Turing level AI 2 decades?
There’s this columnist in Tuesday’s Sydney Morning Herald called Graham Phillipson. He isn’t the most insightful of people, and most of his columns don’t go anywhere. Among his past columns, he’s cited how Linux will go nowhere (a view which he later retracted, a few months down the track). Last Tuesday he wrote about AI. He reckons there will be Turing level AI in 20 years. He reckons that there will be 2 sentient species on this planet. Then he launches into a melodramatic spiel on how “There exists a very real possibility that human supremacy on this planet – and perhaps the universe, given the lack of proof of the existence of life anywhere else – will come to an end, and that the machine age will replace the organic age.” I’m surprised he didn’t mention The Matrix. I don’t think he understands the complexities of intelligence. As far as sentience goes, how are you going to make a computer self-aware? Computers aren’t dynamic. They are quite specialised. One human can perform many tasks, but a machine can’t. You know Robocup? Where they use those $2000 Japanese Aibo robotic dogs to play a soccer game? They estimate it will take 50 more years to create a team of robots that will be able to play a real game of soccer against a human team. And this is still far from achieving real Turing-level AI. 20 years? I don’t think so, Phillipson.
1:40am (GMT +10.00) tomorrow 
