From: hal@finney.org
Date: Tue Oct 17 2000 - 11:58:07 MDT
There is selection pressure for the replicators to advance as quickly as
possible. Robin basically analyzes the steady state, the "climax ecology"
of the replication wavefront. At that point it may be that the best
strategy is a lightweight replication, similar to what Eugene supplies.
A few days or weeks disassembling asteroids, building solar collectors,
masers and probes, and you're off. You then build defenses to keep
competitors from coming in afterwards and using the remaining system
resources to leapfrog you. After the wavefront is far enough away,
even these are no longer needed. Evolution is through with you.
However it seems unlikely that an intelligent being would intentionally
launch such an effort. There would be no point. The wave accomplishes
nothing; it passes through the universe, leaving it largely unchanged.
No one benefits.
What seems more plausible is that a traditional colonization effort could
evolve into a space race, which would eventually culminate in the cosmic
burning that Robin describes. As time goes on, original motivations like
colonization, information collection, and communication with the nucleus
of the effort would be downgraded in importance. Those colonies which
emphasize speed of spreading come to dominate. This slippery slope
eventually leads to the mindless replication wave.
Given that this potentiality exists, wouldn't it be relatively easy
for the colonizing agency to prevent it? Could it "lock in" certain
principles in its daughter colonies, designed to make sure that the
built in motivations are not lost?
Evolution requires variation and selection. But variation can be
controlled through engineering principles. Our own DNA is arguably
designed and evolved to promote successful variation. Engineers can
choose different principles to prevent variation, or to control it.
Given enough time, any engineered system will break. In the long enough
run, built-in constraints will be violated. But it seems plausible that
this can be made much longer than the expected history of the universe.
Foresight recently proposed preventing evolution of nanotech replicators
via cryptographic checking. Such technologies should allow designed
systems to be subject to the whims of evolution only in those parameters
allowed by the creators.
Evolutionary based analyses are powerful and suggestive, but they can't be
taken literally without a mechanism to show that variation and selection
will occur at effective rates. Even without any externally imposed
restraints, it took trillions of generations on Earth for replicators to
evolve significant multicellular complexity. Once you are talking about
engineered systems built by intelligent beings for specific purposes,
the role played by evolution may be much smaller than in Earthly life.
Hal
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 15:31:39 MST