It's also possible that all these approaches (1-5) are different viewpoints of some more fundamental, underlying concept/description of intelligence and will converge at some point (as happened to Turing Machines vs. Lambda Calculus). This would be very interesting.