Re: Why FAI Theory is both Necessary and Hard (was Re: SIAI's flawed friendliness analysis)

From: Cliff Stabbert (
Date: Sun May 11 2003 - 21:24:52 MDT

Hi, you wrote:

CR> Let me put it this way. Imagine a *succesful* Friendly AI
CR> programmer. Someone who really understands the process by which a
CR> Friendly seed AI will change the entire universe forever, making
CR> it a much better place for everyone. Can you expect them to get
CR> involved in the (relatively) pathetic little quarrels that are
CR> presently taught to us as human History?

I do get your point, but still think it's somewhat naive. We, as
humans, don't deign to "get involved" with this trifling little matter
called history. *It* tends to get us involved (cf. Joyce, "History is
a nightmare from which I am trying to awake.").

Presumably, being human and growing up in a human environment, the
people working on FAI will be embedded in human history -- have human
desires, human weaknesses, relationships with other humans, etc.
Becoming free of such involvement, for most people, is what would take
an effort, not becoming involved.

I am not entirely convinced that (what-I-perceive-as) Eliezer's
attempt to be as unencumbered as possible with the extraneous
business of life, i.e. anything not related to achieving a friendly
singularity, is a) successful b) healthy or c) helpful to his goals.
Perhaps it will turn out to be, but I'm certainly not buying an
argument based on imagination.


This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:42 MDT