From: Eugen Leitl (eugen@leitl.org)
Date: Wed Jun 26 2002 - 11:29:52 MDT
On Wed, 26 Jun 2002, Michael Roy Ames wrote:
> Eugen: If you want to be *effective* in your fight, do your homework! Give
> us a good argument, not fuzzy, out-of-focus or generalized critisizms. I
Actually the proposed ethics drift mechanism via coupling between external
referents, the morality metric extracted from them and the coupling via
physical layer was not fuzzy at all.
The metapoint of above: assuming the gene-monkey finds a single genuine
flaw in the design, regardless whether fixable or not, the credibility of
the entire design is nuked for good. The system will see its first acid
test in deployment, and it's irreversible once deployed. If it breaks in
unintended ways, as it is bound to under onslaught of the SI torture
testing, we'd be a bunch of very sorry monkeys. If we're very unlucky, we
might not even die, and love every single second of it.
> do not support FAI because of irrational belief, but because of a detailed
> rational argument. If futher arguments (or evidence) were presented that
> invalidated FAI, then I would not continue with my support, but would move
> to the 'next best' option... maybe one of your suggested options!
I will try to make it readable, and to the point. All I ask for is an open
mind.
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:39 MDT