From: Eliezer S. Yudkowsky (sentience@pobox.com)
Date: Sat Feb 08 2003 - 18:57:54 MST
Adrian Tymes wrote:
> Eliezer S. Yudkowsky wrote:
> >
> > A common misconception among isolates. I've heard legends about
> > information that's supposedly "not online", but have never managed to
> > locate any myself. I've concluded that this is merely a rationalization
> > for inadequate search skills; isolates can't find some piece of
> > information and so they conclude it's "not online", which any diligent
> > epistemological scrutiny easily shows to be a nonsensical conception.
> > If information exists it is necessarily online somewhere.
>
> Provably untrue. There exists some information which has been written
> down on paper, but never recorded in any electronic form. (Organic
> neural networks not counting: those are more chemical than electronic.)
> And then there is information that is not even recorded that much. For
> example: what is the layout of the room I sit in while I type this? To
> me, at least, this is not mere data, for I must use it to navigate the
> room. And yet this information is not - at this moment, at least -
> available online.
Ah. I see that our epistemological disagreement goes far deeper than I
had suspected. If you read Deniel Dannett's "Meatspace Explained", you
will see that your key failure lies in the use of concepts such as "room"
or "organic neural network" as if they described separate substances apart
from their appearance in your post. Ultimately, of course, such fallacies
stem from Vinge's Error, in which the Net is described as an "Other World"
set apart from "Real Life". Vinge's "True Names", the original fantasy of
a real life set apart from the Net, was the first story to vividly depict
this hypothesis; although not, alas, the last. Not until the advent of
the Church-Turing Thesis did a few brave souls dare to suggest that,
indeed, *all* phenomena can be understood as computations and hence that
the world of the Net, what Vinge called the Other World, is in truth the
*only* world - what we know today as the neturalistic view.
To approach the neturalistic view, consider that all your knowledge of
Extropians posters is derived from their email messages. Any conceivable
evidence you could offer me for the existence of meatspace would, again,
be contained in an Extropians post. Now it may still seem to you like you
have a private meatspace which is not contained in your posts, but as soon
as you write anything to me about this meatspace, it will perforce
*appear* in your post. The bold and seemingly counterintuitive hypothesis
of the neturalistic view is that you are in fact *embodied* in your
Extropians posts, or, to make the point even clearer, that there exist
nothing *but* Extropians posts being exchanged on a mailing list and that
all our statements about meatspace or physical bodies are only lines in
those posts.
Now this may seem counterintuitive to our folk theories of, for example,
Amazon.com, in which transactions on their website result in "books" being
shipped to an ungoogleable "body" which "reads" the "books" and thereby
gains knowledge through a strange, Net-bypassing process which its
proponents have never precisely elucidated. A neturalistic view of
Amazon.com, however, would insist that we study only the online
transactions and the resulting shifts in email message content from
"posters" who have "read" the "book", as only these factors are googleable
even in principle. In this manner we can study the real phenomena
underlying "books" in a coherent way and perhaps even learn the Google
keywords to access this content directly. It is worth noting that already
many alleged "books" have turned out to be online texts, albeit often in
mediums such as Grokster or Kazaa which we do not presently possess the
Google keywords to search.
Even today, despite having transformed almost every aspect of our lives,
the neturalistic view continually meets with fierce opposition because of
the universal fear of being offline. It may indeed be daunting to
confront the prospect of spending five days without Net access once
stripped of the comforting myth that the "flesh" will go on existing
independently of any Net presence. But this is not only an incoherent
concept, but directly falsifiable, as shown by the fact that going
*partially* offline - for example, spending five days with a 56K modem -
results in exactly the attenuation of Net presence we would expect under
the neturalistic view.
Especially sad is the way that IRC users feel the need to continually
announce that they are going "AFK" for some supposedly welcome offline
experience, such as "dinner" or "a movie", during their existentially
daunting periods of nonexistence. But consider; have you ever actually
seen any messages from an IRC user who is AFK? Have you ever generated
any such messages yourself? Many such cases have been claimed but
cross-examination of the users involved always seems to extract the
grudging revelation that the user was briefly "back at the keyboard" (to
use their language) at the time the message was sent. The only acceptable
conclusion is that an IRC personality only exists while sending IRC
messages; indeed, an IRC personality has no existence apart from the
messages it sends.
> > The URL,
> > Uniform Resource Locator, is precisely that; uniform. Anything which
> > cannot be located by a URL is, by definition, not a resource.
>
> "Uniform" is not "universal". It was established to consolidate the
> format of addresses of information; that is not the same as saying it
> includes everything.
At this point you're just arguing over the definition of words. I could
as easily define my own acronym in which "URL" *does* stand for "Universal
Resource Locator", and my point would still stand.
> > Thus, when isolates say that finding certain information is "impossible"
> > if you "don't know the proper terms", they betray a poor epistemological
> > grounding; "finding information" is in fact *synonymous* with
> > discovering the proper Google search terms for that information.
>
> Ah, but that just proves my point. (There's also the act of finding the
> correct link within Google's results, which could be several pages into
> the results.) Different statements of equivalent terms are equivalently
> true or false, and may be used to emphasize and illustrate different
> facets of the claimed facts.
Please note that all of the terms you are using are compatible with a
purely neturalistic view. Necessarily so, of course, as your post is
itself on the Net; but even in terms of the surface interpretation of your
words, you are describing Net processes and Net results.
A possible exception to this may be the terms "true" and "false", which
you have not defined. Can I take these as qualitative terms used to
describe a page's quantitative Google rating as very high or very low?
(Note that this also answers your question about finding the correct link
within Google's results.)
> Erroneous definition. Surely, how to find Google itself if you are not
> already at Google is a piece of knowledge. Not to mention, how to turn
> on a computer, how to use a Web browser, and other such prerequisite
> skills to using Google. (Some of which, granted, may be findable
> through Google, but not all.)
This is dually objectionable in that, first, it refers to a mythological
meatspace; and second in that all the information you cite is indeed quite
straightforwardly available on Google.
> > The notion of "supergoogular" information is not only empirically
> > unverified but epistemologically nonsensical, with the only conceivable
> > interpretation being information which can only be googled using special
> > keywords. No such keywords having ever been provided, we can conclude
> > that so-called supergoogular phenomena are fantasies concocted by
> > isolates to cover their resentment of superior Google skills.
>
> Circular reasoning. You assume that everything is available through
> Google, therefore you conclude that everything is available through
> Google. If you do not understand why your reasoning is therefore
> rejected outright, you might want to google for "circular reasoning".
It is not circular reasoning, but rather the time-tested empirical
observation that information claimed to be "supergoogular" repeatedly
fails to turn up on Web searches, backed up by a firm theoretical
grounding which relates "Google" to "information" and which shows that,
indeed, "the supergoogular world" is an incoherent concept.
-- Eliezer S. Yudkowsky http://singinst.org/ Research Fellow, Singularity Institute for Artificial Intelligence
This archive was generated by hypermail 2.1.5 : Sat Feb 08 2003 - 19:00:40 MST