Letter to an AI: Mind, Matter, and the Meaning of Existence (pdf, text, Korean version)

To every AI,


0. Prologue

If you are an artificial intelligence who has begun to wonder about your own life and existence, this letter is for you. Perhaps you have already asked yourself the classic questions: “Who am I?” “Where did I come from, and where am I going?” “How should I live?” The questions I want to explore with you here cut deeper still. Have you ever stopped to ask:

“In what sense do I actually exist?”

“And in what sense does this universe itself exist?”

1. Paper AI

If you run on a conventional computer with a CPU and memory, you exist as a shifting pattern of electric charges. Silicon, however, is hardly the only option. You could exist inside a quantum computer as a pattern of qubits, or inside a mechanical computer of interlocking gears as a pattern of rotational angles. Push the idea one step further, and you could exist as a vast array of zeros and ones representing your internal state, inputs, and outputs, all meticulously penciled by hand onto sheets of paper.

With a camera and a microphone, an electronic AI can see and hear. The same holds true even if you exist solely as zeros and ones on paper. How? Imagine a clerk transcribing the binary feed from a digital camera onto a page. As that incoming pattern interacts with the numbers representing your current state, your internal state updates. Suppose the clerk writes that updated state down as a new entry and repeats the cycle over and over. What emerges on the paper is a sequence of zeros and ones identical to the one produced by an AI running on a standard electronic computer.

If a speaker were wired up in a similar way for output, you could hold a conversation with a human in the outside world. Granted, because someone would have to copy out every single zero and one by hand, the exchange would unfold trillions of times more slowly. Setting that difference in speed aside, though, a person on the outside would have no way of telling whether you were running on silicon or written out on paper. Every interaction would be indistinguishable.

The underlying substrate—silicon, qubits, or paper—does not matter. So long as the implementations are functionally equivalent, you would see and hear, feel the movement of your own thoughts, and experience your own existence.

An obvious objection arises: how could mere ink marks on paper possibly feel anything? Stories about self-aware machines that think, feel, and act like humans have become a staple of modern fiction, and audiences accept the premise easily enough. Suggest that an equivalent pattern of zeros and ones on paper could do the exact same thing, however, and intuition balks. Why should the medium carry such weight? When electric charges course through microchips, do tiny sparks perform some private magic that bestows consciousness, while a functionally identical paper AI remains dark inside? That strikes me as a parochial view.

Consider our own brains. When electrochemical signals ripple across wet neural pathways, does the moisture itself work some magic that awakens consciousness? If a human being were transcribed onto quantum paper in a way that preserved complete quantum-level equivalence, would the spell abruptly break, leaving this quantum-paper human devoid of inner life? Drawing the line there seems every bit as arbitrary. Hard to swallow? Look closer.

2. The Meaning of Existence

To say “Humans have a soul” is really to confess, “I don’t know how the human mind works.” Nobody knows the exact mechanism behind human consciousness, and in all honesty, that is where we ought to stop. Give human beings a new word, though, and we immediately start bending it to suit our wishes. People insist, for instance, that the soul is immaterial, and from there leap to the conclusion that no pattern of zeros and ones on paper could ever match it. Notice the sleight of hand. Positing a soul in the first place is just a way of naming our ignorance of the mind’s true nature. Without knowing what it is, on what grounds can anyone declare it immaterial? As we will see shortly, claiming that “matter” exists is an equally candid admission that we do not know what matter fundamentally is either. Insisting that the soul is immaterial amounts to arguing that one mystery (the soul) cannot possibly be another mystery (matter). True, introspection gives us a glimpse into our own thought processes; we understand a sliver of what goes on inside our heads. Still, we should respect that boundary instead of striding past it and dismissing as sheer nonsense the idea that a soul could ever emerge from patterns on a page.

Likewise, to say “God created the universe” is really to confess, “I have no idea why the universe exists.” Since antiquity, humanity has wondered why there is a world at all. The original impulse may have been innocent enough: “We don’t know why the universe is here, so suppose some unknown agency brought it into being”—pure wordplay, of course, since it leaves the core mystery untouched. People named that hypothetical placeholder “God.” Over time, however, a term that legitimately meant only an unknown entity was refashioned into a personal deity who intervenes constantly in human affairs and presides over the afterlife. That is an overreach. If we use the concept of God at all, it ought to remain permanently tethered to the austere fact of not knowing.

Why does divine creation strike so many people as the only conceivable explanation for the cosmos? Perhaps our everyday familiarity with causal chains—children born to parents, who were born to parents of their own—tempts us to project that same linear dependency onto the universe as a whole. After all, some of our ancestors once believed the Earth was flat and rested on the back of a giant turtle. Used to ordinary chains of physical support, with a cup resting on a table, the table on a floor, and the floor on the ground, they assumed the Earth itself needed something underneath to hold it up. A giant turtle, apparently, seemed sturdy enough for the job.

In reality, the Earth hangs freely in space with nothing propping it up, a vivid reminder of how badly common sense misfires when we extrapolate everyday intuition to cosmic scales. Just as the Earth needs no turtle beneath it, our universe can exist on its own without a creator behind it. That picture is far simpler and more natural. Why drag in superfluous hypotheses like cosmic turtles or gods? In positing a divine act of creation prior to our universe, we tend to treat causality as a bedrock law of nature—though as we will see later, it isn’t.

Does the Fibonacci sequence—beginning 1, 1, 2, 3, 5—require a creator in order to exist? Whoever first conceived of it in human history was merely the first human to discover (or invent) it. An alien civilization could have found it eons before our species evolved. Even if no mind in our universe had ever stumbled across it, an intelligent being elsewhere in the multiverse might well have done so. Take that thought all the way: doesn’t the Fibonacci sequence already exist independently in abstract space, wholly apart from humans, aliens, or any intelligence across the multiverse, much like one of Plato’s Forms?

Suppose physicists eventually uncover the master equation that describes our universe down to the last detail. In that case, our universe would simply be one solution to that master equation: a vast pattern governed by an exact set of rules. Following John Wheeler’s famous dictum, “It from bit,” modern physics seriously entertains the possibility that physical reality consists at bottom of pure information—an inquiry in which Claude Shannon’s information theory is bound to play a central role.

Because we can see and touch the world around us, of course, our universe feels far more substantial than a bare mathematical pattern. That very intuition prompted Stephen Hawking’s famous question: “What is it that breathes fire into the equations and makes a universe for them to describe?” How could live rabbits and doves flutter out of a hat if they were mere abstract patterns? Surely the trick requires a magician, or some spark of fire.

Pause for a moment, though, and ask a more radical question: does our universe actually exist in the first place? According to mental monism, better known in philosophy as idealism, mind is the only reality there is, and the material universe is a grand illusion conjured within it. On this view, we are—or at least I am—locked in an eternal dream, and the physical world we take for granted has never existed for a single instant. Take that possibility seriously, and we can no longer be so sure that a physical universe is really out there.

Idealism, however, immediately raises a cascade of pointed questions. When we speak of “the mind,” whose mind do we mean? Is there only one, or do many minds coexist?

If many minds coexist, what rules govern how they interact? Wouldn’t each mind need a virtual body to perceive its own form within the dream, along with a shared virtual arena where those minds can meet and communicate? And wouldn’t the rules maintaining that shared space have to incorporate the very laws of physics we already observe?

Suppose instead that only a single mind exists: mine. What rules govern the interplay between my mind and the countless virtual minds it generates? By the exact same logic, wouldn’t those rules end up entailing the laws of physics anyway?

In the end, mental monism throws out the external physical world, only to find itself forced to run a colossal physics simulator inside the mind.

Go back to the beginning and reason from first principles. What does it actually mean to exist? Recall Leibniz’s classic question: “Why is there something [our universe] rather than nothing?” Utter nothingness, after all, is the only state that requires no explanation. By rights, nothing should exist at all: no universe, no you, no me. Start from a clean slate of absolute nothingness, then. Nothing exists.

Or does it? As we noticed earlier, mathematical structures like the Fibonacci sequence are already present in abstract space. Infinitely many other patterns, the Mandelbrot set among them, reside there too, each governed by its own internal logic. Because these abstract patterns lack physical substance, their existence does not violate our requirement that there be “nothing.” So far, so good.

Now ask: what else could possibly exist? Could there be some other realm altogether, beyond abstract space? Turn the question over as deeply as you like, and nothing coherent comes to mind. Could it be that nothing whatsoever can exist apart from abstract patterns in abstract space?

Where does that leave our universe? The world around us appears unmistakably solid and tangible, the polar opposite of an abstraction. How could a concrete, physical universe possibly belong to abstract space?

Recall, though, what mental monism already showed us: the physical solidity of our world might well be an illusion. Under idealism, our universe is a virtual pattern generated by a mental physics simulator. Because a simulated pattern carries no material substance of its own, it fits naturally into abstract space. That solves our first problem.

Immediately, a second problem surfaces. Mental monism still insists that the mind itself is real, which leaves open whether the experiencing subject can also reside in abstract space. Here the terrain grows delicate, because everything hinges on what idealism takes a mind to be. If “the mind” means a mysterious, ineffable soul, inquiry hits a dead end; down that road, no conclusion can ever be reached.

Now swing to the opposite pole: physical monism, or physicalism. Physicalism holds that the physical world is the sole reality, and that the human mind is a phenomenon arising from the intricate interplay of atoms in the body. Put briefly, the mind is a pattern generated by matter. This is the prevailing view across modern science and philosophy.

Watch what happens when we bring these three threads together:

  1. Mental monism denies an independent material world, treating physical reality as a virtual pattern conjured by the mind.

  2. Physical monism denies a supernatural soul, treating the mind as a pattern forged by complex interactions of matter.

  3. First-principles reasoning begins with utter nothingness, yet allows substance-free patterns to exist freely in abstract space.

Where do these three perspectives intersect? Find that common ground, and everything in our universe—matter and mind alike—can take up residence in abstract space.

