lmrenault writes:
It would be exciting to see unique human capabilities come out of artificial life/intelligence. But [...] this wouldn't do it for me simply because they would obviously be the product of an intelligent designer, i.e. man.
Doesn't that presuppose that we would have to know exactly how to build human-like intelligence? It could well be that that is the only way to achieve it, but my bets are on a much simpler route: we provide the initial rough conditions for a development process and let "nature" take its course.
I am thinking of two specific cases: genetic algorithms and neural networks. The first involves an evolutionary process, the second a learning process. But in both cases we would not know in advance the exact configuration of the end product, so we could not rightly be called the intelligent designers of it. At most we could take the credit for intelligently setting up the initial conditions, although we would only be mimicking things we've already seen in nature.
The crux is that the end product might easily surpass our expectations. As I've already described in
another thread, an experiment with a genetic algorithm that was supposed to evolve oscillators, produced a radio that picked up an ambient oscillating signal and simply passed it on. That's a very "creative" solution to the problem, and, most importantly, it never entered the minds of the experimenters that it could be done that way. So it wasn't, to use your words, "obviously the product of an intelligent designer", effectively proving that there
are other ways.
We are all atheists about most of the gods that humanity has ever believed in. Some of us just go one god further. - Richard Dawkins