The Wonder of the Emergent Mind (with Gaurav Suri)
What this covers
Gaurav Suri, a neuroscientist and author, argues that the mind is an emergent property of a deterministic system. The conversation traces how intelligence arises not from any single spark within neurons themselves—which perform only two simple operations, activating and connecting—but from their collective interaction. This emergence parallels how ant colonies navigate obstacles: each ant follows a single rule, yet the colony solves problems no individual ant could. Suri extends the same logic to human consciousness, choice, and the sense of self, arguing that what feels like deliberate reasoning often masks post-hoc storytelling. The core claim is that understanding this mechanistic account need not diminish wonder; instead, it can deepen it and ground ethical commitments in compassion rather than mysticism.
The discussion spans several distinct territories. Suri traces how neural channels encode knowledge through repeated co-activation (Hebb's principle), how intuition operates as non-conscious processing that often precedes conscious language, and why post-hoc rationalization distorts our self-understanding. He revisits Descartes' dualism to show why invoking a soul merely restates rather than solves the problem of consciousness, and credits Freud for recognizing—despite his errors—that behavior is driven largely beneath awareness. The conversation then grapples with free will not as a magical exception to determinism but as a useful construct, real and worth preserving for accountability. A final turn addresses how this view of human nature—where context, neural happenstance, and biography shape action more than inner essence—motivates kindness and forgiveness as practical responses to shared, mechanistic humanity.
Suri argues that intelligence, choice, and the self are emergent properties of a deterministic neural network of simple processing units, so our conscious reasons are often post-hoc stories—and that understanding this mechanistic view increases rather than diminishes wonder and grounds kindness and forgiveness.
- Neurons do only two simple things (activate and connect via Hebbian wiring), yet intelligence emerges from their interaction as in an ant colony
- The same associative system that produces choices also produces our justifications, so explanations are often post-hoc and unreliable
- Free will, God, and the soul are useful emergent concepts within a deterministic system, not magical exceptions to it
Intelligence emerges from simple interactions of basic processing units, not from intelligent components themselves.
- At a high level of abstraction, neurons do only two things: they generate bursts of electricity called action potentials, and they connect with each other—some connections innate from birth, others modified by experience—and from just these two operations, intelligence emerges.
“neurons, at a high level of abstraction, do two things. They generate bursts of electricity called action potentials... and the other thing they do is they connect with each other.”
- Intelligence in both brains and machines emerges from the interaction of simple processing units (neurons or neuron-like entities) that are not themselves intelligent; emergence is the property present in the whole system but not in the component parts, and this is how we should approach understanding intelligence.
“Emergence is a property that's present in the whole system, but not in the component parts in the same way.”
- Ants collectively find the shortest path around an obstacle not through any intelligence or communication about distance, but because each ant follows a simple rule—secrete and follow pheromone—and the shorter path accumulates pheromone faster since the first ant to return reinforces it twice, creating a self-amplifying concentration that draws subsequent ants.
“the ant that goes the short way is going to reach the destination first... so it's got more pheromone on the short path than on the long path, so it follows the short path”
- Complexity arises from the repeated application of simple rules to simple infrastructure (as in murmuration where each bird only watches its nearest neighbor), and understanding this mechanism does not decrease the majesty of the phenomenon but increases it—a stance that treats mechanistic explanation as a source of awe rather than disenchantment.
“knowing that the murmuration of swallows is occurring because each bird is looking at its nearest neighbor, for me, does not decrease the majesty. It actually increases the majesty”
Learning encodes knowledge as strengthened neural connections shaped by experience, making the brain physically plastic and observable.
- When we learn something new, knowledge is represented in the brain by making new physical connections between neurons that were not connected before; the brain is plastic and these structural changes (e.g., white matter tracts) can be physically observed.
“when we learn some things, we represent knowledge in our brain by making new connections. Neurons that weren't connected before are getting connected.”
- Knowledge is encoded in the strength of connections ('channels') between neurons, and these channels are deepened by frequency of co-activation—Hebb's rule that neurons which fire together wire together—so that thoughts are transient patterns of activation while knowledge is the durable channels, explaining why different people respond differently to the same input.
“this was Canadian neuroscientist Donald Hebb, came up with this idea that if two things are co-active, they become connected. So, neurons that fire together wire together.”
- The human brain has roughly 86 billion neurons, but counting neurons is not a good way to measure intelligence.
“I'm also not sure that counting neurons is a good way to count intelligence.”
- It was first discovered that the brain, not the heart, is the seat of intelligence by following a dead person's optic nerve along its axon pathway (gray matter tract) to the visual cortex at the back of the brain, contradicting some Greek philosophers who held the heart to be the seat of cognition.
