From: Samantha Atkins (email@example.com)
Date: Thu Apr 07 2005 - 18:28:41 MDT
So are you saying that the VPOP cannot dependably be expected to think
before acting as it simply has no such distinction? If true this
would be a strong reason not to build such a thing.
On Apr 7, 2005 7:06 AM, Michael Vassar <firstname.lastname@example.org> wrote:
> >If the thinking module can decide how to organize its thoughts, then it can
> >decide how to organize the universe.
> I think that the sentiment behind this statement should probably be taken
> farther for emphasis because de-anthropomorphising is hard. For a Very
> Powerful Optimizing Process, there is no natural distinction between its
> thoughts and the universe. "thinking", just like all other actions, is just
> re-arranging molecules for the purpose of predicting and maximizing utility.
> Sometimes, one will think with a digital process, this is closer to what
> we call "formal thought", but is also what we call "running a simulation".
> Sometimes analog processes are used, this encompasses what we call
> "experimentation", but also "intuition". Keeping this lack of self in mind
> is important, because it encompasses many situations where the programmer
> would like to ask the GAI to "figure out X but don't DO anything". Figuring
> out X is a type of doing no different from any other, and like any other
> type of doing, involves taking in inputs and producing outputs. We just
> don't generally think of it that way when we do the sort of cognition that
> seems most like"figuring things out" to us.
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:51 MDT