If we consider ourselves as computational systems, I think there's a fundamental limit on our ability to understand ourselves. With an extremely accurate model

If we consider ourselves as computational systems, I think there's a fundamental limit on our ability to understand ourselves. With an extremely accurate model of our bodies, we could simulate ourselves and see what our mental state would be like in the next moment. But that wouldn't be any use from inside of a computational system, ie as a system of thinking, because the total amount of state space needed to perform such a computation would by necessity be greater than the amount of state space we have available for thinking in any given moment. (We'd have to include the total state of what we're already thinking about plus all additional information describing our current physical state.) This gets exacerbated in two "softer" ways: our brains are nonlinear systems, so trying to understand ourselves with just general principles can still result in significant deviations from predictions (for example, a single neuron could significantly affect our entire mental state in an un-accounted-for way). Secondly, this doesn't even try to account for the rest of the universe... On the plus side, I think this implies some fundamental limits for any artificial intelligence...?