From: Dan Fabulich (dfabulich@warpmail.net)
Date: Wed Jun 04 2003 - 21:39:19 MDT
Harvey Newstrom wrote:
> Actually, Bostrom postulates that these simulations are specifically
> ancestor-based, and specifically realistic. This seems to define them
> as exactly like reality in less exciting times. To me this almost
> requires the simulation to be boring. It must be of things we already
> know and of things we have already seen. Boring!
> [...]
> Nobody I know would build a world just like reality is today. I think
> we all would want to build better worlds that are more exciting and have
> less suffering. As an example, our own goals would lead one away from
> this type of simulation instead of toward it.
The use of indexicals here such as "this world" and "our levels" is
misleading. If this world is a simulation, then this world may be
considerably more exciting than its previous version. I don't feel like I
have a good measure of how "exciting" our world is. You seem to take it
that the answer is "not very exciting", but I don't see it that way at
all.
If the question is whether a posthuman would find us exciting enough, I
think I can say: there's at least a mediocre chance that we might be
interesting *enough*, based on the absurd computational power posthumanity
would have at its disposal. But this is enough to get the sim argument
going; the chances have to be almost *nil* to stop the train.
As for the argument about suffering, I think this is argument is utterly
misplaced. I don't need a posthuman theodicy to think that there's a
reasonable chance that an interested posthuman with arbitrarily large
resources might rehearse the brief history of life on Earth in ver mind,
perhaps in great detail, and that's all it takes to get the sim argument
off the ground.
> As such, simulations that allow us to question it are probably very
> rare.
This is a substantially better argument, but I don't like it in this
version.
In particular, what do you mean by "allow us to question"? If you mean
that most such simulations would probably include some simulated physical
law that would prevent us from questioning whether we were simulations,
that seems much too extreme, and probably wrong.
A simulation of, say, humanity's philosophical development couldn't
*possibly* be complete without allowing people to consider the possibility
that they are in a simulation, God's dream, or what-have-you. This
argument amounts to saying that in most sim universes, movies like The
Matrix would be prevented by physics itself. A cool superhero movie would
be possible, but considering the brain-in-a-vat case wouldn't...? This
seems absurd. Why wouldn't they just let us consider the possibility
without giving us a practical use for this information?
And, if not physical law, then what exactly would be preventing us from
questioning it?
Maybe I'm letting "question" be too weak. There's another analogous
argument which, I think, might be more successful: simulations in which it
would be USEFUL to know that one were in a simulation might be quite rare.
Call this the argument from pragmatism.
Of course, this flies in the face of Hanson's "How to Live in a
Simulation" paper, (which advocates that, in a simulation, you should try
to be interesting, entertaining, praiseworthy, and all around impressive,)
especially in light of your argument that they'd NEVER simulate a world
like ours because it would be too boring. I think if you really believe
that only extra extra interesting worlds would be simulated, you'd reject
my suggested argument from pragmatism, and conclude, with Hanson, that
there WOULD be a use for the simulation argument.
With that said, I think that *because* your excitement and suffering
arguments fail, so does the argument from pragmatism gain force.
Posthumans could, it seems to me, simulate us in *arbitrary* detail; the
more detailed the simulation, the less utility there would be in following
Hanson's advice. Maybe just being a human on Earth at such-and-such a
time is good enough to get you on the list of people who participated in a
pivotal event, the birth of posthumanity, and so no further toadying to
your simulated masters will be of practical use to you.
On the other hand, I said:
> > Here, I think, you have missed the point of this argument. Many
> > transhumanists still seem to believe that *we* will create simulated
> > environments, in which it will be impossible for our simulated subjects to
> > falsify the claim that they are in a simulation. This is a solid
> > prediction that makes claims about what kinds of worlds there will be:
> > namely, that there will be some simulated worlds that we'll make.
... and, of course, you missed the obvious response, which I only just now
noticed myself: the simulation argument does NOT make any predictions on
this point, but only claims to be a refutation of a prediction *if we
aren't in a simulation*.
But it is precisely the point that we'll have no practical tests for
whether we're in a simulation for quite a long time, if ever. Therefore
there is no prediction being made here: only a prediction in combination
with a metaphysical view.
There's something wrong with this, but it's not the verification
principle: the problem is that the information in question is useless.
We should act the same regardless of whether we're in a *detailed
posthuman* simulation. Hanson can't give you any advice for such a case,
and neither can anyone else.
Still, it is interesting to note that there are non-glitch events which
would count as a *confirmation* that we're in a simulation: if we simulate
our own ancestors, that makes it very *very* likely that we're in a
simulation. Not that we'd care, at all, but it'd still be, ya know,
interesting. (Like all metaphysical philosophical discussions.)
> Name me one transhumanist who wants to build a world as mundane as this
> one with our levels of tragedy and suffering.
At times, Tipler seems to argue that we would be simulated (or, as he
calls it, "ressurected") precisely as we were, and/or in certain kinds of
indistinguishable "educational worlds" that would teach us to be better;
hence any Tiplerian transhumanists of that vein would meet my criteria.
Certainly many of us have entertained the idea of ancestor-simulations in
what we took to be a realistic fashion, though I don't think anyone has
gone to the mat arguing that we WILL do ancestor-simulations. If you
doubt me, I suppose I can go archive-diving, but I won't slog through that
pointless task merely for your amusement. ;)
But remember that anybody who simply thinks that there's a good chance (or
even a FAIR chance!) that we'll run detailed ancestor-simulations would
fall prey to the argument. I'd argue that it would include a great many
Extropians of at least one time or another... it'd be hard to prove this,
of course, because, in general, nobody stands on a soapbox and argues for
such boringly wishy-washy propositions, but given that we're all
dynamically practical optimists and that ancestor-simulations would be so
easy for a plausible posthuman, I think I'm on pretty strong ground in
saying "many Extropians."
But remember that nobody needs to offer a convincing argument that
ancestor-simulations are *certain* for any posthuman worth vis salt. As
far as the probable interest in ancestor-simulations is concerned, we'd
just need to argue that it's not that unlikely, i.e. that f[I] is nowhere
near as small as even 100/f[P]/N[I] and so the product f[P]f[I]N[I] is
still huge and f[sim] is nearly 1.
-Dan
-unless you love someone-
-nothing else makes any sense-
e.e. cummings
This archive was generated by hypermail 2.1.5 : Wed Jun 04 2003 - 21:50:41 MDT