From: Wei Dai (weidai@weidai.com)
Date: Fri Nov 15 2002 - 02:30:15 MST
On Thu, Nov 14, 2002 at 06:05:12PM -0800, Lee Corbin wrote:
> That's very perceptive of you. Indeed that does follow. I have
> started several threads on this list under the title of "Repeated
> Experience", and I defend that course of action in certain situations.
> It beats several alternatives, IMO.
I don't remember seeing such threads on this list, but a Google search
shows some posts by you under that title on the cryonet mailing list.
Would you care to summarize your arguments here?
> This is strange. Why shouldn't we want to value those subjective
> experiences that give us great joy or satisfaction?
In the future we'll be able to redesign our motivational systems and
consciously decide which subjective experiences will give us joy or
satisfaction. When we have that technology, your questions becomes "why
shouldn't we want to value those subjective experiences that we decided
will give us great joy or satisfaction?" which no longer makes sense.
Instead, let's ask which subjective experiences should give us joy or
satisfaction? I think it should be those experiences that reliably
indicate that we're succeeding in accomplishing our goals.
But perhaps you have no goals other than to have as many experiences of a
certain class as possible. In order to accomplish that, you'll probably
need to delay having those experiences until you've first gathered as much
resources as you can. And in order to gather resources efficiently, you'll
probably want to program yourself to feel joy or satisfaction whenever you
experience something that indicates you've gained resources. Now go back
and ask yourself why do you value the experiences that are your ultimate
goals? Could it be that you were once programmed by your genes to feel joy
or satisfaction when you experienced them, and then uncritically adopted
them as your ultimate goals even when they've outlived their usefulness,
even when your genes no longer exist?
> > An alternative is for us to develop strong irrational taboos against
> > directly manipulating subjective experiences and then continue to base our
> > decisions on expected subjective experiences.
>
> In your earlier post, I thought you said that all decisions should
> be made objectively. How am I misunderstanding you?
I think that is the ideal, but engineering compromises that only
approximate the ideal may have to be made due to resource and other
constraints. The above mentioned alternative is such a compromise. Of
course I'm speculating wildly about the plausibility of this compromise,
since I have no detailed knowledge of the actual constraints that we'll
face in our decision making procedures.
This archive was generated by hypermail 2.1.5 : Wed Jan 15 2003 - 17:58:08 MST