From: Eugene Leitl (eugene.leitl@lrz.uni-muenchen.de)
Date: Sun Apr 18 1999 - 22:35:51 MDT
O'Regan, Emlyn writes:
> It's interesting how quickly the grey-gooers can write off natural life,
Don't leap to rash conclusions, it's not quick. The debate has been
going on for years and years.
> as if we would be able to invent stuff which could get out of control
> and kill it all off overnight. Biological life is pretty hardy - how
Last time I looked human technology could kill off the bulk of life
pretty efficiently, with the lingering exception of a certain bipedal
primate (but we're working on it).
> long would it be before bacteria or insects or something developed which
> could feed on the alleged goo? After all, it would probably be made out
Bacteria and insects don't evolve rapidly enough to be able to
metabolize novel structures as graphenes, diamond, sapphire and
silicates while they are being busily eaten alive. The autoreplicating
nanomachines themselves could evolve as rapidly as viruses, or
faster.
> of similar stuff to natural life, given that it lives in the same
> environment and has the same resources available.
Species cannot sustainably occupy the same niche while having only slight
differences in fitness. The probabily of the new machines on the block
having the same fitness as biological life is nil, especially
considering the fact that their niche is very broad indeed.
> How quickly would natural life develop, which could outcompete the
> hard-life, and to whose presence the hard-life could not adapt? This is
One of the critical factors of a robust gray goo autoreplicator is its
ability to evolve rapidly.
> especially likely if the hard life was some heavily tuned soft-life
> eater, because it would have to be so specific that it would have
> trouble adapting to new environments.
The autoreplicator creates and maintains its inner inert
environment. Since evolution is open-ended, and a successfull design
will have robust evolvability its adaptability is unpredictable.
> Emlyn
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 15:03:34 MST