I think that I take issue most with you saying that any simulation made by human programming will never be able to conceptualize emotion. What if I created a perfect simulation of a human brain, as is currently being done with various supercomputers?
Can you send/link me some info on that? Even the names of lead researchers would be great. I am intensely interested in this, though doubtful. Modern cognitive science isn't even close to understanding the brain, much less code a perfect reproduction.
Supercomputers, of course, solve everything.
Well, for example there are the people around Henry Markram in Lausanne that try to
simulate cortical networks in great detail on a Blue Gene supercomputer. However, afaik they can simulate maybe hundreds of thousands of neurons, whereas the human brain has a hundred *billion* neurons.
That's partly because they use so detailed simulations of neuronal morphology. However, many neuroscientists think that's getting nowhere anyway, because the brain is just so complicated and full of details that might not actually matter for the "essential" processing, and so much is unknown, so that "we simulate everything from bottom up" might just be futile.
What is more likely IMO is that we will develop AI that shares capabilities of humans but is implemented in different ways, and that exactly will make it so difficult to tell whether it has consciousness or emotions. Personally I haven't thought much about emotions, but for consciousness I start to think that consciousness might actually come about simply from certain types of information processing and representation, and I'm not so sure anymore if my laptop might not have consciousness in some way
Edit: I can give an example of a model with potential for consciousness if anyone is interested, but this is getting somewhat off topic here...