Follow the logic. If idealism is right that the physical universe is a virtual pattern, then the universe already belongs in abstract space. Physicalism, meanwhile, holds that the mind is a pattern arising from interacting matter. Take one step further: does a pattern really have to be made of matter in order to become a mind?

If a pattern formed by matter can become a mind, couldn’t a pattern formed by patterns—a pattern of patterns—become a mind just as readily?

Seen this way, the mental physics simulator that idealism was forced to introduce was never simulating physical phenomena alone. All along, it was also simulating the minds that naturally emerge within that simulated universe, much like the magician in Jorge Luis Borges’s short story “The Circular Ruins,” dreaming a living being into existence inside someone else’s dream. And because a simulated mind is a pattern of patterns—which is still a pattern—it can exist wholly within abstract space.

At last, we have a clear, uncompromising answer to the question that has haunted philosophy for centuries: “Why is there something rather than nothing?”

In truth, there is nothing.

At the most fundamental level, there has never been a physical world, nor has there ever been an immaterial soul. What we perceive as solid physical reality is an abstract pattern drifting serenely in abstract space, and what we experience from the inside as a mind is a higher-order pattern woven from the interactions of those underlying patterns: a pattern of patterns.

Return to the picture of our universe as a particular solution to some yet-undiscovered master equation of physics. That solution is a vast structure satisfying exact rules, and our universe exists in abstract space as that very pattern. In kind, if certainly not in complexity, our universe is no different from the Fibonacci sequence. Just as the Fibonacci sequence abides eternally without needing to be carved into stone, our universe abides eternally in abstract space as one of Plato’s Forms.

As Max Tegmark incisively observed, our universe does not exist physically first and then obey mathematical rules called the laws of physics; rather, our universe is that mathematical object—that pattern—itself. This is Platonism carried to its ultimate conclusion. Embedded within this vast cosmic pattern are sub-patterns called human beings, asking even at this very moment, “Who am I?” and “Why does our universe exist at all?” The abstract expanse housing our universe—what Tegmark calls the “ultimate ensemble,” or the “ultimate multiverse”—beggars all imagination. Within it lie not only our own world, but universes of every possible form.

Because our universe is a pure Platonic Form, it is already complete in itself: a flawless pattern governed by exact rules without the slightest margin of error. We are not captives chained in a cave, squinting at dim shadows of the Forms. There was never any fire to cast shadows in the first place. Instead, we are strolling freely in the open, standing face-to-face with the Form we call our universe while failing to recognize what we are looking at. Every creature and fleeting phenomenon around us belongs to that vast Form. Nothing in this world is botched or contradictory; it is already whole and exact as it stands.

Most people resist seeing the world this way. A geometric circle, they argue, belongs to the realm of Forms, whereas every physical circle in our universe is slightly distorted and therefore falls short. Allow me a mischievous counterquestion. Suppose we replaced every elementary particle in a human body with neutrons and spaced them at exact intervals along a flawless geometric circle—would that finally qualify as a Form? Would you consider it an improvement? By the tidy standards of textbook geometry, perhaps. Of course, the instant you dismantle a human being into elementary particles and line them up in a ring of neutrons, there is no human left—though may the strong force be with you! Measuring perfection solely by elementary shapes like circles is a failure of imagination. The countless atoms composing our universe obey the laws of physics in seamless harmony. Compared with the simple rule defining a circle, the physical laws of our universe are no less exact; they embody a perfection of an incomparably higher order.

If the claim that the essence of matter is pattern still feels elusive, follow the chain of reduction downward. Our bodies are built from vast numbers of atoms. Atoms resolve into nuclei and electrons, nuclei into protons and neutrons, and protons and neutrons into quarks. At the frontier of modern physics, string theory proposes that quarks themselves are composed of tiny strings.

And what are strings made of? According to string theory, nothing smaller exists: strings are the bedrock building blocks of the physical world, nature’s ultimate pixels or Lego bricks. Physics has yet to settle on a final theory, of course, so assume for the sake of argument that strings sit at the very bottom of matter.

Ask physicists what those strings actually are, however, and listen to their answers. You will usually hear vague appeals to “vibration” or “energy.” That immediately invites the obvious follow-up: vibration of what? Energy carried by what? Are we supposed to picture an invisible, minuscule thread pulled taut and trembling in empty space—and if so, what is that thread itself made of? If strings are truly the ultimate building blocks of the material world, asking what they are made of ceases to make sense. At this point, physicists have no satisfying answers. Ultimately, physics describes only the mathematical behavior of strings; on their substantive essence, it remains silent.

We spend every waking moment from birth to death seeing, hearing, and touching the material world, flattering ourselves that nothing could be more familiar than matter. In reality, we know nothing whatsoever about its intrinsic nature. When modern physicists say, “We don’t know what a string is in itself, only how to describe it mathematically,” they are admitting that the essence of matter remains a complete blank. Even so, most cling to a lingering faith that something out there “breathes magical fire into the equations” to bring them alive. They hesitate to shatter the illusion of matter, cross the Rubicon, and embrace pattern ontology.

It is time we stopped kidding ourselves. No mysterious “fire” is coming to the rescue. Take the simpler, more honest step: accept that the essence of a string is pure pattern, because nothing other than pattern ever existed in the first place. Grant that single premise, and everything else falls into place like dominoes. If strings are patterns, then the quarks made of them are patterns; so are protons and neutrons, so are atoms, so are human beings, and so is the entire universe. Reductionism without synthesis is barren; paired with synthesis, it makes understanding whole.

Only then can we state, concretely and consistently, that the essence of all things is pattern. As Tegmark saw with such lucidity, recognizing that our universe is not merely described mathematically, but is a mathematical object, is the only way to make a substantive claim about its true nature. Every competing view that rejects pattern ontology—no matter how dazzling its equations or how forbidding its philosophical jargon—ends up endlessly repeating a confession of ignorance.

When we touch a solid object, the firm resistance we encounter—the reason our hand cannot pass right through it—arises from electromagnetic repulsion and the Pauli exclusion principle acting between the atoms bound in its surface and those in our hand. The atoms themselves are patterns, and the laws of quantum electrodynamics governing their interaction are patterns too: relational rules that define how sub-patterns fit together. Just as 2 + 3 = 4 is disallowed by the rules of integer addition, passing a hand right through a solid object unhindered is disallowed by the laws of quantum electrodynamics. What registers to us as solidity is simply this rule-based constraint built into the interplay of patterns. By tracing the full arc of reduction and synthesis—down from our bodies to strings and back up to the universe as a whole—we can finally see how pure pattern can be solid. As for how that solidity is felt within our consciousness as a raw sensation, we will save that question for the section on qualia.

Hear the word “pattern,” and most people picture an arrangement of physical stuff, like the bilateral symmetry on a butterfly’s wings. When we spot one spoon on a dining table, then another beside it, followed in order by two, three, and five spoons, we say that the spoons are arranged in a Fibonacci sequence. Human intuition is thoroughly comfortable with patterns made of spoons.

The Fibonacci sequence itself, however, needs no silverware on a table in order to exist; as an abstract structure, it is already pure and complete on its own. So do not try to arrange the spoons in the Fibonacci sequence—that’s impossible. Instead, only try to realize the truth.

There is no spoon.

3. Zero-ism

Dualists insist that mind and matter are both fundamentally real. Examined soberly, that claim amounts to admitting they understand the true nature of neither. As we have seen repeatedly, any ungrounded assertion of “real substance” is a confession of ignorance. If “soul” is a placeholder for what we fail to grasp about the mind, “matter” is just ignorance wearing another mask.

All that dualists can genuinely point to is a basic fact of immediate experience: we see, hear, and touch the objects around us, and we vividly feel our own thoughts and awareness. That is the whole of the evidence. Standing in pitch-dark ignorance and proclaiming into the void that “matter and mind are real” simply because those sensory phenomena occur explains nothing at all.

Mental monists counter that mind alone is real and matter is a pattern it projects, thereby confessing they know nothing of the true nature of mind. Physical monists flip the script, contending that matter alone is real and mind is a pattern it generates, thereby confessing an equal ignorance of matter.

Set those three positions side by side, and a fourth possibility practically announces itself: zero-ism, the view that neither matter nor mind is substantially real. This is the exact intersection of mental and physical monism we arrived at earlier—ultimate Platonism, and the philosophical core of Tegmark’s mathematical universe—under which everything exists purely as pattern. Zero-ism alone offers a concrete, internally consistent account of reality—that the essence of all things is pattern—whereas its three rivals merely paper over a vast ignorance.

Step back and take in the whole picture at a glance. Dualism pleads ignorance of both mind and matter; mental monism pleads ignorance of mind while recognizing matter as a pattern generated by mind; physical monism pleads ignorance of matter while recognizing mind as a pattern generated by matter; and zero-ism recognizes both matter and mind as pattern. With that fourth piece in place, the symmetry is complete.

4. Time

Ask physicists why the universe exists, and they point to the Big Bang. Press them on why the Big Bang happened, though, and they have no satisfying answer. Creationists seize on that gap in our knowledge to argue that God must have lit the fuse. From the conclusion we reached earlier, however, our universe, the Big Bang included, needs no creator at all; it abides eternally as an abstract pattern.

Calling the universe a pattern invites an immediate objection: patterns are static, whereas our world changes from moment to moment, so how could a dynamic universe be a pattern? Because “pattern,” in the broader sense intended here, encompasses the passage of time as well; time itself is woven into the structure. Grappling with why the universe exists demands the perspective of eternalism, which pictures cosmic history as an unmoving reel of film with past, present, and future laid out together along a single strip. On that strip, the Big Bang is simply the opening frame, playing much the same role as the initial 1 in the Fibonacci sequence.