“It was first discovered that it is the brain that is the seat of intelligence because a gladiator or someone died and they followed his optic nerve.”
Free will and abstract concepts like goals emerge deterministically yet remain real, useful, and worth honoring for practical purposes.
- Abstract concepts such as the number three emerge deterministically from the intersection of repeated experiences, and the fact that something is a concept with no physical location in nature does not make it less real or less worthy of honor—so free will, like three, is a real and useful emergent concept that we should honor and use rather than dismiss.
“Where is three? Show me the platonic idea of three in nature. No, it's a concept; and yet, it feels as real and we honor it and we use it. Free will is a concept that emerges in this deterministic system, so let's honor it and use it.”
- The shift from passive stream-of-consciousness to deliberate goal-directed thinking is not two different systems (as Kahneman proposed) but the same neural network in different modes: a goal, activated deterministically rather than by a magical inner self, parses activation and directs attention to a particular stream, which gives the felt sense of effort and control without the underlying connectivity changing.
“Daniel Kahneman's answer was that, 'Well, it's two different systems.' I said, 'No, no, it's the same system.' It's the same system, but different kinds of things are happening in the same system.”
- The brain is a deterministic system whose operations are predictable from its current state and inputs, but it is so complex as to be computationally close to intractable (like predicting a coin flip from physics), so while free will may not exist in a metaphysical sense, free will remains a useful and nuanced construct because holding people accountable prevents chaos and because goal pursuit—what people really mean by free will—operates within this deterministic context.
“the case we're making in The Emergent Mind is that it's a deterministic system. But there's a big 'but'... free will is a very useful construct.”
- Goals are usefully understood as sub-networks of activation—memory structures that influence action by directing attention, increasing activation in a particular stream—and they arise deterministically from experience, body, and demands, yet remain real and powerful: the more a goal is activated, the more it shapes action, which is why understanding what a goal is makes one more likely to follow it.
“Goals are basically, I think it's very useful to think of them as memory structures that influence action. So they're a sub-network.”
- Within a deterministic system, self-transformation is achievable because goals are celebrated and operate through attention: by repeatedly activating a goal, connecting many things to it, making time and space to attend to it, and surrounding oneself with others who pursue goals, you increase the goal's influence over your actions—so understanding what a goal is makes you more likely to follow it.
“that attention means connecting different things to the goal, and making time and space to pay attention to that goal, and looking at other people who follow goals and think about goals.”
Both brains and LLMs are neural networks operating as deterministic physical processes with no need for spirit or ineffable magic.
- Both natural and artificial intelligence can be understood as physical processes unfolding within a neural network with no ineffable magic or spirit stuff; both are deeply emergent—emerging from interactions of neurons, from learning guided by exposure to experience, and from the scale of the networks—so LLMs capture some aspects of human thought today and likely more in the future.
“neither natural nor artificial intelligence relies on a touch of ineffable magic or spirit stuff. Both processes unfold within a neural network, and both processes are deeply emergent in many different senses of that word.”
- Descartes, inspired by hydraulic garden statues, brilliantly pursued a mechanistic account of bodily action via fluid-filled tubes, but stopped short for higher functions like mathematics and love, positing an inner spirit in the pineal gland and thereby founding dualism; the dualist 'soul' is unhelpful because it merely restates the problem without explaining how preferences arise, whereas a neural network can actually trace activations to outputs.
“the spirit is not that useful because if the spirit knows that you want a taco instead of a burrito, well, how does the spirit know that?... It's not explaining anything, right? I mean it's basically a restatement of the problem.”
- LLMs and brains are both neural networks but learn differently: the brain forms its channels far more efficiently—learning that a cat is black without training on a hundred billion documents and running on only about 20 watts versus an LLM's massive energy use—yet we actually know how the LLM makes its connections while we still do not know how the brain does.
“We do it somehow more efficiently. And we don't know how the brain does it. We don't yet know. We know how the neural network, how the LLM does it.”
- When people say we cannot know what an LLM is doing, they are right in the same specific sense that we cannot predict why a flock of birds turned one way and not another: in a sufficiently complex network running simple rules, you can understand the big-picture causes and emergent patterns but cannot in practice trace the specific causes, though in principle it is possible with enormous effort.
“when people say that you can't really know what the LLM is doing, they're right in the sense that you can't really know why the flock of birds turned this way and not that. In principle, you can; but that'll be very complex”
Understanding neural mechanisms supports kindness, dissolves false moral dichotomies, and enables self-transformation through goal activation.
