Re: Perversion attacks

Dan Clemmensen (
Sat, 26 Oct 1996 13:04:09 -0400

Anders Sandberg wrote:
> On Wed, 23 Oct 1996, Lyle Burkhead wrote:
> > > A perversion war would be *fast* on a planet (and the chilling thing
> > > is that many nations are gearing up for it - information warfare is
> > > HOT right now), and probably would be over quickly - leaving
> > > more normal kinds of conflict.
> >
> > I don't think so. Memetic warfare is anything but fast. It takes
> > centuries.
> Yes, but the kind of warfare Vinge (and I) talk about is not about memes,
> rather computer viruses, cracking codes, trojan horses and taking over
> operating systems. This could be very fast if AI, expert systems and
> genetic algorithms are used.

Interesting. Both Lyle and Anders have objected to my scenario of
the "transcendence" of an SI by takeover of the internet by a program
that uses the computing resources to augmemt its own intelligence.
This scenario occurs in Vinge's "True Names" as the result of a defense
against a perversion attack. One of my underlying assumptions is that
the internet is vulnerable to such an attack, and that the first atacker
will win. It isn't really a war at all. Further, the result may not be