[extropy-chat] Newcomb's Paradox (was Superrationality)

[Lee Corbin]

Russell writes

On 5/20/06, Lee Corbin <lcorbin at tsoft.com> wrote:

> > Yes, but then if we want to reason outside the boxes (as it were),
> > then one may wish to take only one box in Newcomb's Paradox in order
> > to show that one is a nice guy, or that one is not greedy, or some
> > other irrelevant consideration. 

> But that's the whole point.

Good heavens!  It is not!  As I wrote, the incentives in Newcomb's Paradox
(a million 1960 dollars in one box against a thousand dollars in the other)
were specifically designed to overcome these side considerations such as
what one's mother will think of one.

You are supposed to consider that you are *only* interested in walking away
with the greatest amount of money. An argument pro taking two boxes: "What
is in two boxes is logically more than what is in one box";  an argument
in favor of taking one box "It will turn out that I'll get more money by
just taking the one box". 

Nowhere in there  are there any considerations of whether one is a nice
guy, or that one is greedy, or that one is concerned about what other
people think.

> 1) Solution(Equation X) = Y. So what?

> 2) What we should do = defect.

You've lost me.  Aren't you talking about Newcomb's Paradox?  Indulge me,
and let's keep *this* new thread on the NP (thanks!).

> What the fuck? No way! People are foul hideous monsters if they
> contemplate that!! I hate humanity!!! Mathematics isn't true!!!!
> (Hofstadter's words, in fairly reasonably accurate paraphrase if
> my memory is even vaguely on the same continent as the mark.)

Dear, dear.  This really is one case where we shouldn't get carried
away. People are not "foul hideous monsters" (although there are a
few exceptions, of course). As Solzhenitsyn wrote, "the line between
good and evil goes through every human heart".

And I just have to let your "Mathematics isn't true" pass without comment.

> But #1 does not in any way imply #2. The only way you can
> conclude anything about real life from #1 is by putting in
> the "outside the box" stuff that makes it realistic. And
> when you do that, #2 stops applying.

None of this was supposed to be about real life. At least not until
we've solved the strictly theoretical cases of what the optimal
strategies are.  THEN, maybe----and only maybe---should we wonder
whether this has implications for real life.

As I begged in my original post

> > P.S. Fooey: it needs stating to some people that real-life situations
> > are *far* removed from mathematical theory, and that in fact I'd
> > cooperate with any human being I thought had a heart, and I advise
> > others to do likewise, because the other person's good is a factor too
> > (not shown, of course, in the payoff table).

Lee