From: Jeff Bone (jbone@jump.net)
Date: Sat Dec 08 2001 - 21:29:44 MST
"Eliezer S. Yudkowsky" wrote:
> There's a very big difference between saying "science moves forward,
> therefore your scientific disproof of my silly idea should not be trusted"
> and saying "science moves forward, therefore our extrapolation of the next
> billion millennia changes every fifty years".
True, but...
There is also a big difference between saying "long-term predictions made from current
scientific understanding *may* be inaccurate" and "long-term predictions made from
current scientific understanding *must* be inaccurate." What you are stating is much
closer to the latter than I am comfortable admitting, and again I will claim that
those kinds of comments are most often heard from people that are either ignorant of a
given field or predisposed to be antagonistic to the logical conclusions of a
particular field. Given that you're neither, Eli, then I am rather surprised to hear
you make arguments like this.
The science of cosmological eschatology is *much* more firmly grounded in real,
quantifiable, testable laws with a mass of available data and a track record of
prediction --- and an equal track record of gradual refinement rather than revolution
--- than, say, the relatively new art of predicting where technological progress may
lead.
Ahem. ;-)
Having said that, Dyson's a great person to look to for many of these issues. The
link earlier is a good place to start. BTW, I didn't point it out, but the answer to
the question of why Dyson's attempt to avoid the heat death conclusion is inaccurate
in the face of new information is actually *found* in his paper as a note: most
current models predict proton decay. This is consistent with some of the
information-theoretic implications of 2LT. (To conclude with a nice bit of symmetry.)
> Any attempt to establish "ultimate physical limits on technology" is a
> case in point. Do you really think that an attempt to argue an ultimate
> physical limit - as a reality, and not just as one model's interesting
> extrapolation - should be treated under the same rules as a disproof of an
> exploded flat-Earth theory? Disproving creationism or flat-Earth does not
> require that our current model be absolutely complete and that it stand
> for all time, merely that we take into account the overwhelming
> preponderance of negative evidence against a *disproven* theory.
I'm having a hard time trying to figure out how to explain to you why this scenario
you've just sketched out is broken, so I'm going to bail on the attempt. I think
we're butting up against an incomplete understanding of some fundamental
epistemological constraints, here. Suffice to say: no, I do not believe that such
things are subject to the same rules of disproof. On the other hand, this does NOT
speak to the unknowability of ultimate physical limits, technological or otherwise.
> Establishing a physical limit is the hardest kind of result to establish,
> since anything we don't know about could represent a possible loophole.
> Considering the rate at which this trick has failed in times past, one is
> justified in insisting that a given "physical limit" remain firm for, oh,
> at least a couple of centuries, before being accepted as anything other
> than an interim approximation.
So here's what's wrong with this argument: just as technology has been accelerating
non-linearly (perhaps asymptotically) over history, so has scientific understanding
(the accuracy of our models for making predictions at longer terms and finer levels of
"resolution") been accelerating similarly. (Indeed, the latter has driven the
former.) So over time, our ability to use scientific models as predictive mechanisms
has improved, yielding longer and longer-term detailed predictions with higher and
higher accuracy. At the same time, the "churn" in implications has been slowing
inversely --- i.e., the emphasis has shifted from the "what" to the "how" and even the
"how come?" This isn't just an opinion, it's an observation about the nature and
history of scientific (rational) thought.
> The historical record of "technological
> impossibilities"
Note we aren't talking about "technological impossibilities," rather logical and
physical constraints. Apples and oranges. "The world market for computers is around
five," "we will never put a man on the moon," etc. are all dumb statements. OTOH,
things like QED aren't about impossibilities, they are probabilistic models for actual
physical events.
> dying off almost as fast as religious text can't be used
> as an argument for the latest crackpot theory of perpetual motion, but it
> does mean that we should limit the confidence level of statements about
> *all possible* future technologies.
Modulo accepted physical and mathematical constraints that form the most essential
underpinnings of our most accurate theories: things like c, the Beckenstein Bound,
2LT, Godel's theorem, the Halting Problem, Chaitin's incompleteness theorems, etc.
While I place a high probability on at least one of those being incorrect, there's
good reason to believe that the probability of all of them being incorrect is
approaching zero.
To the extent that your world-view requires you to blissfully ignore the implications
of such things, you should recognize it as the former: a psuedo-religious expression
of faith used to support belief in highly speculative hypotheses. I want to believe
in Friendliness, Eli --- indeed I do believe that superhuman AI is an inevitability,
and I'd like for it to be benign. But honestly, your arguments are inspiring less
confidence rather than more. :-(
Hoping for a better tomorrow.
jb
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:37 MDT