10:57:14 From Kelly Smith to Everyone : @ Eric - perhaps yes, perhaps not.  I don’t think we are really well positioned to say with any confidence at the moment

 

10:58:45 From John Malloy to Everyone : @Eric I agree with you, and I’ll add that to be useful, any complexity measure must be strongly based on some physical system. A theoretical or arbitrary measure (Geoffrey West’s power laws on technological innovation, for example) are analytical on human systems, but not too useful to me because they’re arbitrarily based.

 

11:01:32 From Eric Hughes to Everyone : Kolmogorov complexity, based on computation, is invariant (within a constant, that is, O(1)), but it's contingent upon the symbolic representation. What's much less studied is the notion of a set of admissible symbolic systems on a physical situation

 

11:04:47 From Eric Hughes to Everyone : ... and I should add that there's a problem of infinite regress in thinking about this, because if you model "physical reality", say, as some kind of mathematical object, you've already reduced the physics to a symbolic system.

 

11:15:09 From Eric Hughes to Everyone : That illustrates my whole point: for simple probability with dice, a world where coincidences are different can't exist, not without positing even less plausible explanations.

 

11:27:17 From Dan Capper to Everyone : Excellent presentation, Robert.  I'll never look at "20 Mule Team" the same again.