What makes you think evolving machines wouldn't have at least as much entropy and diversity as humans?
When you understand that from my writings, then you will have the "a ha" eureka epiphany. I think I already explained my hypothesis:
In order to increase entropy, there must be more chance, i.e. more cases of failure not less. Remember the point I made about fitness and there being hypothetical infinite shapes and most failing to interlock. Computation is deterministic and not designed to have failure. Once you start building in failure into robots' CPU, then you would have to recreate biology. You simply won't be able to do a better job at maximizing entropy than nature already does, because the system of life will anneal to it. Adaption is all about diverse situations occurring simultaneously, but don't forget all the diverse failures occurring simultaneously also.
I don't see sufficient utility to rewrite what I already wrote in my blog and explained further in this discussion.
The next step would be to develop it formally with some math, formal logic, and write a more comprehensive paper or book. I don't have the time right now.
Thanks for the persistence and (your) cordial discussion. May I opt out now?
P.S. the big take away is how most people view life as deterministic and can't get their perspective adjusted to the reality that life is all about widespread failure as trial-and-error to attain chaotic fitness thus resilience and adaptation.
This is why socialists think they can improve things as they fail to understand that inequality (i.e. diversity) and failure are integral with survival and fitness.