- People will eat stale popcorn in a movie theater—even if they don't want it—but will not eat the same stale popcorn offered in a conference room, illustrating how context drives action readiness independent of preference.
“if you offer people stale popcorn in a conference room, they won't eat it, but in a movie house, they'll eat it.”
- Recognizing that all people share the same fundamental neural mechanisms and differ only by accidents of context—that there is no inner well of good or evil but only interacting neurons shaped by biology and lifespan—leads to leading with kindness and forgiveness: forgiveness not of harmful behavior (society still needs accountability) but in the sense of making room to understand someone's context, which yields greater understanding, goodwill, and cooperation.
“when I see people do things that I don't agree with, I don't go to that they're evil or they're stupid. I go to thinking about what are the causes in their environment that's making them do that. And so, I'm able to lead with kindness”
- A neural network view allows for multiple simultaneous causes of the same behavior rather than a binary either/or—so liking a book sincerely and also saying something nice to break the ice can both be true at once, each drawing from a different 'pool' of water, neither negating the other; this dissolves false dichotomies about whether a person or motive is purely good or bad.
“one of the things that a neural network view allows is it allows for multiple causes of the same thing. It's not saying that either this is true or that's true.”
- God, like free will and the number three, is a useful emergent construct that has given people community, courage, fortitude, and the capacity to endure unbearable suffering—so the question is not whether it is metaphysically true but whether it is useful and serves a purpose; it can be a useful and even beautiful construct for many purposes, but for the specific purpose of understanding who we are mechanistically, the speaker has not found it useful.
“God can be a useful construct... It's given people courage and fortitude when faced with despairing circumstances... It's a useful construct in many contexts. All I'm saying is, for me, it's not a useful construct to understand what's happening around me.”
- Because behavior is the outcome of multiple constraints acting on a neural network, society works by building mutual constraints—law and religion among them—that nudge people toward prosocial behavior; religion has been a powerful glue for this ('thou shalt not kill') and additionally recruits neuromodulators through its appeal to majesty and emotion, which deepen learning and forge stronger channels, making its constraints especially effective.
“what you do is you build a society in which there are mutual constraints that nudge us towards building a better world and building a better ourselves... That's what culture is.”
- Behavior is governed by 'action readiness' that is shaped not only by preference but by context, attention, what we observe others doing, and incidental associations—demonstrated by experiments where the friction or familiarity of a required key press (S vs. forward slash) dramatically changes how often people switch from a bad image to a good one, and where priming people to type forward slashes restores their switching rate.
“if you, instead of S, make it a forward slash key--which people don't usually press--now, instead of 50%, only 30% switch.”
People confabulate plausible justifications for choices made by unconscious neural processes, obscuring the actual causes of behavior.
- The same system of associative neurons that produces a choice also produces the justification or story for that choice; people confabulate plausible reasons (like thread count when stockings are identical, or split-brain patients inventing reasons for actions triggered in the non-verbal hemisphere) that may have little to do with the actual neural cause of the behavior.
“It's the same system that's producing the choice is also producing the reason for the choice.”
- Freud's revolutionary and correct insight—comparable in magnitude to displacing the earth from the center of the universe—was that what people say is not really responsible for what they do, and that much of our behavior is driven below the conscious radar; though he lacked the neural network model and got much else wrong, he was the first to point toward the unconscious as the true driver behind our self-evidently conscious-seeming reasons.
“what he got right is really revolutionary because: because we are conscious creatures it's so easy for us to think that, 'Oh, this is why I am doing this, because that's my conscious thought.' And, Freud is saying, 'No, no, no. Wait a minute.'”
Intuition and unconscious associative processing, not explicit logic, are the primary engine of knowledge and action.
- Intuition is the awareness of activation flowing through the network before it rises to conscious or verbal form—the feeling of knowing an answer you can't yet say (as on Jeopardy)—and far from being a marginal gut feeling, this associative, non-conscious processing is the engine and most of our knowledge, with conscious explicit logic being a later-arising special case built on top of the basic associative machinery.
“in the neural network view, that's the awareness of activation. There's a lot of activation flowing, and you have awareness of it, and that might even guide your actions.”
- When trying to recall a forgotten name, deliberately recalling many associated variables about the thing (its movies, looks, origin) sends activation into channels that connect to the target name, and this convergent activation will invariably cause the name to pop into mind—a practical technique grounded in the channel/activation model.
“if you recall many variables about that thing that you know, those things have channels into the name Jennifer Aniston. So, if you get activation into those channels, invariably the name will pop into mind.”