From: Peter de Blanc (peter@spaceandgames.com)
Date: Wed May 20 2009 - 11:50:16 MDT
Matt Mahoney wrote:
> (1) we would want to be rational,
> (2) we would want to be immortal,
> (3) we would want to know that we are immortal.
>
> And I claim you can't have both (1) and (3).
>
> By (1) I mean that if you are given a choice of two utilities (summation of numeric future time series denoting reward) and you can compute them, then you would choose the higher value.
>
> By (3) I mean there exist two computable utilities such that if you believed that you were immortal with probability 1 that you would make a different choice than with any other probability distribution over life expectancy.
Matt, you're confusing an agent A's knowledge of its own immortality
with other agents' knowledge of A's knowledge.
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:01:04 MDT