The next A.I. winter will be a very cold one.
For those wondering what I mean exactly: we're seeing the term A.I. being used in marketing, in the papers, in the news. Yes, we are making great strides in weak A.I. but strong A.I.? The kind we read about in stories? The kind of A.I. the public thinks of when we say A.I.? Asimovian Robotics A.I.?
Smoke and mirrors[1]. People develop new techniques and algorithms which are moderately self-learning in a focused way. The general public presumes this to be the basis of a general intelligence which can evolve (magically) to be like another form of life. Soon, everyone jumps on the A.I. bandwagon. The future must just be around the corner!
Then the uncomfortable details emerge; strong A.I. is not a matter of faster processors, more memory, or even more advanced/well-designed programming models. Rather, it is that there is still some fundamental aspect of real, human-like, or even animal-like intelligence that, to this day, eludes our understanding of intelligence.
A.I. winters have occurred many times before in many countries. The United States in the early 80s, for example, was pulling hair over the cybernization of the Soviet Economy. The US highest levels of government gloomily predicted that massive mainframes, given enough information and processing power, would become self-learning and turn the communist laggard economy into a powerhouse.[2]
I think maybe one day A.I. could happen. I think one day I will be proved wrong. Regardless of how A.I. comes about, it will not be due to the label of "A.I." slapped on any kind of product that remotely resembles intelligence. [3]
[1]https://en.wikipedia.org/wiki/AI_winter
[2]http://nautil.us/issue/23/dominoes/how-the-computer-got-its-...
[3] Between the winters, people call their stuff A.I. for the sexiness factor. When called out on the implications of the term, those same people retreat to the textbook definition. "It's A.I!.... well, technically it's weak A.I..."