From: Wei Dai (weidai@eskimo.com)
Date: Fri Nov 30 2001 - 20:00:04 MST
On Fri, Nov 30, 2001 at 11:41:43AM -0800, hal@finney.org wrote:
> I was thinking more in terms of survival, replication and evolution
> rather than how it would feel to the agents. We should be able to make
> predictions about their actions based on what behaviors lead to maximal
> reproductive fitness.
I think this is a very attractive approach, but there's an issue of whose
behavior are we interested in? If we look at software copies then we'll
probably find that most of them think identity is preserved by uploading
and copying. If we look at physical copies then we may find that most of
them think identity is not preserved by uploading and software copying but
is preserved by physical copying. If we look at originals then we may find
that most of them think identity is not preserved even by physical
copying.
Another way to think about this is what does fitness means in a posthuman
world? For example someone who is very promiscuous about allowing his mind
to be copied might end up having many copies of himself, but most of them
may become permanently enslaved by others. Do these slaves count toward
his reproductive fitness? Or consider a superintelligence that manages to
maintain mind coherence as it expands in space and stays a single entity
(think the Blight). Does it have low fitness because it doesn't reproduce
in the traditional sense? Or consider a physical copy versus a software
copy, or a running copy versus a static stored copy. How do each of these
count towards reproductive fitness?
This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 08:12:18 MST