I also read Prof. Smalley's web page (http://cnst.rice.edu/dallas12-96.html), but, in the context of the Chronicle reporting that "Smalley and Rice President Malcolm Gillis say the Rice center will do no research into the assemblers Drexler advocates, that they aren't possible according to the laws of nature. Smalley adds that the name 'the Center for Nanoscale Science and Technology' was chosen partly to distance it from popular conceptions like Drexler's." I am much less convinced that Smalley regards engineered devices which are capable of self-reproduction as feasible. I do think that he NEEDS them for HIS proposals to succeed. He spends about half of his web page describing energy problems, and proposing solar energy as a solution.
There is a problem with this. We have solar cells, we've had them for decades. They work "...even on a cloudy day wherever you put these solar collectors...". The problem is cost, and a good chunk of that cost isn't even in the high-tech parts of the collector, but in things like glazing, installation, maintenance, wiring, and so on. A very key part of Drexler's proposals is that you can drive incremental manufacturing costs down to close to the costs of raw materials if you build a manufacturing system that is capable of self-reproduction. For this application, reducing the manufacturing (and, to a lesser extent, the deployement) costs is CRUCIAL. Smalley wants to make "clean energy that is so cheap that there is no longer any incentive to burn coal, gas, or oil."
Consider the following analogy: Suppose that some sort of non-replicator nanotechnology produced the best possible result for solar energy utilization. What might such a thing look like, and what is its closest analog today? Imagine some ideal self-assembled monolayer which turns incident sunlight into electricity, with practically no capital cost for the monolayer itself. Is there any analog today? Well, when solar energy is used for space heating, it is basically absorbed in what amounts to black paint. It works, it is available, it is reasonably cheap. Nonetheless, while houses heated with solar power exist, and aren't really rare, they are not particularly common either. I think that this is about the best outcome Smalley could expect from applying non-replicator nanotech to solar power. I think that he should either