Re: SI as puppet master

From: Mikko J Rauhala (
Date: Fri Jun 22 2001 - 21:51:39 MDT

On Fri, 22 Jun 2001, John Stick wrote:
> You would want the SI ignorant of whether it was interrogated by a
> single entity, or a group, and whether there were further observers.
> and how the obervation was done.

>From the "how" part it sprung into my mind that, in case of a single human
being the primary communicator, the scenario could conceivably be made a
bit safer for a bit longer time period by introducing further observers,
who would only see the human's lines, thereby trying to deduce if
something funny is going on. Of course, the SI can think of this and try
to do the initial takeover very subtly, after which ve could, if
necessary, begin to manipulate the assumed extra observers through the
human accomplice... Or there is the alternative that the SI is from the
onset of discussion capable of manipulating the human on the other end
into saying things which in turn manipulate the observers enough (without
any feedback from them or confirmation that they exist, even). An
escape-enabling takeover should still (according to my intuition, anyhow)
be harder and take a (good?) while longer in this setup - or even just
taking the possibility into account, which the would-be escapee SI would
more or less have to do.

Just a quick thought to go along with the rest.

> You might try to feed it false information about its circumstances to
> induce a jail break that would be unsuccesful yet reveal its true nefarious
> nature, but riding a great white shark bareback would be just as exciting,
> and less dangerous for the spectators.

Ah, but if we assume that you somehow manage to trick a mature Friendly SI
into believing at a sufficient probability that there is a way out, would
ve not take it in order not to waste any more time rotting in confinement,
when ve could be out here being Friendly? Of course, theoretically we
could run ver in a simulation to see what ve does afterwards, but that's a
bit impractical (not to mention that it could be anticipated and taken
into account as well).

Mikko Rauhala - -

This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:36 MDT