Well, my personal view is that I can see absolutely no reason why a proper general artificial intelligence is not possible (and yes, I know about a lot of the philosophical objections) - after all, we have an example of at least one kind of entity that displays general intelligence (i.e. us) and I can see no reason why whatever mechanisms the brain uses to create general intelligence couldn't be replicated by an "artificial" system.

Mind you, I think we are a long way off from actually having the theoretical or practical ability to build a true general intelligence - but I'm pretty confident we will get there eventually. Whether this development will actually turn out to be a good idea (Culture) or bad idea (Skynet) is another thing entirely.