Michael S. Lorrey wrote:
> The concept would also be useful for spreading neural
> processing kernels across the net, to work together as a test AI..
That might be OK for doing small-scale tests during the development of an AI, but I don't think it would work very well for actually running it - the latencies involved in Internet traffic are just too long. It might, OTOH, be useful for doing very large-scale decision tree searches on well-defined problems.
Billy Brown, MCSE+I