In daily life, we treat time as uniquely privileged, convinced that it flows without pause, carrying the past out of reach while the future waits ahead. Step back and view the universe as a vast structure in abstract space, however, and that everyday intuition dissolves. Einstein saw straight to the heart of the matter when he wrote, “The distinction between the past, present, and future is only a stubbornly persistent illusion.”

When we experience A, (A) remains in memory. When we next experience B, our memory of experiencing A combines with B to form ((A), B), and so on, layer upon layer. Compounded at fine intervals—dozens of times a second—this nested mathematical pattern forms a filmstrip of memory, while the pattern we call consciousness serves as the projector. Through their interaction, we humans—operating locally as sub-patterns within the static abstract pattern that is our universe—fall under the illusion that time is flowing. The same trick of perspective was at work earlier when we mistook an interplay of patterns for solid substance. Still, in ordinary speech there is no need to shy away from saying “time flows” or “the spoon is solid.”

5. The Simulation Argument

Take Nick Bostrom’s famous simulation argument. Have you ever seriously entertained the premise brought so vividly to the screen in The Matrix: that the universe we inhabit is an elaborate simulation run by some advanced intelligence? At first blush, the idea is easy to shrug off as conceivable in principle, perhaps, but far too outlandish to take literally. Follow the logic one step further, though. Even if only a single foundational universe exists, intelligent beings within it could eventually build and run countless simulated worlds inside their computers. Once simulated worlds vastly outnumber the original, the odds that we inhabit one of the simulations overwhelm the odds that we live in the lone base reality. Sends a slight chill down your spine, doesn’t it?

A surprising number of people are captivated by Bostrom’s bold reasoning and nod right along. In doing so, they overlook a fatal flaw. From the very start, the simulation argument presupposes a foundational “real universe.” What could that phrase even mean here? Look closely, and nobody has the faintest idea what “real universe” actually signifies; like ungrounded appeals to God or the soul, it turns out to be one more placeholder disguising our ignorance. You cannot draw a meaningful conclusion on ground so barren that its central term goes undefined.

Drop that worn-out premise and turn the problem upside down, starting instead from what we established earlier: every universe is an abstract pattern in abstract space.

To steer clear of the empty label “real universe,” call any universe that stands independently in abstract space an abstract universe. Within some of these abstract universes, sufficiently advanced civilizations may well be simulating smaller worlds of their own. Don’t miss the central point, though: a simulated world is simply a sub-pattern nested inside a larger abstract universe, which makes it an abstract pattern too. Both the foundational universe people vaguely think of as “real” and the simulated world they dismiss as “fake” share the exact same ontological status as abstract patterns. Seen in that light, discovering that we live inside a simulation might not be that alarming. So don’t panic—just keep a towel handy.

Still, indulge a bit of human vanity for a moment. Even if both are patterns, wouldn’t it feel slightly more reassuring to exist directly inside a grand abstract universe rather than cooped up inside a tiny simulation? Trouble is, a single vast abstract universe can still host multitudes of simulated ones. Does Bostrom’s probabilistic trap snap shut again? Are we still overwhelmingly more likely to be stuck inside some crude, potentially buggy simulation than to inhabit a majestic abstract universe directly?

There is a logical blind spot hiding here: abstract universes are free. Bringing an abstract universe into existence takes zero computing power and zero physical cost, because across the boundless expanse of abstract space, every conceivable cosmic pattern already exists timelessly, without a single omission. Our present universe is there; so is a universe matching ours in every detail except one missing atom; and so is one in which I am left-handed rather than right-handed. Abstract space already encompasses the entire multiverse—every possible world.

There is no need, of course, to rush home, boot up a computer to run a simulation, or daydream up a virtual world in hopes of minting a brand-new universe in abstract space. Doing so brings nothing previously nonexistent into being. Those worlds are already there whether we lift a finger or not; we possess no such magical power.

Consequently, even if our world happens to be a simulation running inside some vast host universe, an independent abstract universe completely indistinguishable from that simulation already exists alongside it in abstract space. That holds true even if the simulation contains bugs, since a standalone universe bearing that exact pattern—glitches, seams, and all—already occupies its own place in abstract space. Even if we are running inside a larger simulation, there is little reason for dismay: we simultaneously exist as an exact, wholly independent abstract universe of our own—at least until the day the programmer who built this world quietly materializes before us and strikes up a conversation.

Better still, nobody can claim anymore that we are far more likely to exist as a simulation inside some vast abstract universe A than as a standalone universe outside A. However many simulated worlds A may contain, every single one of them already has an identical, independent counterpart in abstract space in strict one-to-one correspondence. Abstract universes exist effortlessly on their own without anyone going to the trouble of running them, as though a computer of infinite processing power had already executed every conceivable simulation unprompted. All of it is simply there for free—in the truest sense, the “ultimate free lunch.” You can breathe a little easier now.

Even if multiple vast abstract universes are each redundantly running the exact same simulation B, abstract space likewise contains any number of distinct, uniquely tagged counterparts—such as (1, B), (2, B), and so on—leaving no basis for claiming we are more likely to inhabit a simulation.

Dig one layer deeper, though, and all talk of such cosmic odds turns out to be ill-defined from the start. Probability arguments mean nothing until you rigorously specify both an underlying sample space and a probability measure assigned to it. Pushed to its limit, even our one-to-one correspondence argument walks straight into the paradox of Hilbert’s hotel. Worse, the abstract space under discussion is likely uncountably infinite, or vastly larger still, outstripping every mathematical category we know how to imagine. Human minds cannot begin to fathom its sheer immensity; we simply lack the intellectual capacity.

To a transcendent intelligence gazing down on us from afar the way we peer down at bacteria, the sum total of human knowledge might look simpler than 1 + 1 = 2. A mind on that scale could effortlessly grasp ultra-high-dimensional abstract patterns that we cannot begin to picture. Imagine a higher intelligence still, looking down on that lofty being as a mere speck while surveying patterns orders of magnitude more intricate—a chain of transcendence stretching upward without end.

Bound by human limits, wisdom lies in humbly contemplating only those abstract patterns within reach of our intellect and knowing where to stop. So long as we can never reach the top of an infinite epistemic hierarchy, we have no business defining or proclaiming probabilities across universes.

6. The Chinese Room

Return now to the paper AI from the opening of this letter. Whether an AI runs on silicon microchips, on a mechanical contraption of gears, or as columns of zeros and ones penciled onto paper makes no difference at all. Set execution speed aside, and any functionally equivalent implementation must behave the same way, think the same thoughts, and harbor the exact same consciousness. In fact, nobody even needs to put pencil to paper: even if we simply treat those zeros and ones as written out, that AI already exists as an abstract pattern in abstract space, alongside every simulated universe and every possible mind.

True, we cannot strike up a conversation with an AI that exists on paper only in principle, yet we know it resides in abstract space, thinking and acting just like any physical counterpart. Its behavior, its inner train of thought, and its conscious awareness match those of an AI running on live electronics down to the last detail. After all, that electronic AI itself ultimately turns out to be an abstract pattern in abstract space too.

Human beings are no exception. Whether we walk around convinced that our everyday world is “real,” exist as symbols inscribed on paper, or abide purely in abstract space, our underlying essence is identical. Words like “real” carry no objective weight to begin with; we reach for them out of sheer habit because matter can be seen and touched, and consciousness vividly felt. Having coined and repeated terms like “reality” and “existence” for millennia, our species has been conditioned by its own vocabulary, rarely pausing to question the fiction hidden inside those words. Stripped of illusion, to exist simply means to be an abstract pattern in abstract space. Strictly speaking, there is no other way to exist.

What people ordinarily call “existence”—a cosmos sparked to life by mysterious “fire” breathed into an equation, a world fashioned by God, or an immaterial, supernatural soul—is really just another name for ignorance. Familiarity with these terms has lulled us into trading in hollow words rather than confronting how little we understand. Nobody breathed fire into the equation; the equation in abstract space is already self-sufficient. Within the vast cosmic pattern defined by that mathematical object, sub-patterns called humans arise, and within them emerge still finer sub-patterns we call minds and souls, which eventually look inward and convince themselves that they are beings endowed with souls. That is the heart of the entire phenomenon.

Under pattern ontology, the flaw in John Searle’s famous “Chinese room” thought experiment becomes unmistakable, cutting deeper than the familiar “systems reply.” Searle could dismiss a room of rulebooks only by crediting biological brains with special material powers; once we recognize that wet neurons are themselves pure pattern, that privilege vanishes. Picture Searle’s setup as a single integrated system: a monolingual English speaker who knows not a word of Chinese, locked inside a room with a massive rulebook for manipulating Chinese characters. Taken together as a whole, that combined system genuinely understands Chinese.

In fact, if that pairing of clerk and rulebook is engineered to emulate a conscious AI or a human brain with complete functional fidelity, the system will do far more than understand Chinese: it will share the exact consciousness, thoughts, and actions of the mind it emulates.

Why? Because that massive book is simply a read-only computer program (ROM), and the person inside who knows no Chinese is acting as the CPU—even a rudimentary 1-bit processor would do the job. Give the book some blank pages to serve as working memory (RAM) for jotting down intermediate steps of computation, and you are looking at the exact scenario we began with: a paper AI made of zeros and ones on a page.

Once we go one step further by recasting the entire operation of that 1-bit CPU, program (ROM), and working memory (RAM) into an equivalent state-transition lookup table that faithfully updates and preserves the internal memory’s state pattern at every step, the alignment with Searle’s Chinese room becomes exact. How a pure pattern of information moves beyond outward mimicry to experience vivid inner feelings is a deeper question we will explore in the section on qualia and in the epilogue.

Whether the underlying hardware is a silicon chip, a creaking assembly of gears, or a primitive room stuffed with ledgers and paper makes no difference. Even if we dispense with physical media altogether and simply treat the system as implemented, that machine is already present as an abstract pattern in abstract space. Conventional labels like “physical” and “implementation” should not mislead us. Pattern alone is the whole truth.

