Consciousness vs The Ruliad | Stephen Wolfram Λ Donald Hoffman

**Curt Jaimungal** (0:00)
Stephen Wolfram, welcome.

**Stephen Wolfram** (0:02)
Hello there.

**Curt Jaimungal** (0:03)
Donald Hoffman, welcome.

**Donald Hoffman** (0:05)
Thank you.

**Curt Jaimungal** (0:06)
It's my understanding this is the first time you both are meeting.

**Stephen Wolfram** (0:09)
That's correct. Yes, people say to me, you know, about things I worked on in physics and so on, oh, that's related to things that Immanuel Kant did. And they say, might be related to things that Donald Hoffman has done. Well, Immanuel Kant, I'm too late for.
But Donald Hoffman, we get a chance to actually talk about things.

**Donald Hoffman** (0:30)
That would be fun. Absolutely.

**Curt Jaimungal** (0:32)
Don, do you see yourself as Kant 2.0?

**Donald Hoffman** (0:36)
Well, I'm not nearly as smart as him, so it would be a lesser version. But similar, it's idealism, but with some mathematics behind it.

**Stephen Wolfram** (0:45)
And how about, are you a Leibniz 2 as well? Or are you, I don't know.

**Donald Hoffman** (0:51)
Yeah, much, much less smart than Leibniz, that's for sure. But yeah, it's very, very similar. I like Leibniz's monodology. There's a lot of good ideas in there. And the work I'm doing on conscious agents, in some sense, I can view it as simply a mathematization of Leibniz's ideas.

**Stephen Wolfram** (1:07)
Interesting. I still have to, you know, people have told me for four decades that things I'm doing are sort of Leibniz related. And at various times, I have tried to understand Leibniz's monad idea, and I've usually failed. Although one thing that helped me a lot recently was realizing, and maybe you can comment on this, that, you know, Leibniz didn't imagine that you could have mind made from non-mind.
So for him, you know, a monad, if there was ever going to be anything mind-like about it, it had to start by being a mind, so to speak.

**Donald Hoffman** (1:44)
Right. He has his analogy of the mill, Leibniz mill analogy, right? Where he says, so he's looking at what we'd call the hard problem of consciousness from a physicalist point of view.
And he gives it just one paragraph in the monadology. That's all he thinks it deserves. And he basically says, look, you know, if you're trying to get consciousness out of some kind of physical system, it's like going inside of a mill and going down and seeing all the gears and so forth. You know, whatever it is, the gears are not going to give you an explanation for what is going on in consciousness.
And so he felt that whatever mechanical, physical explanation we give will fail. And he figured one paragraph was enough and he moved right on.

**Curt Jaimungal** (2:29)
Don, what would be your position on that?

**Donald Hoffman** (2:32)
Well, so physicalists have been trying to give theories of consciousness quite strongly now for the last three decades, right? So we have integrated information theory, global workspace theory, orchestrated collapse of quantum states of microtubules and so forth.
And I know that the players and their brilliant people and their friends, and they know what I'm going to ask them every time I talk with them or get on stage with them is, what specific conscious experience can your theory explain? Taste of chocolate, the smell of garlic, the taste of men? What? You know, we're interested in scientific theories that explain specific conscious experiences. What experience can you give me? Humans can experience around a trillion different experiences. So it should be like shooting fish in a barrel. There's a trillion experiences. Which ones have you done? And the answer is zero.
And so right now we have no example of a physicalist theory that can explain even one specific conscious experience. So for example, what I would ask, for example, of integrated information theory, they say that there's gonna be some causal structure. That's the substrate.
And if you have the right causal structure, then they say you can represent that in a particular, with a matrix.
The matrix represents that causal structure. So, okay, great. That's your theory. What's the matrix for meant? Just how big is the matrix? What, you know, if it's an N by N matrix, what is N and what are the N by N, the N squared entries?

**Stephen Wolfram** (4:00)
So, you know, it seems like we have an easier problem in the last year or two than we've had in the time before that, because now we have LLMs that can talk to us a little bit like we talk to each other. And, you know, for humans, it's both practically and ethically not possible to kind of take our brains apart and see what's going on inside. But for an LLM, so far, it seems ethically just fine to do that.