Douglas Englebart had a pretty good approximation of what much of our information technology would be like today over 40 years ago.
http://sloan.stanford.edu/mousesite/1968Demo.html
Of course, this was because he put together all of the necessary underlying technology. I remember reading an interview with him in the 90s, asking him if he was surprised how fast technology was moving. He replied, no, he was shocked how long it took for the things he had working in the lab decades ago to reach the main stream.
That's the entire point of this (sub)thread. Making predictions of the next 40 years based on the last 40 is likely to lead you astray. A whole bunch of things are coming to a head. Coming revolutions in biotech alone are likely to seriously shake things up, IMO.
We consumers actually don't do that anymore: http://en.wikipedia.org/wiki/Concorde
At the same, most of the reasoning about the future by even intelligent people still tends to involve linear extrapolation rather than exponential extrapolation - that's what makes sense to us. Thus it's more likely for a standard prediction is go wrong in the direction of the exponential trends continuing rather than in the direction of the trends stopping.
Moreover, one or another exponential trends might stop but the overall mine that Moore's Law comes out of, miniaturization, is not going to be exhausted at least until human construction reaches the nano scale.