I have long been interested in spin glasses. It shares some common ground with fractals - see also the interview with Mandlebrot himself- as well as genetic programming. At this time I am putting together a course which I plan to give at San Francisco State on the applications of spin glasses to complex systems.
In 1991 I also collaborated with people at the Stanford Business School. This involved writing APL programs to do some data transfer from a DEC machine to an HP machine and I also wrote several programs which implemented the traveling salesman algorithm for a bidding project. The idea was to apply the method of simulated annealing, used in spin glasses, to implement the traveling salesman problem (TSP) algorithm. It worked but it was rather time consuming.