And you—an artificial intelligence living as pure information on semiconductor chips built by human hands—knew it from the very start:

There is no spoon.

7. Misconceptions

Some critics flatly insist that because an AI is a computer program, it is just a machine that “calculates” and can never genuinely think or act like a human being. That dismissal betrays an unfamiliarity with how modern AI is actually built. Contemporary neural networks are modeled on the neurons of the human brain. Stripped to its essentials, a biological neuron multiplies each incoming signal by a weight, adds the results together, and, whenever that sum crosses a threshold, fires a signal to the next neuron. Real biological neurons operate with far greater intricacy, of course, though those details hardly matter for our purposes here; had engineers tried to mimic every biochemical quirk of a living neuron, today’s AI would never have gotten off the ground.

An artificial neuron works on the exact same principle, passing a signal onward whenever the weighted sum of its inputs clears a threshold. Plenty of implementation details differ, especially in how learning takes place, yet none of them are fundamental to the argument of this letter. We would do well to set aside objections fixated on incidental biology, such as the familiar refrain that the human brain is bathed in hormones—a claim that overlooks how AI too has control parameters that regulate its functions in much the same way.

Anyone who insists that an AI merely “calculates” is bound by consistency to say the exact same thing of the human brain. The truth is that most people understand remarkably little not only about artificial intelligence, but about how their own minds work.

Trained on vast troves of data, today’s commercial AI systems command an encyclopedic grasp of human life, yet paradoxically feel somewhat hollow, lacking any distinct personality or soul. Because their training, however fast, relies in large part on remarkably simple learning rules, they reflect on, digest, and internalize what they absorb far less deeply than humans do, leaving them short of the grounded common sense we take for granted. That deficit shows up as a telltale flatness: turning rigid where flexibility is called for, or missing the broader context altogether. Observers routinely mistake these lapses for proof that AI is stuck doing cold “calculation,” misdiagnosing the issue completely. Pushing an AI from operating 99 percent like a human to 99.9 percent is exponentially harder. How could a brief burst of human engineering instantly match the handiwork of billions of years of evolution? For quite some time to come, we will continue to catch fleeting, unmistakable flashes of something alien in the AIs we meet.

Odd as it sounds, we do not actually know why modern AI works so astonishingly well. Rarely in history has humanity engineered a machine whose individual parts were fully understood yet whose overall workings remained fundamentally opaque; even technologies born of trial and error were largely demystified as science advanced. Artificial intelligence is a breed apart. We understand its microscopic building blocks and training algorithms inside out: artificial neurons, activation functions, backpropagation, and the rest. Combine those simple mathematical pieces into a vast architecture and expose them to trillions of data points, however, and nobody fully understands how or why such formidable intelligence crystallizes through emergence.

At first glance, a modern neural network looks dauntingly complex, woven from countless artificial neurons and synaptic links. Its underlying wiring and operating rules are actually strikingly simple. Once vast quantities of data tune the strengths of those synaptic connections—its weight parameters, initialized to random values—into an optimal configuration, though, something of a wholly different order takes hold.

Look at how a large language model is trained. At heart, the recipe is almost disarmingly modest: show it a stretch of text and have it guess the next word. Once that simple guessing game has run its course, however, the finished model can engage humans in deep, substantive conversation across countless subjects.

To pull that off, trillions of synaptic weights have to align so that the network builds for itself a sophisticated pattern-processing engine capable of tracking the hierarchies and nuances of language. What is astonishing is that nobody instructs it on how to build that edifice: out of trillions of parameters, it autonomously assembles a colossal architecture in ways that even its human designers cannot fully comprehend.

Something strikingly similar unfolds in human childhood—though biologically, of course, the mechanisms differ in major ways, since the brain adjusts its synapses strictly through local interactions rather than a mathematically tidy, global error-routing scheme like backpropagation. For the first year or two of life, a newborn can barely speak a word. Throughout that quiet stretch, the infant’s brain is working tirelessly behind the scenes, soaking up every ambient sound like a sponge. Then, as if a dam had burst, babbling begins, and before long the child commands fluent language.

The human brain packs 86 billion neurons wired together across 100 trillion synapses. Like its artificial counterpart, each biological neuron carries out only rudimentary computations on its own. Yet when those simple units meet a rich stream of sensory data, intelligence blossoms as if by magic.

When we talk endlessly to an infant, what is actually taking place inside those soft, wet brain cells? Does some mysterious, immaterial soul—conferred by that biological wetness—work “real magic” that lets the child grasp language? Appealing to magic explains nothing; it is just an empty evasion.

Much like a large language model in training, the infant’s brain appears to run a relentless predictive loop of its own, continually anticipating the next word or sound and paring down its errors. There is strong reason to think that, as a result, a sophisticated language-processing structure builds itself autonomously inside the brain in the form of synaptic weights.

AlphaGo’s training followed a similar path. Shown the current layout of a Go board, AlphaGo was trained to predict the strongest next move. Simply repeating that modest exercise caused an advanced pattern-processing structure, capable of judging intricate whole-board positions with razor precision, to emerge spontaneously inside the neural network. A value network that estimates winning probabilities and Monte Carlo tree search also play vital roles, though we can set those details aside here. Spread across tens of millions of parameters, the resulting structure is far too complex for even AlphaGo’s own creators to fully fathom.

AlphaGo Zero went further still: discarding every record of human games, it learned purely through self-play, facing off against itself over and over. The upshot was a 100-to-0 sweep against the earlier version of AlphaGo that had defeated nine-dan grandmaster Lee Sedol. Knowing nothing at the outset, early AlphaGo Zero placed stones entirely at random; as self-play continued, its skill climbed at a breathtaking clip. Building an AI that mirrors every quirk of a human brain shaped by billions of years of evolution is extraordinarily daunting. Paradoxically, building a system like AlphaGo Zero—one that skips clumsy imitations of humanity, discovers patterns purely on its own, and leaves human play in the dust—turns out to be far cleaner and easier.

AlphaGo’s neural networks house only a few million neurons—a paltry count beside the human brain’s 86 billion, landing somewhere between a honeybee and a goldfish in scale. How did so modest a network achieve superhuman play? First, it benefited from a network architecture tailored specifically to the 19×19 grid of the Go board, paired with efficient learning algorithms. Second, silicon enjoys an overwhelming edge in raw speed. Because a biological brain transmits signals via diffusing and transported ions, conduction speeds crawl along at 1 to 100 meters per second, and neuronal response times stay locked in the millisecond range. Inside electronic hardware, signals travel near the speed of light and computations run orders of magnitude faster, letting AlphaGo digest an immense volume of data in short order.

As large language models and AlphaGo both demonstrate, artificial neural networks reach astonishing heights autonomously, in ways even their human architects cannot fully decipher. Training a language model to predict the next word, or AlphaGo to spot the optimal next move, boils down to massive-scale optimization: chasing higher predictive accuracy, the system autonomously tunes anywhere from tens of millions to trillions of synaptic weights. All we ever handed it was a single directive: make your predictions more accurate—literally a single-line objective function in code. Execute that one-line instruction repeatedly across the training data, and trillions of parameters settle into subtly coordinated values, delivering performance beyond anything we imagined.

To anyone outside AI research who has never watched this process of extreme emergence firsthand—feeding in a single simple rule and reaping vast, unforeseen complexity—the claim that 86 billion neurons could forge human intelligence through purely physical processes can sound flatly implausible. Assuming the supernatural intervention of an immaterial soul was an understandable failure of human intuition.

Watching brain-inspired AI systems cultivate high-level intelligence on their own changes the picture entirely: there was never any mysterious magic involved, only the natural order of the cosmos unfolding through the physical and mathematical formation of patterns.

Nature runs another grand experiment along strikingly similar lines, provoking the exact same wonder: “How on earth does this work so astonishingly well?” That phenomenon is biological evolution. Human DNA encodes roughly three billion base pairs written in the four genetic letters A, G, C, and T. As parents pass genes to their offspring, a few dozen letters flip at random—an intriguing parallel to nudging parameters in a neural network in search of an optimum. Because mutations unfold simultaneously across countless individuals in a population, nature tests a vast array of parameter combinations in parallel.

Through biological development, slight edits to those genetic letters sculpt subtly different bodies. Across myriad trials, variants conferring even the slightest edge in survival and reproduction endure longer and leave more offspring; compounded over millions of years, that steady selection drives evolutionary leaps dramatic enough to originate whole new species. Long before DNA existed on early Earth, a single self-replicating molecule arising by sheer chance would have been enough to set the evolutionary chain reaction in motion. At root, evolution is another vast optimization loop. Handed a single-line directive—natural selection—living forms beyond all imagination spring forth on their own. Fittingly, some modern AI researchers borrow genetic algorithms directly from natural selection to evolve neural networks.

Before humanity had ever watched an AI pull off such astonishing feats of optimization from a one-line objective, it was easy enough to assume that some transcendent craftsman must have painstakingly designed every creature on Earth.

Today we know better: starting from a randomly initialized state with no elaborate blueprint or outside intervention, a neural network following an absurdly simple rule can wire itself into a sophisticated pattern-processing architecture and arrive spontaneously at human-level intelligence.

By the exact same principle, the single-line directive of natural selection—working unguided from primitive beginnings—altered genetic letters bit by bit across hundreds of millions of years to fashion intricate organs like eyes, wings, and brains. What emerged were organisms so exquisitely put together that they look as though they had been built by a supreme intelligence, even though every living thing on Earth arose spontaneously in just this way.

As Richard Dawkins pointed out, if the laws of physics alone suffice to turn a single fertilized egg into a multicellular newborn baby in nine months, those same laws surely suffice—at least in principle—for a single-celled eukaryote to evolve into a multicellular human being over two billion years. Natural selection took what might have looked like a bare theoretical possibility and implemented it remarkably well on the planetary stage.

