnected library of stories and patterns that it can retrieve and recognize in a flash. That feels like a much more elegant system. Christopher: It's incredibly elegant. And that sequential, pattern-based memory system is the key to the brain's real superpower, which Hawkins argues is prediction.

Lucas: Okay, prediction. That sounds like fortune-telling. Isn't that just a fancy word for guessing? How is that the foundation of all intelligence? Christopher: It's less about seeing the distant future and more about constantly, unconsciously predicting the next microsecond. Hawkins uses a brilliant thought experiment: the Altered Door. Imagine you come home from work. You reach for your front doorknob. You do this every day, without thinking. Lucas: Right, it's pure muscle memory. Christopher: But what your brain is actually doing is making hundreds of low-level predictions. It predicts the exact feel of the cold metal, the specific resistance as you turn it, the familiar click of the latch, the weight of the door as it swings open. It's a symphony of tiny, expected sensations. Lucas: I'm with you. So what happens? Christopher: Now, imagine that while you were out, I snuck over and moved the doorknob two inches to the left. You walk up, reach for the knob where it's supposed to be, and your hand finds... nothing. What happens in your brain at that exact moment? Lucas: Whoa. Instant alarm bells. I'd stop, be confused, and my full attention would snap to the door. I'd be consciously aware of it for the first time all day. Christopher: Precisely! That jolt of surprise is what Hawkins calls a "prediction error." Your brain predicted "doorknob here," but the sensory input was "no doorknob." That error signal shot up your cortical hierarchy, demanding attention from your conscious mind. Hawkins argues that this is the essence of how the brain works. We are constantly swimming in a sea of correct predictions, which we call "understanding" or "reality." It's only when a prediction fails that we become consciously aware. Lucas: That's a wild idea. So, intelligence isn't the thinking we do; it's the thousands of things we don't have to think about because our brain predicted them correctly. It's like missing the last step on a staircase in the dark! Your body lurches because it predicted a step that wasn't there. Christopher: That's the perfect example! Your brain is an organ of prediction. And the evidence for this is stunning. There was a famous experiment where scientists rewired the brains of newborn ferrets. They took the nerve that goes from the eye and, instead of letting it connect to the visual cortex, they plugged it into the auditory cortex—the part of the brain that's supposed to process sound. Lucas: You're telling me they made a ferret see with its hearing-brain? That's insane. Christopher: That's exactly what happened. The auditory cortex, when fed with visual patterns, learned to process them as vision. The ferret could see. This shows that the cortex isn't a collection of specialized tools—a vision tool, a hearing tool. It's a general-purpose learning algorithm. Its job is to find patterns in whatever data it's given and use those patterns to predict what will come next. Lucas: So the brain tissue itself doesn't care if it's getting light waves or sound waves. It just sees patterns and starts building a predictive model of the world based on them. Christopher: Yes. Your entire perception of reality is a model built on predictions. When you look around the room, you feel like you're seeing a stable, continuous world. But your eyes are actually darting around three times a second in jerky movements called saccades. With each saccade, your brain is predicting what it's going to see next. The feeling of stability is an illusion created by a constant stream of successful predictions. Lucas: My mind is a little bit blown right now. So if our brain is just this biological prediction machine, where does something like human creativity come from? Or consciousness? It feels like there has to be more to it.

Christopher: That's the big question, and Hawkins has a surprisingly simple answer. He argues that creativity isn't some magical, separate faculty. It's a direct, logical extension of the memory-prediction framework. Specifically, creativity is prediction by analogy. Lucas: Prediction by analogy? Break that down for me. Christopher: Think of a mathematician staring at a new, difficult equation. She's stuck. But then, she notices a small part of it looks vaguely similar to a different type of problem she solved years ago. Her brain makes an analogy: "This new thing is like that old thing." It then predicts that the methods that worked on the old problem might work on the new one. That leap, that analogical prediction, is the creative act. Lucas: So it's not pulling an idea out of thin air. It's accessing a memory of a different pattern and applying it to a new situation. Christopher: Exactly. It happens all the time. You go to a new restaurant you've never been in before and need to find the restroom. You don't panic. You make an analogy to all the other restaurants you've been in and predict it will probably be in the back, near a sign with a little person on it. That's a small, everyday creative act. Genius is just the ability to make these analogies across very distant and abstract domains. Lucas: Okay, that demystifies creativity a lot. But what about the future of AI? If this is how intelligence works, why are we still building AI that feels so... un-intelligent? Christopher: Because, Hawkins argues, we've been trying to build the wrong thing. We're obsessed with the sci-fi image of a humanoid robot that talks and acts like us. But human intelligence is deeply tied to our bodies, our emotions, our evolutionary history. A "real" AI, a machine built on cortical principles, wouldn't need any of that. Lucas: So what would it look like? Christopher: Think of a smart car. You could give it sensors—cameras, radar, microphones—and hook them up to a hierarchical memory system, just like the cortex. Then you'd train it by having it "experience" millions of miles of driving. It would learn the invariant patterns of traffic, roads, pedestrians. It would build a predictive model of the world of driving. It wouldn't "feel" anything, but it could predict an accident before a human could react because it has seen the precursor patterns a million times. Lucas: That's a much more powerful—and frankly, less scary—vision of AI. It's a tool, a specialized predictive engine. Christopher: And this is where the book, written back in 2004, feels incredibly relevant today. Hawkins was critical of AI for ignoring biology. Now, we have things like ChatGPT, which are essentially massive prediction engines. They predict the next most likely word in a sequence. Lucas: Right! So is that what he was talking about? Is ChatGPT the kind of AI he envisioned? Christopher: It's a step in that direction, but there's a key difference. Hawkins's model is grounded in a hierarchical understanding of the world built from real sensory data. It's not just predicting text; it's building a model of how things are structured. This is a point of debate, and some critics do argue that Hawkins's theory is a bit too neat, especially when it comes to the messy problem of consciousness, which he defines simply as what it feels like to have a neocortex. Lucas: A bit of a hand-wave on that one, maybe. But the core idea is still so powerful. It reframes everything.

Christopher: It really does. The entire book is a journey toward one profound realization: your brain doesn't just passively observe reality; it actively creates it. Every second, it's making a hypothesis about the world—"this is what's about to happen"—and then it just checks to see if it was right. Lucas: So the big takeaway isn't just that our brain predicts things, but that our entire reality—what we perceive, what we understand, even who we are—is a model built from those predictions. Our world is a story our brain tells itself. Christopher: Exactly. And the most powerful thing we can do is question that model. Hawkins points out that something like a negative stereotype is, at its core, just a faulty prediction. It's a lazy, over-generalized pattern that our brain uses. A truly intelligent person, in his view, is someone who is constantly seeking new data to refine and update their predictive models of the world. Lucas: That's a powerful thought. It puts the responsibility on us to be better predictors, to be more curious and less certain. What's a prediction your brain made recently that turned out to be wrong? We'd love to hear your 'altered door' moments. Share them with us on our socials. Christopher: I love that. It’s a challenge to us all to be more aware of the stories our brains are telling us. Lucas: A fantastic and mind-bending read. Thanks, Christopher. Christopher: This is Aibrary, signing off.