Still Twenty Years to AI

BBC News highlights a prediction by Raymond Kurzweil that we will achieve human-level artificial intelligence in twenty years.

You know, for all we know, maybe he's right.

As for myself, I'll make the far more secure prediction that in each of the twenty years between now and then technoboosters for artificial intelligence will continue to predict that we will achieve human level artificial intelligence "in twenty years."

I say that this prediction is secure, because it matches perfectly the predictions that we will achieve human-level artificial intelligence in twenty years that have been made in each of the twenty years prior to this one, and in each of the twenty years prior to that as well.

Again, maybe this time Kurzweil is the lucky AI technobooster who managed to get it right. Eventually, I suspect somebody who makes this prediction will be. But it is hard for me to see why this prediction is newsworthy, particularly, though it is quite obviously the sort of stuff that sells.