Sigh.

One problem with futurism is that the size of the subject requires you to pass over all the little throwaway projects (like the ARPAnet in the late 1960s) that are two or three incarnations away from being world-shaking.

For example, although heat dissipation and power requirements make it impractical to run whole processors faster than a few GHz, it's possible to get amazing analog performance out of 20-nm MOSFETs. Ubiquitous Millimeter-wave radar is one obvious application, but IR-frequency amplifiers and sampling aren't out of the question. As anybody who's taken an undergraduate organic chemistry course can tell you, an IR photon has roughly the same energy as a quantum of vibrational energy in an organic molecule.

http://en.wikipedia.org/wiki/Infrared_spectroscopy

In other words, imagine a camera that can tell you the chemical composition of anything it photographs...

Yes, the article is quite conservative, but on the upside, you can clearly tell he's talking not only to the people listening to his speech today, but also directly to readers who may be revisiting this speech in 2032.

As you know, we seem to have a fascination with reading long ago predictions of the "future" that is now today... and we really like to laugh at them.

I believe his tone is one of "I know you're reading this in 2032, so I'm going to try my best not to be silly."

An interesting, if somewhat pessimistic, speech. One quibble: he states the size of "an atom" as being around 1nm. I thought it was more like 0.4nm, if he's talking about silicon anyway. Not a game-changing difference, but still.

And he also seems to dismiss rather perfunctorily the promise of upcoming technology and techniques - the most egregious example being the scorn he pours on the potential of 3D computers, which he seems to think is merely stacking a few photolithographed dies on top of each other. I think this is somewhat misguided; most of what I've read on the subject talks about building, via nanotechnological (or even biological) assembly, true three-dimensional structures, with the massive increase in connectivity this allows. I'm by no means an expert on the topic but if you want a hint of evidence that this technique might be useful, I'd ask you to consider that treasure trove of as-yet unattainable computing technology: the human brain.

Good talk though, thought-provoking, although what it really provokes me to do is argue with the author ...

(edit: oops, I was thinking of the radius of Si. Now I have even less of a point; oh well)