(November 21, 2015 at 3:55 am)Darkstar Wrote: Of course, I am under the impression that in order to perfectly simulate all of the neurons in the human brain you would need to discover how all how all of those neurons work first. But then again, maybe not. Maybe you really could just put them together and get an apparent mind out of it even if you don't really know why it works. There would still be other chemicals you would need to take into consideration, though, like neurotransmitters, so maybe that wouldn't work after all...
Yeah. One problem is that we don't actually know what is there because evolution was limited with working with certain materials, and what aspect of it we need to implement to reproduce the same functionality on a computer and why. Generally when implementing AI on a computer it will be for a far more specific problem whereas the same brain is adaptable to many different problems and it may be that it can do this because it can reconfigure itself in so many different ways. Or it may be that if we understand how and why the brain manages to adapt we might not need computationally expensive simulations and will be able to reimplement it in a way that is more suited to a computer.
