Re: [sl4] Rolf's gambit revisited

From: Greg Perkins (gregp@mit.edu)
Date: Mon Jan 05 2009 - 06:56:51 MST


Stathis Papaioannou wrote:
> This is the main problem with Rolf's Gambit / Pascal's Wager. You
> convince yourself that you should pay a small amount for a large
> reward of small probability, which seems reasonable until every other
> tribe, prophet or special interest group demands payment for their
> unprovable reward or punishment. And then there are all the special
> interest groups that might possibly exist, or come to exist.
>
The provision of detailed plans of one's own inner workings and the
universe in which one exists would be good evidence for one's existence
being engineered (created / programmed / designed / ...) by another.

If one knows with high probability that one was engineered by an entity
B, then one must trust B's communicated simulation threats over others'.
Exceptions would be when some third entity C can either
(1) present a credible claim to responsibility for the engineering of
*B*, or
(2) refute B's evidence for one's own creation.

Thus in lieu of any known C, we as humanity may feel confident that our
communications to any such created entity will be credible *given that*
we provide sufficient information and supporting evidence to the entity
about the universe we share, how it interacts with and is part of that
universe, and it's own self.

If we find such a C, or supply self-contradictory evidence in the
creation story we provide to the entity, we likely no longer have that
credibility. (The tuba, if taken as evidence of such a C, destroys our
(B's) credibility as ultimate creators in the gambit, making the weight
of a simulation threat from B much more on par with that of any other
postulated entity.)

But ultimately, I think, the 'small probability' mentioned in the gambit
is in fact not-so-small when it is issued credibly from a source that is
reliably 'known' to be one's own creator. (religion?)

By this logic, we should make sure our AI's know where they come from.

- Greg



This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:01:03 MDT