I agree with Musk
Hear me out.
The semantic web is where AI will learn about the world of facts and things. Robotics will give it a variety of bodies. Machine learning and inferencing has been cracked I'd say. The law of accelerating returns will do the rest.
I read about AI in sci-fi when I was younger and I dreamed that maybe some day there'd be AI but that it would be in the very distant future. Then I read about Turing machines and computational equivalence and I asked myself, "is there something inherent to biology (carbon) that makes it superior to artefacts (silicon) in nurturing intelligence". I reasoned that no there isn't. I read The Emperor's New Mind by Penrose and wondered how someone so smart could come up with so fallacious a tome. Then I realised that people make up their minds first and then logic their way to their own position after.
AI will happen. The question is when. Kurzweil and Musk are the most optimistic among futurists. But if you read Kurzweil's book you see that he calculates the capability of the average human brain and plots the rate of increase of computing power and sees where the numbers intersect. He has been saying 2029 for AGI consistently for over 20 years. It is my belief that Moore's Law will continue until we get the raw computing power for a super-computer to surpass the raw compute of the human brain before 2030.
The next stage will be teaching and training this raw computational entity about the world it has woken up in. The semantic web, sensors, machine learning and inferencing is how that is going to be accomplished. How long that takes is anyone's guess but given that it takes a human about 18 years to be considered an adult and another 10 years after that to become fully educated then I think that 30 years will be the outer limit. So 2060 is the latest date. Given the DeepMind have demonstrated an algorithm that learns chess in 4 hours to a level greater than any human, it may not take the machine as long as 30 years at all. But chess is one thing, common sense, general knowledge, language abilities – not to mention wisdom is another kettle of fish. But if you see the advances that are being made in all these areas: computational linguistics, knowledge representation and reasoning you'd be mad to think otherwise.
Musk is an unusual person. Not only is he smart. He has a social conscience. He is not afraid to speak his mind. He is deeply philosophical. Jobs has nothing on this guy. Jobs got his minions to build shiny toys. Musk is trying to alter society by making it environmentally sustainable transport-wise and wholly sustainable by making it multi-planetary. People like him come along once in a couple of generations I think. Anybody here who dismisses him is either a fool, an idiot, or a bag of resentment.
My default position on AI these days is. You think it won't happen? Tell my _why_ exactly. Which bit of my reasoning is faulty?