It's easy to dismiss this as so far off that it's not worth thinking about, but there are plenty of brilliant people with PhDs who think robots with sufficient capabilities to resemble strong AI in domain-specific ways (human-competitive in limited specific tasks) will come in the next few decades.<p>It would be impossible to overstate the magnitude of the consequences of such developments.<p>A few years ago Prof. Robert Hecht-Nielsen, director of research at the Fair Isaac Corp (they created the FICO credit score) and the confabulation lab at the University of California San Diego, gave a talk to my neural networks class about his very firm belief that real strong AI would come in < 30 years. He didn't convince me that it would happen, but he correctly pointed out that were it to come it would make economics obsolete and could lead to some kind of grand utopia if we don't fuck it up.