Donald Hoffman

the agents that are available in it to, to create that.

So there'll be, maybe, maybe I'll have like a trillion agents interacting, but I arrange for dynamics of say a billion of them to be creating a particular space time headset.

And then what I would do – so I know because Markovian kernels are computationally universal, I can do that.

I can use the whole language of conscious agent dynamics to create a projection into like a space-time headset.

I know that they're computationally universal, so I can do that.

Now the question is, once I've got that dynamical system and I have –

some of the agents creating this headset.

What happens when I turn that headset and have it look at

the whole agent system, and in particular, the part of the system that creates that headset.

So I'm taking a bunch of agents and their dynamics is creating a space-time interface, but now I use the interface to look back at the only thing that's available, which are agents.

That's all that's available to look at.

But I want to look at the agents that were particularly interested, that were involved in the creation of the interface.

What will that set of agents look like?

They'll look like neurons and brains.

Well, for us.

In other words, I'm saying- I thought they didn't map like that.

Remember, neurons are just artifacts of the interface.

They're symbols in the interface.

Well, so consciousness won't look like neurons, but the...

One could build... I'll put it this way.