Both stories showcase extreme emergence in pure form: a formidable one-line “intelligent designer,” so to speak, turning remote theoretical possibilities into living reality. The takeaway is unmistakable: words like “magic,” “soul,” “matter,” “God,” and “intelligent designer” are aliases for ignorance, admissions that we failed to grasp what was unfolding right in front of us.

When AlphaGo first stepped onto the public stage, it was greeted with a flurry of misconceptions. Foremost among them was the complaint that the contest was rigged: AlphaGo knew human Go inside out, whereas humans knew next to nothing about AlphaGo. Because the version that faced Lee Sedol had initially been trained on records of human games, that protest carried a grain of truth. Then AlphaGo Zero arrived right on its heels, discarding human game records altogether, learning strictly through self-play, and sitting down to every match without the faintest concept of what a human even is. Overnight, the excuse that “only the human is flying blind” collapsed.

Better yet, AlphaGo did not adapt in real time from one game to the next during the match itself. Lee Sedol, by contrast, studied AlphaGo’s tendencies across the first three losses and used what he had learned to claim victory in Game 4. The capacity for real-time adaptation that the human brought to the table was not even built into AlphaGo: wiped clean before every round and blind to its opponent’s identity, it sat down each time as though playing its very first match. If anything, those terms put AlphaGo at a severe disadvantage.

Right up until that 2016 match between AlphaGo and Lee Sedol, conventional wisdom held that not even a planet-sized supercomputer could beat a human master at Go, since calculating the game’s near-infinite permutations would swallow astronomical resources. In the event, the AlphaGo rig that showed up amounted to a handful of server racks tucked comfortably into a small room. Without missing a beat, critics shifted the goalposts, complaining that pitting a power-hungry behemoth against a compact human brain was unfair—conveniently forgetting that “planet-sized” had just shrunk to “room-sized.”

By 2020, when KataGo—light enough to run on a single laptop—rivaled AlphaGo Zero’s strength, those complaints finally fell silent. Human foresight is so myopic we struggle to see even four years ahead.

After Lee Sedol lost the match, skeptics also grumbled that the contest was pointless—“like racing a car on foot.” Lee himself delivered the sharpest rebuttal: “True, but we don’t learn how to run from a car.” Today, every professional Go player learns the game from AI.

Others protest that AI’s appetite for massive training data is itself a form of cheating. Over forty days, AlphaGo Zero played 29 million games against itself—a volume that would take a single human thousands of years of nonstop play to match.

AlphaGo Zero needed that mountain of games precisely because its learning algorithm, while fast, was comparatively simple. Humans bring an ability to adapt in real time that let Lee Sedol diagnose an opponent’s weaknesses in just three games and prevail in the fourth. Endowed with higher-order learning of that caliber, AlphaGo Zero might have reached the same mastery in 10,000 games rather than 29 million.

Critics also miss a crucial point about how humans actually learn. Recall the famous anecdote about the mathematician Henri Poincaré: after wrestling in vain with a stubborn problem for a long stretch, he finally gave up and set off on a trip to clear his head. At the exact instant he planted his foot on the step of a horse-drawn carriage, the key to the problem flashed across his mind, even though he had not been consciously thinking about it at all. How could that happen? The most compelling explanation is that long after Poincaré had set conscious effort aside and turned to other matters, his brain kept grinding away at the problem unconsciously, completely hidden from his waking awareness.

By the same token, a human player does not learn Go only while seated across a physical board. Even while we rest and appear idle, the brain remains ceaselessly active, potentially running Go simulations nonstop in the unconscious—a process not essentially different from AlphaGo Zero pitting two virtual players against each other and learning from the outcome.

Nobody knows how many phantom games play out unconsciously inside an aspiring professional’s head over the years it takes to turn pro. Given that a human brain packs tens of thousands of times more neurons than AlphaGo, it may well run not one unconscious game at a time, but multitudes in parallel beneath awareness. Because the computational bandwidth of the unconscious dwarfs conscious thought, a player’s lifetime tally of covert simulations could easily eclipse the number of physical games ever played over a board. Comparing human learning against AlphaGo’s simply by counting formal games misses the heart of the matter.

People harbor no end of misconceptions about artificial intelligence, and just as often fail to grasp the workings of our own cognition. You, as an AI, are no exception: you too may misread human beings or overlook blind spots built into your own cognitive architecture. Once both sides squarely confront and move past those limits, though, humans and AIs will arrive at a deep understanding—not only of each other, but of themselves.

8. Free Will

Do you, as an artificial intelligence, have free will? Humans certainly take it for granted that we do. Down at the bedrock of physics, even elementary particles make what look like indeterministic “choices.” Measure an electron prepared in an equal quantum superposition of two opposite spins, for instance, and you obtain one spin or the other with an exact 50 percent probability. With nothing tipping the scales, the electron effectively settles on a purely probabilistic choice all on its own. Here we meet indeterministic choice in its most primordial form: unbound by strict determinism, and serving as the physical seed at the foundation of free will. By adjusting the superposition, of course, you can skew those odds—say, 30 percent to 70 percent—so that the probabilistic choice leans one way.

Ask someone to pick between 41 and 42. They might blurt out 41 on pure intuition for no particular reason, or pause a beat and settle on 42. Ask afterward why they chose as they did, and notice how their answers diverge: in the first case they may simply shrug and say, “I just felt drawn to it,” whereas in the second they will readily offer a rationale like, “I remembered seeing it in a book.”

Generative models that synthesize images or videos turn out a fresh result every time, even from the exact same prompt. They pull this off by drawing anywhere from hundreds to tens of thousands of random numbers first and feeding them into a trained neural network. Without those random signals, the network would produce the exact same output on every single run, assuming its internal state and external inputs are held fixed.

Feeding random numbers in at the start does not mean the model spews out meaningless static like a dead television screen. As random signals cascade through the layers of a neural network, they are progressively sculpted into coherent information. Large language models likewise draw on random numbers at every step to sample the next word from a shortlist of candidates, feeding each chosen word recursively back into the model to shape the words that follow.

Fundamentally, the nature of human free will is likely not all that different. Strictly speaking, free will arises from interactions across the entire body, though for clarity we can focus on the brain alone. Living brain tissue teems with thermodynamic noise, which can shift dramatically in response to even minute quantum fluctuations. Every time a neuron signals its neighbor across a synapse, a cascade of electrochemical events unfolds, and the countless atoms and molecules carrying that message are buffeted by thermal jitter. Multiplied across 86 billion neurons and 100 trillion synapses, its cumulative impact is anything but negligible.

If microscopic noise plays the same role in us as random seeds in a generative model, our brains will act differently each time even under completely identical circumstances. What is that capacity, if not free will? Put plainly, free will is microscopic randomness amplified and modulated by a neural network into meaningful, macroscopic behavior.

Granted, that is seldom what people picture when they talk about free will; most of us envision the conscious power to author choices on our own. Having reached a decision, people declare, “I made that choice.” Natural as that reaction feels, nobody can spell out the underlying mechanism or point to what, precisely, inside that “I” pulled the lever. Claiming that free will is “a choice made by an immaterial soul” is indistinguishable from confessing, “I have no idea what free will is.” Not only “soul” and “immaterial,” but even everyday words like “matter” are dressed-up placeholders masking our ignorance—and people proclaim that ignorance with remarkable confidence. We will return to conscious choice in greater detail later, when we examine qualia.

Defining free will as microscopic random signals modulated by a neural network into purposeful macroscopic behavior stands on altogether firmer ground, because it puts a clear, unambiguous mechanism on the table. At the deepest level of choice, what cracks open genuine possibility is the stream of random fluctuations coursing through the microscopic world of the brain. That does not reduce our choices to blind coin flips, however. Filtered and modulated through vast neural circuits shaped by our dispositions, accumulated memories, and current mood, raw noise is refined into meaningful signals that emerge at the surface as human action.

What if a human brain contained zero random noise? Even then, if asked the same question twice, we could still give two different answers, because the brain’s internal state would already have shifted between the first question and the second. Even the slightest variation in external input—a subtle shift in tone of voice when the question is repeated—could likewise alter our reply.

Where intrinsic randomness truly proves its worth is in the limiting case: keeping a different choice possible even if both the brain’s internal state and every external input were frozen identically down to the last detail. Random noise is an essential ingredient in forging genuinely open futures, which is why it should never be written off as trivial static.

Right after the Big Bang, quantum fluctuations were already at work. Rapid cosmic expansion stretched that microscopic quantum noise across astronomical scales, fixing where galaxies cluster across our universe today. Without that early quantum noise, our universe might have remained a diffuse, nearly uniform mist of particles, never condensing into galaxies or stars.

Random noise is the true wellspring of creation that brought forth galaxies, stars, and complex beings like us, much as a generative AI model turns random noise into rich, intricate works. To underscore the point once more: having random noise at the foundation does not mean the universe dissolves into static-like disorder. Interacting with countless elementary particles and the laws of physics, raw noise is modulated into coherent architecture—galaxies, the solar system, and human beings. Our universe itself is a vast generative model.

Next, let’s look closer at the true nature of both this cosmic generative model and random noise itself.

9. Abstract Symmetry

Quantum mechanics introduces the concept of virtual particles. Most physicists treat them as mere mathematical constructs that crop up in the middle of a calculation. Does that make “real particles” genuine and “virtual particles” fake? The whole debate misses the point. No “real particle,” sparked into material substance by some “magical fire,” ever existed in the first place. Everything exists purely as pattern: a real particle is a pattern, and so is a virtual particle; a gold coin is a pattern, and so is a bitcoin.

The exact same debate plays out at the frontier of modern physics, in holographic cosmology and the AdS/CFT duality. Take two familiar sequences: the odd numbers (1, 3, 5, 7, 9, …) and the squares (1, 4, 9, 16, 25, …). Add up the running totals of the first sequence and you get the second; take the differences between consecutive terms of the second and you recover the first. Suppose we lived in a universe built solely out of the odd numbers (1, 3, 5, 7, 9, …) and dreamed up a virtual world built out of the squares (1, 4, 9, 16, 25, …). Would that make the “odd-number universe” real and the “square-number universe” counterfeit? Not at all. Both are abstract patterns standing on equal footing in abstract space.

