Note

Freeman Dyson gave a short introductory talk at the physics student's lunch today (March 7, 2001) and answered a few questions about his upcomming lecture on whether life is analog or digital in nature. During the session I mentioned Adrian Thompson's work with geneticaly "bred" FPGAs and made the observation that network circuits like Thompson's FPGAs and other evolved neural network type circuits have a very strong propensity to incorporate and exploit properties of their environment to accomplish whatever task they are evolving to accomplish. In the case of Thompson's FPGAs the properties in question are things like stray capacitances, inductances, etc. In the case of our own neural network based brains the items up for exploit might be microscopic systems exhibiting quantum behavior to be used as quantum computing subsystems. While these systems might not have originally been involved in the networks to start with, if they are present they will almost certainly be coupled to and used. His response to my suggestion was:

Physics Department Colloquium
Organized by Prof. Jonathan Wurtele

Date: Wed, Mar 07, 2001 5:45 p.m.

Location: Pauley Ballroom

Speaker: Professor Freeman Dyson of the Institute for Advanced Study

Abstract:

I have been enjoying a lively argument with Laurence Krauss and Glenn Starkman, who disagree with my conclusion that life might survive for ever in a cold universe by adapting itself to low temperatures. It turns out that both sides in the argument are right, because we made different assumptions about the nature of life. I assumed that life is analog, Krauss and Starkman assume that it is digital. The talk will explain what this means and why analog does better.