Once Blue Gene can design de novo proteins with reactive groups in the correct positions

We do not know that Blue Gene will in fact be able to do this. A good number of biophysicists are of the opinion that the effort is hopeless, in particular due to the strong effect of solvent-ordering interactions and the likelihood that proteins are selected for conformational lability, "at the edge of chaos." Of course, this opinion may just as well turn out to be wrong. It does seem reasonable, as Drexler argued so many years ago, to suppose that protein engineering for nonbiological applications is a much easier problem than computational protein folding in biology. But how much easier, we don't know. There still are very few published success stories. Blue Gene may or may not provide a breakthrough. However, if it fails, in spite of its large scale, special-purpose-hardware, and task-specific optimized design, it may demonstrate that we are still a long way from having such a capability. A priori, it is hard to see why the most powerful protein-folding computer we are able to build at this point in time should just happen to be first one exactly equal to the task.

I'd like to think that IBM were aware of this

Oh yes, they are very aware of the potential use of such a machine in nanotechnology as well as biotechnology, and if the former turns out to be easier than the latter, I think they may not be too disappointed.

Protein may not form the backbone of future nanomachines, but it may well be a way by which we can make nanoscale components. Think enzymes: Proteins that catalyse a reaction by putting reactive components in the right place on the surface of a protein, causing them to self-assemble in a required way.

Once Blue Gene can design de novo proteins with reactive groups in the correct positions, a template for producing nanocomponents will be produced. Although we may well be able to join them by AFM by this point, it would also be possible to cause two proteins with appropriate binding sites to join together, thus bringing their cargo of nanocomponents into alignment.

I'd like to think that IBM were aware of this when they considered the design of Blue Gene.

Vik :v)