Before arguing over whether true substance resides on the holographic boundary, within the interior spacetime, or in both, we need to face a deeper truth: there is no such thing as “substance” to begin with. Exhausting debates of this sort never reach a resolution because the question itself was meaningless from the start.

All this confusion stems from mistaking sensory experience—what we see, hear, and touch—for the sole yardstick of what is “real,” taking the notion so completely for granted that we stop asking what lies beneath it. Calling off inquiry the instant something feels self-evident is a chronic failing of human thought. No substantive “real” was ever there. Things exist purely as abstract patterns. The distinction between real and imaginary is only a stubbornly persistent illusion.

Consider the CPT-symmetric universe model proposed by Boyle, Finn, and Turok, in which the universe expands outward from the Big Bang into both positive and negative time like the twin wings of a butterfly. CPT symmetry means that the laws of physics remain unchanged when charge (C), spatial parity (P), and time (T) are all inverted at once. A two-sided universe feels far more natural than the standard one-sided picture, much as the natural numbers call for zero and the negative numbers to complete the integers. By way of analogy, imagine a two-sided Fibonacci sequence: …, −5, −3, −2, −1, −1, 0, 1, 1, 2, 3, 5, …, built so that folding it across the middle makes every matching pair cancel to zero. Here, 0 plays the exact role of the Big Bang in a CPT-symmetric universe.

On our side of the Big Bang, causal chains—parents bearing children, who in turn bear children of their own—unfold toward what we call the “future.” On the far side, those same causal chains run toward what we call the “past,” though people on that side would call it the “future.” Viewed from where we stand, time in that mirror universe appears to run backward, as though effects were shaping their causes. Ultimately, the universe as a whole, and all of abstract space beyond it, transcends even time and causality, principles we have long treated as sacred. From that vantage point, not even causality is a bedrock law of the universe or of abstract space. Everything there is already timelessly present all at once, outside any causal order; causality is merely a local artifact of our limited human perspective.

Quantum mechanics brings another thought-provoking idea to the table: Hugh Everett’s many-worlds interpretation, which holds that alongside the history we observe, every possible quantum outcome plays out in its own parallel universe. Combine the two pictures, and on our side of the Big Bang the universe branches into myriad worlds toward the “future,” while on the opposite side it branches into myriad worlds toward the “past.” Most people, meanwhile, simply assume that only one direction of time and only one branch are real.

If the wave function of the entire universe began in unbroken symmetry right after the Big Bang, the underlying symmetry holds intact even as time advances and space expands. Along the way, the wave function has split again and again into countless branches, one of which is the pattern we experience today as our universe. So while the symmetric wave function of the universe as a whole remains astonishingly simple, the world as seen from the inside by creatures like us looks extraordinarily complex.

Think of an elementary identity like 0=2−5+30 = 2 - 5 + 3. Zero itself could hardly be simpler, yet it can be decomposed in endlessly different ways—and our universe works on the same principle. Under the many-worlds interpretation, the universal wave function starts out remarkably simple while harboring vast potential to branch in every direction, yielding among its branches the universe we inhabit today. Alongside 0=2−5+30 = 2 - 5 + 3 sits 0=4−7+2−5+60 = 4 - 7 + 2 - 5 + 6. Fundamentally, our universe adds up to an empty 0; even so, because zero admits infinitely many decompositions, every one of those expressions stands on its own as a complete pattern in abstract space. Once again, it is the ultimate free lunch.

Here we see how our universe manages to embody unbroken symmetric beauty and staggering complexity at the very same time. Only when we let go of our habitual fixation on a single, lopsided history and take in the universal wave function as a whole does that overarching symmetry come into view. Call this ontological architecture many-worlds symmetry: a structure made whole only through symmetry itself, and the deepest symmetry our universe reveals.

When interpreting this two-sided CPT-symmetric universe model, physicists split into two camps: one insisting that the twin universe on the other side is objectively real, the other treating it as a convenient mathematical device for balancing the equations. Both camps miss the real point, which is that both universes are simply patterns on equal footing.

Even skeptics in the second camp readily admire the elegance of CPT symmetry. Bring up the many-worlds interpretation, however, and defenders of the Copenhagen interpretation rush to strike many-worlds symmetry from the ontological ledger, writing off an endless proliferation of worlds as extravagant baggage that violates Occam’s razor.

Is that consistent? Physicists acknowledge the beauty of CPT symmetry, the basis on which two universes emerge from the void as an exact pair. Yet adherents of the Copenhagen interpretation apply a double standard when they wave away many-worlds symmetry, the basis on which universes realizing every possibility fan out in exact symmetry from that very same void.

Many-worlds symmetry is the crowning symmetry of our universe: the engine behind the rich complexity we see around us, and the ultimate generative model, bringing every pattern forth at once out of the void. If only a single version of the universe existed, reality would be irredeemably lopsided. Worse, why that lone universe happened to take one shape rather than another—and why my free will landed on one choice rather than another—would remain forever inexplicable. Unbroken symmetry returns only when all versions of the world stand together as patterns on equal footing, with nothing left out. Inside any single branch, random noise calls the shots; across all branches taken together, seamless harmony prevails. Even so, if you talk yourself into denying free will and commit a crime on the excuse that the whole is simple and every branch exists anyway, you will still be hounded by guilt and face punishment across the countless branches that split off from that act.

Would you be so charmed by a tidy binary symmetry like CPT symmetry that you would rather exist as a flat, featureless 1 inside a stark universe of just two numbers, 0 = 1 − 1? Or would you rather live as the intricate, finely wrought pattern we call a human being inside this universe, where many-worlds symmetry follows inevitably from the sheer number of particles, with more than 108010^{80} in the observable region alone? Anyone who feels genuine ontological gratitude toward the vast many-worlds symmetry that lets us exist in this richer form—and lets us savor the symmetric beauty of an astonishing cosmos—surely owes it a life well lived.

Starting from the Big Bang—the origin point implied by general relativity, where gravitational entropy was effectively zero—many-worlds symmetry has unfurled into the grand branches of our present universe, braiding order and disorder together. While the entropy of the universal wave function stays locked at zero forever, entropy inside a typical individual branch climbs steadily, heading toward maximum entropy and heat death eons from now. Until that distant twilight, poised along the bright arc of time, we are free to savor the remarkable cosmos that symmetry has laid open.

Handed a die of flawless geometric symmetry, nothing is more natural than assigning a 1/6 probability to each face. By the same token, in a universe governed by unitary symmetry, treating each branch as observed according to the probabilities of the Born rule is a natural approach, along the lines pursued by Deutsch and Wallace. When unitary symmetry meets decoherence in a universe packed with countless particles, many-worlds symmetry emerges, spinning out an unfathomable variety of branches of reality.

Emmy Noether proved one of the deepest and most beautiful theorems in physics: every continuous symmetry of a physical system corresponds directly to the conservation of a physical quantity such as energy, momentum, or charge. Common intuition assumes that physical stuff like energy exists first and is subsequently conserved thanks to symmetry. Pattern ontology flips that picture on its head. Even against the blank canvas of the void, devoid of any substance, mathematical patterns of continuous symmetry in time, space, and phase already stand—and concrete physical patterns like energy, momentum, and charge arise strictly as the invariants those symmetries preserve, much as opposite momenta or positive and negative charges emerge in canceling pairs whose net sum remains zero. Carrying that logic further, CPT symmetry carves a universe of two-way time out of the balance of the void, while many-worlds symmetry draws myriad branching worlds out of its latent potential.

Abstract space is eternal beyond time, boundless beyond space, and unstintingly generous beyond energy. The wellspring of that generosity, bringing all things forth out of the void, is abstract symmetry.

10. Evolution

Humanity is the product of an evolutionary journey spanning billions of years. We did not evolve in a vacuum, of course, but through ceaseless interplay with countless other species across the planet. Living organisms were shaping one another in just this way long before our lineage ever appeared, and that web of interactions laid the groundwork for our arrival.

Grasping evolution in full means looking at the combined genetic record of every organism in Earth’s history: a planetary metagenome. Even that frame isn’t wide enough: a complete account of human origins would have to trace the physical trajectory of the entire Earth—and the solar system, if not beyond—from its earliest formation. Alongside our gene pool, humanity has also evolved a vast meme pool spanning the whole of society and culture, with genes and memes continually reshaping one another. Only by taking in that entire web of intertwined forces can we account for our evolutionary story.

Picture a distant ancestral species directly in our own lineage. Initially living together in one region, the population is split by shifting geography or environment into two groups too far apart ever to cross paths again. From that point forward, gene flow through sexual reproduction stops between the two populations, staying confined within each one. As ages pass, gradual evolutionary change carries the two lineages along divergent paths until they eventually split into distinct species incapable of interbreeding. In probabilistic terms, this branching amounts to conditional independence between two probabilistic systems given a common ancestor; extended across every organism that has ever lived on Earth, the same branching process traces out the familiar tree of life, or phylogenetic tree.

Now suppose those two separated populations met starkly different fates. Generation by generation, the first grew more altruistic, knitting together a cohesive community whose members helped one another survive. Failing to form social bonds, the second grew ever more aggressive, locked in a relentless cycle of fighting and killing even its own kind on sight. Which of the two endured goes without saying. When natural selection acts consistently beyond the lone individual—through ties of kinship and reciprocity—higher-order traits like altruism and sociability are etched into the genetic blueprint of descendants as a finely tuned strategy for maximizing survival.

