The only thing that is clear is that there is an enormous technical race
going on.  I would tend to think that the decentralized nature of the
Internet favors those who want to subvert government censorship and
monitoring.  But I think the article missed a couple of important issues.

First, the nature of language makes the job of "thought police" more
difficult.  If there's anything I learned while managing search and
filtering products (the very ones used by the CIA, NSA, etc.), it's that we
don't understand language very well at all.  The ability to understand
language probably equates to intelligence, which means that we fundamentally
don't know how to write software that can reliably figure out what is being
said.  That helps the dissidents.  If nothing else, they can invent new
language -- as revolutionaries usually do -- that software alone will not
figure out.

Second, working in favor of the police, is the kind of technology my last
startup developed, which uses traffic analysis to spot patterns.  That
approach doesn't require the software to figure out causality, only
correlation.  Look at the Brin-L visualizations and you can tell who the
influential people are -- and that was generated without any linguistic
analysis at all.  All the Chinese have to do it scrutinize the people who
are highly central or are key links between groups by analyzing who talks to
whom.  Unlike other methods, that is scalable -- if you have the data.  But
it's a game of catch-up, inevitably.  A person has to become influential to
be identified; by then, they've had influence, by definition.  The question
is how disruptive they could be before being identified.

IMO, the battle between blocking software and work-arounds will go on for a
long time.  What's much scarier is the notion that the Internet quickly
allows anyone with access to the traffic statistics to figure out who the
leaders are.  If you have the resources to silence quickly anyone who
emerges as a leader, that's certainly a historically validated means of
suppressing dissent.  On the other hand, self-organizing wireless technology
might be extremely disruptive to traffic analysis, making many of the
interactions impossible to detect.

Strong pseudonyms, which are the only real defense against traffic analysis,
will come only out of networks that are so densely interconnected that
traffic analysis becomes impractical.  But I have no doubt that in China,
they can keep the network interconnection density to a level where the
analysis is practical.  Then it becomes a Darwinian issue, I suspect -- can
a poorly interconnected nation be competitive.

As for the U.S. companies being willing to do anything for a buck, I don't
expect that to change.  Too many, especially in technology, believe that
technology is neutral, inherently valuable or seductively "sweet," as Robert
Oppenheimer argued, or that it is simply inevitable, as Edward Teller
argued.  Of course, I'm deliberately referring to the inventors of atomic
weapons.

Nick

> -----Original Message-----
> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On
> Behalf Of Gautam Mukunda
> Sent: Sunday, February 17, 2002 9:05 AM
> To: Brin-L
> Subject: SCOUTED: China and the Internet
>
>
> http://www.weeklystandard.com/Content/Public/Articles/000/000/000/
> 922dgmtd.a
> sp
>
> I don't really have the technical background to assess some of the claims
> made here.  Nick?  You probably have some experience with the issues
> involved.  Anyone else?  It seems to be a pretty strong rebuttal to the
> techno-utopians.
>
> Gautam

Reply via email to