There’s a nice piece in Slate by Paul Boutin reporting on his trip round the Hewlett-Packard labs in Palo Alto. The opening stresses the evolutionary, short term, character of the research going on there, stressing that these projects only get funded if they are going to make a fast return for the company, usually within five years. The first projects he mentions are about RFID (radio frequency identification), and these are discussed in terms of Walmart, supply chains and keeping track of your pallets. I can relate to this because my wife used to be a production planner. She used to wake up in the night worrying about whether there were enough plastic overcaps in the warehouse to pack the next week’s production, but she knew that the only way to find out for sure, despite all their smart SAP systems, was to walk down to the warehouse and look. But despite these mundane immediate applications it’s the technologies that are going to underlie RFID that also have such uncomfortable implications for a universal surveillance society.
The article moves on to talk about HP’s widely reported recent development of crossbar latches as a key component for molecular electronic logic circuits (see for example this BBC report, complete with a good commentary from Soft Machines’s frequent visitor, Philip Moriarty). The author rightly highlights the need to develop new, defect tolerant computer architectures if these developments in molecular electronics are to be converted into useful products. This nicely illustrates the point I made below, that in nanotechnology you may well need to develop systems architectures that accommodate the physical realities of the nanoscale, rather than designing the architecture first and hoping that you’ll be able to find low-level operations that will suit your preconceived notions .