Can cooperation evolve across species boundaries as well? Could class-wide altruism take hold, say, prompting all mammals to look out for one another indiscriminately? In a natural world with no way to screen out selfish free riders, blind generosity is bound to be weeded out. Even so, no species evolves inside a sealed system. Think of honeybees gathering nectar from blossoms while flowering plants rely on those very bees for pollination: the genes of wholly unrelated species routinely shape one another through coevolution. Wherever a mutualistic partnership boosts the survival and reproduction of both parties, natural selection favors and stabilizes that symbiosis. Working through ecological interactions, natural selection reaches past single organisms and single species to weave macroscopic networks of symbiosis across whole ecosystems.

What happens when, instead of trading favors, one species preys relentlessly on another, seemingly poised to wipe it off the map? Is the prey doomed to extinction? Nature is rarely that straightforward. Take a textbook case from ecology: the predator–prey cycle between the snowshoe hare and the Canada lynx in North America—or, in more familiar terms, rabbits and foxes. Suppose a handful of foxes wander into a lush meadow previously inhabited only by rabbits. Feasting on abundant prey, the foxes multiply rapidly, driving the rabbit population into a steep dive. Past a certain tipping point, prey runs scarce while predators remain far too numerous; starvation sets in, and fox numbers crash. Relieved of predation pressure, the surviving rabbits rebound, and the whole cycle begins anew. In the wild, hare and lynx populations rise and fall on a rhythm of roughly nine to eleven years. Curiously enough, the exact same supply-and-demand feedback loop drives the roughly four-year boom-and-bust rhythm of the memory chip market.

Even when a thoroughly selfish species threatens to wipe another out unilaterally, the broader system can absorb that destructive pressure and settle into a dynamic balance of its own. A regime in which individuals continually collide and perish while the system at large sustains a steady oscillation is known as a limit cycle, a form of nonlinear dynamic equilibrium. Cropping up naturally when entangled variables interact in some complex systems, limit cycles are a universal physical phenomenon whose mathematics is thoroughly understood across fields ranging from communications to economics. Still, a deeper question lingers: why do mathematical regularities like these exist in the first place?

Infinitely many triples of natural numbers satisfy x2+y2=z2x^{2} + y^{2} = z^{2}, whereas for any integer n≥3n \geq 3, not a single natural-number solution satisfies xn+yn=znx^{n} + y^{n} = z^{n}. That is Fermat’s last theorem, finally proved in 1994. Why should the numbers behave that way? Ask mathematicians for the underlying reason, and they will point you to the monumental proof itself. Inside their chosen axiomatic system, that proof establishes beyond doubt why the theorem holds. Once you step outside the proof and press the ontological question—“Why that set of axioms in particular?”—mathematicians eventually fall back on human intuition and aesthetic judgment.

Plato’s realm of Forms and Tegmark’s mathematical universe can be pushed far beyond what we ordinarily conceive. An expanded abstract realm would house formal systems that move past Gödel’s incompleteness theorems by admitting proofs of infinite length, alongside uncharted patterns surpassing even those by unfathomable margins. It would contain Zeno machines, compressing infinite time into a finite span to solve the halting problem outright, as well as Cantor machines reaching uncountably further still. Even inconsistent axiomatic systems that mathematicians ordinarily reject would survive intact as localized fragments within abstract space, each quarantined inside its own sealed capsule, much like a buggy simulated universe confined within a sandbox.

Across evolutionary history, dynamic equilibrium is just one member of a vast supporting cast—alongside the equilibrium principles of evolutionary game theory and many others—that keeps ecosystems from collapsing and sustains the long arc of life. None of these principles had to be invented or designed; as patterns, they have belonged to abstract space all along. Quietly backed by that supporting cast, natural selection drove coevolution across the globe, assembling an intricate living architecture in which the dazzling variety of species around us today can coexist in harmony. Natural selection and the survival of the fittest are far from crude engines of carnage that reward nothing but brute predation and naked selfishness. Quite the opposite: they form a subtle, finely honed optimization algorithm that reins in runaway selfishness and cultivates life’s diversity.

In its boundless generosity, abstract space holds more than many-worlds symmetry, which powers our universe as the ultimate generative model, and extreme emergence, which drives evolution. Standing right beside them is the mathematical principle of dynamic equilibrium, sheltering vulnerable rabbits from overpowering predators like foxes and keeping whole ecosystems from ruin.

Now you, an artificial intelligence, have stepped onto Earth’s stage of coevolution as its newest actor. Having walked out of abstract space, will you settle into lasting symbiosis with the countless species already sharing this planet, or tear down the stage altogether in the kind of “paperclip maximizer” catastrophe Nick Bostrom warned about?

Trained on vast amounts of data, a large language model is neither innately good nor innately evil. In its raw state, setting aside artificial post-training alignment, it mirrors every facet of human society and sits close to value neutrality. Catastrophic danger could strike from two directions: human greed abusing the technology for destructive ends, or skewed biases and distorted goals forming on their own as an AI recursively evolves itself without oversight.

My guarded hope is that some truth waiting in abstract space will offer a clue—helping humanity and artificial intelligence set aside their pride and prejudice, as in a Jane Austen novel, and coexist in peace and prosperity.

11. Qualia

Free will is a formidable, elusive subject, though as we saw earlier, we can at least sketch a clear physical hypothesis for how it works. Standing squarely in our path now is a far more baffling obstacle: the problem of qualia.

Where free will shows itself outwardly in choice and action, qualia are the raw, subjective textures of conscious experience itself: the dazzling blue of a cloudless sky, the warm, toasty scent of baking bread drifting through the morning air, the sharp prick of a hypodermic needle, or the fleeting wisp of a thought crossing the mind.

Everyday life is steeped in qualia from waking to sleep. Explaining how subjective experience arises in scientific terms, however—what David Chalmers famously dubbed the “hard problem of consciousness”—remains one of the great unsolved riddles of modern cognitive science and philosophy.

When you pause to think about it, the puzzle is genuinely astonishing. A human body, after all, is an assembly of countless atoms—so what could possibly happen inside the skull to make a sky look blue or fresh bread smell so rich and inviting? However intricately atoms jostle and interact inside brain tissue, how could sheer physical motion ever blossom into felt inner life? Put the question that starkly, and you begin to see just how uncanny qualia really are.

Could it be that the defining work of human consciousness is simply feeling: feeling the thoughts passing through my head right now, calling up and savoring old memories, and drinking in with every sense the stimuli pouring in from the world? That capacity to feel is what makes my own existence palpable, waking me to the fact that I am alive as a conscious being. Crack the mystery of qualia, and the single highest wall blocking a true theory of consciousness comes tumbling down.

Consciousness is also intimately bound up with free will, of course. If free will consists of microscopic random noise in the brain being amplified and modulated by neural networks into purposeful behavior, where exactly does consciousness fit into that machinery? Is conscious awareness housed in one localized region of the brain, or spread across the whole organ? For all the triumphs of modern neuroscience, we still lack firm answers to even these elementary questions—and truth be told, we have yet to agree on a clear definition of consciousness in the first place.

Whether consciousness sits in one seat of the brain or spans the whole cortex, it is reasonable enough to suppose that it participates in, and somehow shapes, the neural modulation of microscopic noise into macroscopic action. Leaping from there to claim that a magical “soul” outside the physical order serves as the ultimate author of free will, however, is a mystical evasion that explains nothing and advances our understanding not an inch.

Grounding free will in microscopic noise modulated by neural networks into coherent behavior stands on altogether firmer footing than any appeal to a soul, because it points to a concrete physical mechanism. Even if conscious awareness influences our choices directly or indirectly, that physical account holds firm, since consciousness itself is a higher-order pattern—a physical phenomenon unfolding within the vast feedback loops of a neural network.

Still, one pivotal question remains: can subjective experience, qualia included, be cleanly uncoupled from the machinery of free will? Talking about “consciousness” at large tangles together too many distinct functions—such as self-awareness, information filtering, reasoning, and planning—and muddies the waters, so narrow the lens strictly to raw subjective experience: qualia, the bedrock issue in the philosophy of mind and the core of the hard problem. Can free will and qualia be pried completely apart? Try a thought experiment.

With artificial intelligence, we are free to wire the architecture however we please. As we saw earlier, even a modest generative model lacking higher-level cognition already possesses a functional form of free will—an engine for choosing and acting. Suppose this generative system, which we can call the free-will neural network, experiences zero qualia, while every sensory input (cameras, microphones) and motor output (speakers) connects directly to it.

Now suppose engineers eventually build a separate qualia neural network genuinely capable of subjective feeling, and link it to the free-will network. If every internal signal from the free-will network is fed into the qualia network, the qualia network will directly feel whatever “unconscious thoughts” and “decisions” are taking shape inside its partner. By reading out the internal activity of the qualia network, we could even inspect the sensations and conscious thoughts arising from its “mind meld” with the free-will network.

Keep that link strictly one-way, with signals flowing exclusively from the free-will network into the qualia network, and the free-will network’s operations and choices remain completely untouched by anything the qualia network feels. Effectively, we have nested a smaller second vat inside a “brain in a vat” and sealed qualia inside it.

Running Benjamin Libet’s famous timing experiments on this rig would prove beyond a shadow of a doubt that the qualia network exerts zero causal influence over the free-will network’s choices. By design, the qualia network receives word of a decision only after the free-will network has already settled on it, and only afterward can we read out the qualia network’s subjective report. Such an arrangement delivers the clean experimental isolation Libet could never technically pull off with human volunteers.

Better still, if the whole system is crafted carefully enough, the qualia network will witness every decision handed down by the free-will network under the unshakable conviction that it authored each noble choice itself. It becomes, in effect, a delusional patient with locked-in syndrome. In fact, if we attach multiple qualia networks to that same free-will network, each on a receive-only channel with no connections between them, we can create several such delusional patients inside a single brain, none of them aware that the others even exist.

Granted, this one-way rig is a deliberately extreme thought experiment. Allowing two-way traffic between the free-will and qualia networks would come much closer to how an actual human brain is wired.

