I respect Sterling's opinion; he's very talented as a futurist, and his writing's quite enjoyable. I see his point with a lot of what he says, but I think he's somewhat misinterpreting the singularity idea.

Unlike two of his historical examples - the A-bomb and the .com era, the coming singularity is not just an unexpected event that will eventually either die out for lack of interest, or be assimilated into popular culture.
It is a true asymptote, where the curve representing rate of technological change goes vertical!

At some point, even if computers don't "wake up" and start thinking for themselves, they will be able to simulate the thought process well enough so that it does not effectively matter whether or not they are conscious. They will become better problem solvers than we are, and will leave us in the dust. When knowledge becomes something that can be digitally copied and transferred effortlessly, the laborious process of "learning" becomes obsolete.