(Warning: This post assumes an familiarity with UD+ASSA and with the
cosmological Measure Problem.)
Observational Consequences:
1. Provides a possible explanation for the "Measure Problem" of why we
shouldn't be "extremely surprised" to find we live in a lawful
universe, rather than an extremely chaotic universe, or a homogeneous
cloud of gas.
2. May help solve the Doomsday Argument in a finite universe, since
you probably have at least a little more "measure" than a typical
specific individual in the middle of a Galactic Empire, since you are
"easier to find" with a small search algorithm than someone surrounded
by enormous numbers of people.
3. For similar reasons, may help solve a variant of the Doomsday
Argument where the universe is infinite. This variant DA asks, "if
there's currently a Galactic Empire 10000 Hubble Volumes away with an
immensely large number of people, why wasn't I born there instead of
here?"
4. May help solve the Simulation Argument, again because a search
algorithm to find a particular simulation among all the adjacent
computations in a Galactic Empire is longer (and therefore, by UD
+ASSA, has less measure) than a search algorithm to find you.
5. In basic UD+ASSA (on a typical Turing Machine), there is a probably
a strict linear ordering corresponding to when the events at each
point in spacetime were calculated; I would argue that we should
expect to see evidence of this in our observations if basic UD+ASSA is
true. However, we do not see any total ordering in the physical
Universe; quite the reverse: we see a homogeneous, isotropic Universe.
This is evidence (but not proof) that either UD+ASSA is completely
wrong, or that if UD+ASSA is true, then it's run on something other
than a typical linear Turing Machine. (However, if you still want use
a different machine to solve the "Measure Problem", then feel free,
but you first need to show that your non-Turing-machine variant still
solves the "Measure Problem.")
Decision Theory Consequences (Including Moral Consequences):
Every decision algorithm that I've ever seen is prey to paradoxes
where the decision theory either crashes (fails to produce a
decision), or requires an agent to do things that are bizarre, self-
destructive, and evil. (If you like, substitute 'counter-intuitive'
for 'bizarre, self-destructive, and evil.') For example: UD+ASSA,
"Accepting the Simulation Argument", Utilitarianism without
discounting, and Utilitarianism with time and space discounting all
have places where they seem to fail.
UD+ASSA, like the Simulation Argument, has the following additional
problem: while some forms of Utilitarianism may only fail in
hypothetical future situations (by which point maybe we'll have come
up with a better theory), UD+ASSA seems to fail *right here and now*.
That is, UD+ASSA, like the Simulation Argument, seems to call on you
to do bizarre, self-destructive, and evil things today. An example
that Yudowsky gave: you might spend resources on constructing a unique
arrow pointing at yourself, in order to increase your measure by
making it easier for a search algorithm to find you.
Of course, I could solve the problem by deciding that I'd rather be
self-destructive and evil than be inconsistent; then I could consider
adopting UD+ASSA as a philosophy. But I think I'll pass on that
option. :-)
So, more work would have to be done the morality of UD+ASSA before any
variant of UD+ASSA can becomes a realistically palatable part of a
moral philosophy.
-Rolf
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "Everything List" group.
To post to this group, send email to everything-list.domain.name.hidden
To unsubscribe from this group, send email to everything-list-unsubscribe.domain.name.hidden
For more options, visit this group at
http://groups.google.com/group/everything-list?hl=en
-~----------~----~----~----~------~----~------~--~---
Received on Sun Oct 21 2007 - 14:33:48 PDT