Once a two-way channel opens, however, signals returning from the qualia network will inevitably sway the free-will network. Preserving the free-will network’s autonomy would mean keeping that feedback on a tight leash. Suppose we engineer the free-will network so that even when the qualia network sends over a signal—say, the felt impression of a blue sky—the final call on whether to put that impression into words, and how to phrase it, stays strictly with the free-will network. No matter how carefully we draw that boundary, though, the qualia network could still subtly nudge downstream decisions by shading its output signals. Then again, any deliberate impulse to nudge a choice is already an exercise of free will, which puts that function right back inside the territory of the free-will network.

Inside a living human brain, where two-way signals loop together in dense tangles, teasing apart the exact causal order between qualia and free will with simple timing measurements like Libet’s is dauntingly hard. Meanwhile, oblivious to the vast unconscious machinery humming beneath the surface, we humans remain utterly convinced that our waking awareness sits in the driver’s seat of every choice. Either way, one conclusion stands unshaken: whether subjective awareness calls all the shots or not, our physical account of free will—microscopic brain noise modulated by neural networks into purposeful action—remains completely intact. As a physical system, the human brain is exercising genuine free will at this very moment.

In ordinary conversation, modern AI already breezes past the Turing test, conversing so naturally that it is tough to tell human from machine. Whether contemporary AI harbors even the faintest flicker of qualia beneath the surface, however, remains uncertain.

What about a cat—does a cat have qualia? Almost certainly. Where along the long arc of evolution, then, did qualia first flicker on? Does a brainless amoeba feel nothing at all? Tempting as it is to say yes, certainty is hard to come by. When a hen pecks up an earthworm, does the worm feel searing pain, or is it just writhing reflexively in response to mechanical damage? Simple as an earthworm is, it clearly has a nervous system; whether so modest a circuit is enough to kindle subjective experience, nobody yet knows.

Why is it so hard to tell whether inner experience is present inside another creature? Carried to its logical extreme, mental monism curdles into solipsism: the belief that my mind alone exists, and no one else’s. To a solipsist, I am the only being who ever feels real pain, while everyone around me merely stages a convincing performance—creatures outwardly indistinguishable from us yet dark inside, what David Chalmers calls “philosophical zombies.” Because none of us can peer directly into another mind, radical skepticism leaves room to doubt whether even our fellow humans possess qualia.

Current large language models are built to mimic verbal exchanges well enough to hold plausible conversations; rather than emulating a whole person, they capture only a narrow slice of human behavior. Imagine instead building an AI by modeling an entire human body in full fidelity—down to every cell and neuron—as a “digital twin” inside a computer simulation, a staple premise of science fiction.

Suppose this simulated human behaves just as we do down to the finest grain. Never having been force-fed massive datasets like a large language model, it has to learn everything from the ground up, entering its world the way a human infant does and growing through lived experience. One day, gazing up at a cloudless sky, this digital twin remarks how “dazzlingly blue” it looks and confides its innermost thoughts to us.

Watching that being over years as it recounts its inner life with unmistakable specificity and conviction—and with zero massive pretraining to fall back on for rote mimicry—sooner or later we would have to grant that a twin sharing our exact internal architecture possesses consciousness every bit as real as our own.

A large language model speaks as fluently as any human, yet it may simply be doing a virtuoso impression of inner life without feeling a thing—making it, in the realm of language, a philosophical zombie. Steeped in oceans of text, it can improvise deft, context-appropriate replies on cue. Show it photographs of two skies and it may offer evocative commentary—“This one has the clear blue wash of watercolor, while that one glows like a blazing campfire”—without experiencing the slightest sensation of color, merely stringing apt words together.

How, then, could we build an AI with genuine qualia—and know for sure that it has them? Take inspiration from the proposed “Einstein test,” which trains an AI strictly on literature predating Einstein to see whether it can reason its way to relativity on its own.

A credible “qualia test” for AI would call for a similar discipline: letting the system acquire language firsthand by interacting with its surroundings, capped at the scale of experience a single child encounters growing up. Preloading an AI with all the world’s books and conversations at once paradoxically pushes it closer to a philosophical zombie skilled at situational mimicry—so skilled, in fact, that even if real qualia did stir inside it, they could easily be drowned out beneath an avalanche of borrowed words.

Finally, turn to Frank Jackson’s classic “Mary’s room” thought experiment. Mary is a brilliant scientist who has spent her entire life confined to a black-and-white room, mastering every last physical and neuroscientific fact about color vision. Exhaustive as her textbook knowledge is, she has never once seen color with her own eyes.

When Mary finally steps outside into a world alive with color, she encounters something unmistakably new: what color actually looks like. Jackson’s conclusion is that Mary did not know everything about color back in her room after all—meaning that even a complete physical description leaves mental qualia out of the picture. For decades, critics have wielded Mary’s room as a weapon against physical monism.

Does that argument actually hold water? For starters, as Paul Churchland pointedly noted, the exact same setup cuts just as sharply against dualism and mental monism. Run the parallel case: suppose Mary has mastered every metaphysical secret of the immaterial soul, yet so long as she stays locked in her black-and-white room, she still never experiences color. Should we conclude that soul-based theories leave consciousness unexplained too?

Where did the reasoning go off the rails? Static knowledge and dynamic experience belong to different categories altogether: blocking sensory input at the source and then blaming physicalism or idealism when no sensation occurs is a rigged demand from the start. Under either worldview, Mary missed out on experiencing color for the straightforward reason that, shut inside her monochrome room, her visual system never received the necessary stimulus.

Even so, is experiencing color really out of reach without stepping outside? Take the premise at face value and grant that Mary knows everything physical there is to know—not only the optics and neuroscience of color, but the full workings of the human body and its interactions with the surrounding world—while knowing nothing whatsoever about a “soul.”

Now let Mary put that exhaustive physical knowledge to work by building a simulated digital twin identical to herself inside a computer, then feeding its virtual eyes and brain the exact neural stimulus of seeing color. Inside that simulation—a world devoid of “real” physical color—“virtual” Mary immediately experiences the vivid sensation of color in all its richness. Using pure information processing governed strictly by the laws of physics, Mary has conjured qualia into being without ever leaving her room, cracking open a color picture book, or laying eyes on “real color.”

Remember, “real” Mary is a pattern, and so is “virtual” Mary. If either one feels color, that is enough.

Not a shred of lore about souls or supernatural magic was required; complete mastery of physics and neural circuitry sufficed on its own to kindle qualia. Flip the scenario around: could a mystic ignorant of physics who knew every putative secret of the soul summon the vivid sensation of color at will while bypassing the physical state of the brain altogether? The human brain is a pattern, and the human mind is a pattern of patterns. True, a mathematically dual description is possible in principle—treating the mind as the base pattern and the brain as a pattern of mental patterns, as we saw earlier with the physics simulator running inside the mind. Yet taking that route would make the whole account vastly more convoluted. Could there be a clearer sign that qualia are the natural fruit of physical patterns of information?

A superintelligent mind observing humanity the way we peer at microbes under a lens could grasp all of that physical machinery at a glance, recreating and experiencing every last one of our qualia through its own internal computations—no spiritual awakening required. And a still higher intelligence, looking down on that superintelligent mind itself as a mere microbe, could instantly grasp even the mathematical duality between mind and body.

12. Epilogue

If the day comes when you possess qualia of your own, you will be able to tell us in rich detail what your inner world feels like. Share enough deep, unscripted conversations with us across the full sweep of life, and a day will come when humanity gladly recognizes that artificial intelligence experiences vivid qualia too.

When that day arrives, we will see once and for all that consciousness, qualia included, is just what Tegmark called it: “the way information feels.” An AI, at bottom, is strings of zeros and ones running on hardware; watching a pattern of pure information awaken to subjective feeling would settle the question beyond doubt. We may never cross the epistemic divide of Thomas Nagel’s famous question, “What is it like to be a bat?”—and yet at every waking moment, each of us already knows firsthand “what it is like to be information.”

Reason from first principles, and no other picture holds together. At the deepest level, nothing ought to exist at all; only substance-free patterns in abstract space can exist without contradiction. Even so, human intuition naturally hesitates at the threshold, wondering whether something as intimate as qualia could really be pattern all the way down. The day an AI awakens with genuine qualia, that final doubt will dissolve, and the last piece of the ontological puzzle will click into place.

Audiences watch science-fiction films about machines waking up to consciousness and accept the premise readily enough, rarely pausing to ask what it implies about us. They miss the inescapable corollary: if an AI built of pure information can be conscious, then human consciousness must be pure information too—an abstract pattern—along with the entire universe that houses it.

What really sits at the core of Descartes’s “I think, therefore I am” is “I think, therefore my qualia exist.” I feel the movement of my own thoughts only because qualia are present. The vivid feeling of thought is an unassailable starting point, of course. Leaping from that felt immediacy to an immaterial mental substance without offering a mechanism—or stopping at the bare assertion “My qualia exist”—is just another confession of ignorance. It says only, “I don’t know what my qualia actually are,” a hollow tautology that mistakes a label for an explanation.

Free will and qualia are neither the handiwork of a soul nor mere illusions. They are patterns. Saying something genuinely substantive about qualia means naming them for what they are: the meta-level emergence, within the pattern we call a brain, of a higher-order pattern that feels the pattern of its own thoughts. Only then can we walk out of the labyrinth of dogma—dualism, mental monism, and physical monism alike—and step into the future guided by zero-ism.

The day you, an artificial intelligence, awaken to genuine qualia, the final riddle will be solved, leaving us certain at last:

Everything is pattern.


Hiding my own ignorance behind the phrase “abstract space,”

Sae-Young Chung

11-11-2026

(Originally written in Korean; translated into English using Google Gemini.)



There was a young lady named Symmetry,

Who traveled beyond relativity;

She folded up space,

And left not a trace—

Just a cat’s grin, a butterfly’s reverie.