From: Eliezer S. Yudkowsky (sentience@pobox.com)
Date: Sat Apr 05 2003 - 18:42:29 MST
Lee Corbin wrote:
> Eliezer writes
>
>>When I teleport, I want to anticipate a 100% subjective probability
>>of being where I'm going, not a 50% subjective probability of being
>>where I'm going. I therefore want to be moved, not copied.
>
> I think that it is the deepest mistake I've been able to
> identify in a decades-long examination of the identity
> paradox to suppose that it's a question of probability.
On the contrary; if you accept at least one level of parallel worlds then
"probability" is a word which refers to exactly this kind of branching.
That's why I analogized to many-worlds. If I flip a quantum coin before
getting in a car, and either many-worlds or open cosmology is correct,
then around half of the entities that are future histories of me get in
the car and end up at the grocery, while half of the entities that are
future histories of me stay at home. I'd just as soon get in the car
without flipping the coin, thanks.
> In other words, if you teleport to point B and your point A
> version is not disintegrated, then with 100% probability
> you arrive at B *and* with 100% probability you remain at A.
> Needless to say, this is highly counter-intuitive, and although
> evolution has prepared us for understanding being at two times
> in the same place, we are unprepared for the logically equally
> real possibility of being in two places at the same time.
There's a 100% probability that an Eliezer arrives at B, and a 100%
probability that an Eliezer stays at A, which means that my "subjective
probability" of going to B is 50%. Just as in many-worlds theory, there
is a 100% chance of the split occurring, and in this split some of my
measure definitely goes to "Eliezer at grocery" and some of my measure
definitely goes to "Eliezer at home", in two definitely existing parallel
worlds. Therefore, my "subjective probability" of going to the grocery is
|Eliezer-B|^2 and my "subjective probability" of staying at home is
|Eliezer-A|^2. Both branches definitely exist, of course. Perhaps you
have not sufficiently internalized this branching aspect of selfhood if
you cannot understand why I would want as much of myself as possible to go
into the best available branch; a nonbranching transition to Disneyland is
better than a branching transition to Disneyland, for the same reason that
a million dollars cash is better than a quantum lottery ticket. Or to put
it another way, if I were to flip a quantum coin as to whether to go to
Disneyland, I would prefer it to be weighted as much as possible toward
Disneyland - both copies of me will definitely exist, of course, but I
want as much measure as possible to go to Disneyland. So if I can send
practically all of my measure to Disneyland by not flipping the coin, or
through a nonduplicative teleport, why not?
Although you may have prepared yourself for understanding being in two
places at one time, you seem unprepared for the logically equally real
possibility of renormalizing continuous subjective probability measures
applied to multiple branches of yourself.
> That you *should* anticipate being in both places is demonstrated
> by the possibility that later the memories conceivably could be
> merged; at that point, it would be apparent to the future you
> that it had been a mistake not to anticipate both possibilities.
I *do* anticipate being in both places. That's *why* I want a move rather
than a copy. I *want* be in the grocery and I anticipate the frustration
of the Eliezer who finds himself still at home. When I snap my fingers I
should just go where I want, dammit!
-- Eliezer S. Yudkowsky http://singinst.org/ Research Fellow, Singularity Institute for Artificial Intelligence
This archive was generated by hypermail 2.1.5 : Sat Apr 05 2003 - 18:54:38 MST