poly: ESS for HPLD

From: carl feynman <carlf@atg.com>
Date: Fri Dec 05 1997 - 11:50:50 PST

At 11:53 PM 12/4/97 EST, CurtAdams wrote:
>
>In a message dated 12/4/97 1:17:33 PM, carlf@atg.com wrote:
>
>>"What is behind me no longer matters!" ... cinders and derelict
>>machinery.
>
>That's only a rational strategy if you have a meaningful chance of
>outracing all other and colonizing the entirety of a wedge of the universe.
>Two groups which race each other forever leaving a trail of cinders
>obviously do worse than one which stops eventually - at least that
>group gets something.

Ah, but does it want that something? Evolution creates desires; it does
not conform to them. It does not maximize overall good, either, so even if
everyone obviously does worse under a given circumstance, that doesn't
indicate that evolution won't get to that circumstance.

Suppose we had two adjacent replicators on the frontier, one of which
agrees with your argument and decides to drop out of the race. Its
neighbor gets to take up twice as much of the frontier, and the ideological
purity of the frontier is preserved. Evolution favors replicators which
are pigheaded enough that they absolutely cannot be talked out of putting
all their resources into replication.

>Another point is that once the frontier is
>nearly a plane this hypothetical races is just giving up one oasis for
>another.

Well, once the Serengeti is filled with grazers, they're just giving up one
grass blade for another. But there are always more grass blades, just like
there are always more oases. The Serengeti also supports parasistic
nematodes that live and reproduce inside a single grass blade; it's a
perfectly acceptable lifestyle. It's just that evolution also rewards the
replicators that use up as well as husband their resources.

>Basically, I doubt that the groups would ash out their oases.

They might not have time to wreck them completely. If oases are star
systems, then the operative times are on the order of years. In a few
years, you can pretty thoroughly trash all the matter in the planetary
system, by burning all the fusible isotopes in the planets, but you can't
affect the star itself much if at all. It's just too big, and puts out too
little power, to be modified in less than on the order of 10^5 years. So a
more patient ecosystem, that was willing to subsist on solar power rather
than controlled fusion, could recolonize the 'cinder' solar system and live
there for a long time.

Similarly, if oases are galaxies, it takes a few times 10^6 years to travel
between them, which is long enough to extract most of the nuclear energy
from their stars. However, the gravitational potential energy of the
galaxy itself would take much longer to release, so a 'cinder' galaxy could
be recolonized by something willing to wait 10^8 years for a payoff.

>Have you consider intergalactic space? Oases might be farther apart there,
>necessitating alternate strategies.

No doubt. I've been trying to analyze the problem at a level of
abstraction where I don't get bogged down in particular cases. If oases
are comets, interstellar space is a big gap. If oases are stars,
intergalactic space is a big gap.

--carlF
Received on Fri Dec 5 11:50:20 1997

This archive was generated by hypermail 2.1.8 : Tue Mar 07 2006 - 14:45:29 PST