Amid all the phantasmagoric special effects and puzzling symbolism of Stanley Kubrick's 1968 film "2001: A Space Odyssey," the one detail that sticks in everyone's memory is the HAL 9000 computer that ran most of the spaceship Discovery One's operations. Not only did HAL speak and possess a humanlike personality, but it actually was superior to humans, because it never made mistakes.
But when the year 2001 actually rolled around, what inventor and futurist Ray Kurzweil calls "strong AI" -- that is, a machine that possesses self-awareness and is equal or superior to humans in intellectual ability -- was still closer to science fiction than reality.
That's with a few caveats. In 2011, the supercomputer Watson went head-to-head with "Jeopardy!" contestants and won mightily. However, just because a computer can answer questions -- and Watson is a very sophisticated "question-answering machine" -- it doesn't necessarily mean it's smarter than a human [source: Markoff].
In a 2005 essay, Kurzweil, who estimates that a computer would need to be capable of performing 10 quadrillion calculations per second to match all the regions of the human brain, predicted that threshold would be reached by 2020 [source: Kurzweil]. (Watson uses 80 trillion operations per second, the slowpoke [source: Deedrick].)
Others, such as Microsoft co-founder Paul Allen, have expressed doubts about whether machines ever will be able to even approximate human intelligence. He noted that neuroscientists still don't really know enough about the brain yet to hope to duplicate it. "You can't create an artificial intelligence," Allen insisted in a 2012 Forbes interview, "unless you know how the real thing works."
Check out our other countdowns of fascinating inventions from throughout history.