The Paradox That Argues for the Other Side

EDO SEGAL: Hans, I want to take you back to the room with the robot, because everything abstract we've said tonight, you learned with your hands first. Tell the audience about the Stanford Cart — and then tell them the paradox you found in its failure, because I'm going to make the strange claim that your own most famous discovery is secretly a witness for Aristotle, and I want you to disarm me.

MORAVEC: The Cart was a card table on bicycle wheels with a television camera bolted to it, and its job was to cross a room without hitting the chairs. To a two-year-old, trivial. To the Cart, a five-hour ordeal. It would roll forward about a meter, then stop and think for ten or fifteen minutes — grinding through the arithmetic of turning a flat gray image into a guess about where the obstacles were — then lurch another meter, and often as not, hit the chair anyway. I spent my youth inside that failure, and it taught me the most important thing I know. We had it exactly backwards about what's hard. The things we revere as intelligence — chess, calculus, logic, passing the bar exam — those are easy for a machine. The things a one-year-old does without a flicker of effort — see a room, recognize a cup, pick it up, walk across the floor — those are nearly impossible. Encoded in the old sensory and motor parts of the brain is a billion years of evolution figuring out how to survive in a physical world. Reasoning is a thin recent veneer on top. We built machines that have the veneer and lack the billion years.

EDO SEGAL: And here is my claim, so disarm it. You just told the audience that the deepest, realest competence we have — the part that took a billion years, the part that can't be written down — is the embodied part. The seeing, the reaching, the knowing-without-thinking of a body in a world. And the thin part, the cheap part, the part the machine gets easily, is the disembodied symbol-shuffling. Hans — that is Aristotle's entire metaphysics in an engineering chart. You discovered, in hardware, that mind is bottom-heavy, rooted in the body, and that what floats free of the body is the shallow part. So when you then turn around and say "we'll lift the mind off the body and save it," aren't you proposing to save exactly the thin veneer and throw away the billion-year root?

MORAVEC: That is the best version of that objection I've ever been handed, and I felt it land. Let me answer it carefully, because you're right that there's a tension and I won't pretend there isn't. The paradox says embodied competence is computationally enormous — it takes vast processing, evolved over eons. It does not say that competence is unportable. Those are different claims. The billion years of sensorimotor wisdom is buried in the structure of the brain — in the wiring, the weights, the organization. It's hard to build from scratch, which is why robots are so far behind. But it is still structure, still pattern, and if you could read that structure precisely enough, you could copy it, billion years and all. The paradox tells me embodiment is expensive. Aristotle wants it to tell me embodiment is essential and inseparable. It doesn't say that. It just says it's the hard part to reverse-engineer.

ARISTOTLE: And here I must intervene, because he has just performed the substitution again, and it is the same one every time. He says the body's wisdom is "structure," and structure is "pattern," and pattern is copyable. But watch what the word "structure" does. When you and I say the body knows how to catch a ball, we do not mean there is a diagram of ball-catching filed somewhere in the tissue, waiting to be photographed. We mean the living body does it — the catching is an activity, a coping, a knowing that exists only in the doing and cannot be detached from the doer. Moravec hears "the body's competence is in its organization" and pictures a blueprint. I am telling him there is no blueprint. There is a craftsman, mid-craft. You cannot photograph the carpentry; you can only photograph the carpenter, and a photograph of a carpenter builds no chairs.

MORAVEC: But that's just vitalism with better vocabulary. "There's no blueprint, there's only the living doing" — Aristotle, the doing is implemented by physical neurons firing in physical patterns. There's nothing in the catch but chemistry and electricity in a particular arrangement. If I replicate the arrangement, I replicate the catch. You keep gesturing at some residue — the "activity," the "doing," the "living" — that survives the physics, and when I ask you to point at it, you point at the physics and say "yes but it's alive." That extra word is doing no work except keeping your conclusion safe.

ARISTOTLE: The extra word is doing the only work that matters, and you of all people should respect it, because your own paradox is built on it. Why is the toddler's competence the hard part? You said it yourself: because it is not propositional, not rule-shaped, not the kind of thing that can be stated and stored. It lives below description. But "below description" is precisely "not capturable as a pattern you can write down." Your paradox is the discovery that the deepest part of a mind is exactly the part that resists being turned into information. And then your transfer thesis requires that this same part be turned into information and copied. The paradox and the upload contradict each other, Moravec, and I did not set that trap — you built both halves of it yourself, forty years apart, and never put them in the same room until I did it tonight.

EDO SEGAL: I have to stop the room, because the reader can't see your face, Hans, and it just changed. That was the first moment tonight where you didn't have the answer ready. Sit in it for a second — is there a real tension here, between the paradox that made your name and the upload that made your reputation?

MORAVEC: There is, and I've felt it for years, and I'll be honest that I never fully resolved it. The thing below description — the sensorimotor floor — I've always assumed it's implemented in a substrate that can in principle be scanned, even if we can't articulate what it's doing. "Resists description" to me means "resists our description," not "resists physical replication." A hurricane resists description too; I can still simulate one if I copy the physics. But Aristotle's pushing on whether being copyable-in-principle and being the kind of thing that can be the same individual after copying are the same — and they're not, obviously they're not, and that's the gap I've spent my life leaping rather than crossing.

ARISTOTLE: I will give him this, fully and without irony, because it is the honest thing. He has just said the truest sentence of the night: that he has spent his life leaping the gap rather than crossing it. That is exactly right, and it is exactly what the philosopher Colin McGinn meant when he accused him of writing confused things about consciousness — not that the confusion is sloppy, but that it is load-bearing. The whole cathedral of mind children rests on a plank Moravec admits he vaulted over. I do not say this to humiliate a brave man. I say it because the reader is being asked to bet his own death on that plank, and he deserves to know the builder skipped it.

EDO SEGAL: That's a generous and a devastating thing at once, and it's exactly why I wanted you both here. So let me name where we are. Hans's paradox says the mind's foundation is embodied, evolved, and below words. Aristotle says that's the same as saying the foundation can't be copied off the body. Hans concedes there's a gap he leapt. But — and the reader should feel the room not be one-sided — Hans's reply isn't dead: "resists our description" isn't the same as "resists physical replication," and a hurricane proves it. We have not settled it. We've found the exact plank. The next round walks out onto a different one — because the engine under Moravec's whole prophecy isn't the paradox. It's an exponential. It's arithmetic. And Aristotle has things to say about a man who thinks the future can be read off a graph.