To those who say that LLMs absolutely cannot be conscious because there is nothing in there that can be identified as the “conscious” component, you’d be awfully surprised when you slice open a human and the readily-identifiable “consciousness area” is also nowhere to be found¹.

What a bummer, right? That we just can’t have absolute certitude here. And so many conclusions that so many people are certain about just have no basis in anything at all — not neuroscience, not computer science, not pure reason, and certainly not philosophy. It’s all just someone’s naïve and nearly-always worthless feeling about something.

Embodiment matters. So do other things that shall go unnamed here as it’d be just too long to have any sort of proper discussion.

But many properties in this world of ours are emergent. They do not appear “in” anything. They are the thing itself. And this is what all these not-even philosophers miss: The outward properties — the perception — is often all we have. It’s all we can really know. I’m not arguing that we should treat Claude as conscious just because it says it is. But recall how I said (correctly) that many features of the world are emergent? That means that it could be conscious in some important way, even if fleetingly, even if much different from how humans experience their own self-awareness, and we’d not know it because we do not have any way at all to evaluate that.

In other words, we — humans, AI, your cat — might be illusions that become real by us believing in one another. That belief reified is what we call “consciousness.”

1. And even if it could be found, and even if you say, “It’s the whole brain!” what would that prove? Probably far less than you think.