Re: Cognitive Systems Conference, CCortex, and alternative friendliness aproaches

From: J. Andrew Rogers (andrew@ceruleansystems.com)
Date: Thu Jul 08 2004 - 00:23:20 MDT


On Jul 7, 2004, at 2:25 PM, Bill Hibbard wrote:
> It seems to me that efforts to control
> nuclear, biological and chemical weapons are a healthy
> precedent for controlling AI weapons.

I disagree, there is no parallel.

The control of biological, chemical, and nuclear weapons is generally
based on the control of necessary exotic infrastructure and materials
that are neither widely available nor generally distributed in consumer
retail channels. It is NOT based on controlling information, as that
is a fool's game and far more difficult at the very least.

AI weapons can be safely manufactured from ubiquitous consumer
technology with nothing more than a little information. As others have
noted in many different contexts, physical things are scarce and
therefore costly to replicate, but information can be replicated at
extremely little cost. Control of biological, chemical, and nuclear
weapons isn't based on controlling the replication of information, but
on controlling the replication of scarce physical items, something
basic economics makes far easier to do.

j. andrew rogers



This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:47 MDT