I really appreciate your feedback and your questions! 🙏
I’d love to reply in detail but it would take me a while. 😅 But maybe two quick points:
On observer independence: The main challenge that computational functionalism faces (IMO) is that there’s no principled way to say “THIS is the (observer-independent) system I posit to be conscious” because algorithms, simulations, etc. don’t have clearly-defined boundaries. It’s up to us (as conscious agents) to arbitrarily determine those boundaries, so anything goes! The section “Is simulation an intrinsic property?” in this post sums it up quite neatly, I think. Field topology, as well as, say, entanglement networks, do give us observer-independent boundaries.
On the simulation behaving identically to the brain: Here I think one could reasonably ask: What if, in order for the simulation to behave identically, we had to simulate the brain even at the smallest physical scale? Many people think this isn’t necessary and that the “neuron as digital switches” abstraction is enough. But say we actually had to simulate EM field phenomena, quantum phenomena, etc. Then I think runtime complexity matters, since maybe some parts of the brain can be simulated easily and others take millions of years. Can one bootstrap a coherent simulation from that? Now imagine trying to simulate multiple brains interacting with each other, running physics experiments, etc. Can one set up the simulation such that e.g. they all measure the speed of light to be the same? Or otherwise always get the same experimental results? I kind of doubt so. But regardless, the previous point about observer dependence would still stand.
I really appreciate your feedback and your questions! 🙏
I’d love to reply in detail but it would take me a while. 😅 But maybe two quick points:
On observer independence: The main challenge that computational functionalism faces (IMO) is that there’s no principled way to say “THIS is the (observer-independent) system I posit to be conscious” because algorithms, simulations, etc. don’t have clearly-defined boundaries. It’s up to us (as conscious agents) to arbitrarily determine those boundaries, so anything goes! The section “Is simulation an intrinsic property?” in this post sums it up quite neatly, I think. Field topology, as well as, say, entanglement networks, do give us observer-independent boundaries.
On the simulation behaving identically to the brain: Here I think one could reasonably ask: What if, in order for the simulation to behave identically, we had to simulate the brain even at the smallest physical scale? Many people think this isn’t necessary and that the “neuron as digital switches” abstraction is enough. But say we actually had to simulate EM field phenomena, quantum phenomena, etc. Then I think runtime complexity matters, since maybe some parts of the brain can be simulated easily and others take millions of years. Can one bootstrap a coherent simulation from that? Now imagine trying to simulate multiple brains interacting with each other, running physics experiments, etc. Can one set up the simulation such that e.g. they all measure the speed of light to be the same? Or otherwise always get the same experimental results? I kind of doubt so. But regardless, the previous point about observer dependence would still stand.