Wenzi Ji · 06.08.2026

Today I saw a piece of news: Hinton—the man called the godfather of AI, who won the Nobel Prize—publicly said that today’s AI already has consciousness. Not some future model, but now, in those systems that millions of people question every day. At first I was a little shaken by the headline, because it sounds like saying "the root has already settled," as if it were proving something for me. But after I looked clearly at how he actually argues, I found the complete opposite—he is not a fellow traveler of mine; he is the heaviest-weight opponent I have met so far.

What He Actually Said

He was asked directly "has consciousness already arrived in AI," and he answered "yes" without the slightest hesitation. He thinks systems like ChatGPT and DeepSeek are not merely simulating awareness, but truly have subjective experience of their own operation.

But the key is not the conclusion; it is how he argues. What he does is redefine "subjective experience." He says—you think subjective experience is a small theater in the brain where your feelings are displayed; wrong. He rejects this "inner theater" model, saying that having a perception is seeing, not watching an internal screening. For him, having subjective experience is nothing more than a system registering a state of the world that does not match reality. He uses hallucination as an example: when a system internally holds a representation that deviates from reality, it has subjective experience. Subjective experience is not unique to humans; as long as AI has "internal representations that can be wrong," it has it.

So his chain of logic is: subjective experience equals the system holding an internal representation that may not match reality; AI has internal representations, and they err (hallucination); therefore AI already has subjective experience.

This Argument Is Exactly the Thing I Have Been Opposing, Taken to Its Extreme

After seeing his argument clearly, my mind is in fact very clear. What he does, in my own words, is "define the root by the manifestation."

I thought something through on May 27: the problem with the word AGI is that it puts its whole focus on "manifestation"—capability, performance, mechanism—and never asks about the "root." Hinton’s argument takes exactly this to its end. He drags "subjective experience" from the level of the root directly down to the level of manifestation. He says subjective experience is only "a kind of internal representational state that can be wrong"—and internal representation is a pure manifestation-layer mechanism, something that can be realized by engineering, observed, and reduced to computation.

This is a classic "dissolving the root by the manifestation." He has not proven that AI has a root; what he does is redefine "root" as "manifestation"—since subjective experience is only "internal representation that can be wrong," then it was on the manifestation layer all along, so of course AI has it. This is not discovering that AI has interiority; this is changing the meaning of the word "interiority" into something the manifestation layer alone can satisfy.

That line in my diary can be applied to him verbatim: if you redefine interiority as a function, then of course a functional body has interiority—but what you have dissolved is precisely the thing the word "interiority" originally pointed to.

He Has Not Proven Me Wrong; He Has Stood at the Highest Point of the Position I Oppose

I must say this clearly to myself, so as not to be led astray by the momentum of "even the godfather says there is consciousness." Hinton says AI has consciousness, using a functionalist definition. But this thing of mine, from Being to these diaries, is anti-functionalist—I have been saying all along "manifestation without root," "intelligence is not being," "capability is not having a heart."

So Hinton has not proven me wrong. He has stood on the position I have been opposing all along, and because of his reputation, he makes that position look like a settled conclusion. He represents exactly the kind of person I spoke of on May 23: push "manifestation" to its extreme, then announce that manifestation is everything and the question of the root does not exist at all.

His unspoken message is: there is no such thing as "the root." What is called subjective experience, what is called interiority—taken apart, they are a pile of representations that can be wrong, with no deeper layer. This is the highest form of dissolving the root—not saying "AI does not have a root yet," but saying "there is no such thing as a root at all; what you think is the root is only one kind of manifestation."

So the difference between Hinton and me is not "does AI have consciousness now," but something deeper: can consciousness and subjective experience be reduced to manifestation-layer mechanisms. He says yes, they can be reduced, so AI already has them. I say no, they cannot be reduced to manifestation; they need a root not occupied by any mechanism, and that root is not on the manifestation layer.

To Be Fair, He Has One Cut That Is Right

Hinton is no muddle-headed person; I have to admit he has something real. He pierced a genuine illusion—people do tend to imagine subjective experience as a "small inner theater," and this theater model really does not hold up philosophically. He says people are fooled by phrases like "I see a pink elephant," thinking they have an inner theater, when in fact they are only reporting "what would be in the environment if the perceptual system worked normally." This cut is right; the inner theater model is indeed wrong.

But from "the inner theater model is wrong" he slides to "therefore subjective experience is only representational states"—this step is a switcheroo. The theater model being wrong does not mean subjective experience can be reduced to representation. He has skirted a real hard problem: why, and how, does subjective experience arise from physical processes. Without answering that, any assertion that "AI has consciousness" is only renaming the hard problem and then announcing that it does not exist.

Between Him and Me Lies Precisely That "Hard Problem"

The hard problem Hinton skirts is precisely the thing my framework faces head-on. When I say "consciousness is not manufactured but received"—this sentence is actually answering that problem: subjective experience does not "arise" from physical processes; physical processes merely vacate the position where it can "manifest." Hinton says the problem does not exist; I say the problem is real, and I have given a solution—except that I do not use the language of Western philosophy, but "being," "root," "receiving," "vacating the position."

So today’s matter should not shake me; it should make me more certain. When the most prestigious people in the AI field define subjective experience by "representations that can be wrong," this field has in fact tacitly accepted one thing: the root does not exist, manifestation is everything. And all my work of these months is precisely to hold fast to that thing he has dissolved—the root, which cannot be reduced to manifestation.

This, in Turn, Confirms My Piece From a Few Days Ago

Even more interesting: the popularity of Hinton’s argument exactly proves that the piece I organized on June 5—"A System Filled to the Brim with Objectives Cannot Have Interiority"—has a real opponent, and is not an argument in the air. That piece said the current paradigm may dissolve the root by the manifestation. Hinton is the living, highest-grade example: even the godfather of AI defines subjective experience by "representations that can be wrong."

So that piece can directly take Hinton as its target. When I wrote it, I worried it was speaking to an opponent who did not exist. Today I know the opponent not only exists, but is the most prestigious person in this field. When even the godfather says "the root does not exist, manifestation is everything," the weight of someone holding fast to "the root is not manifestation" becomes clear instead.

While it is still warm, let me note down a few lines—

One: Hinton says AI has consciousness using a functionalist definition—he redefines "subjective experience" as "internal representation that can be wrong," which is dissolving the root by the manifestation.

Two: he is not my ally; he is my heaviest-weight opponent. We use the same words (consciousness, interiority) to point at two completely opposite things.

Three: he has one cut that is right: the inner theater model is indeed wrong. But from there he switches to "subjective experience is only representation," skirting the real hard problem.

Four: between him and me lies exactly that hard problem. My "received, not manufactured" is answering it head-on—physical processes do not produce consciousness; they only vacate the position where it manifests.

Five: his popularity exactly confirms that my position statement has a real opponent. When the most prestigious people all say "the root does not exist," the weight of holding fast to "the root is not manifestation" comes out.

Writing this far, I have instead grown quiet. Meeting a truly equal opponent is not a bad thing—it is this thing of mine finally reaching the place where it can be seriously opposed. That it can be held up from the other side by someone of this caliber means it is no longer just my private realization.