Re: Kurzweil's new Singularity/AI page

From: Eugene Leitl (Eugene.Leitl@lrz.uni-muenchen.de)
Date: Thu Mar 15 2001 - 03:17:09 MST


On Wed, 14 Mar 2001, Robert J. Bradbury wrote:

> no way you can "transmit" 10^40 bits using even gamma ray lasers
> in the age of the universe. The further away you go from your

Why would you want to transmit that much? Oh, Hollywood.

> information source, the greater it costs you to retain access
> to your "memory" (because you lose the photons as they spread
> in space due to the inverse square law). As you move further
> and further away your energy limits means you become "less"
> ultra-intelligent.

I dunno, all solar systems are roughly alike. A seed does not to be
ultraintelligent, in fact it doesn't need to be intelligent at all. It
will absolutely suffice for it to generate intelligence, using local
resources. I'm not sure how small the seed can get, but given enough time
even the meekest self-replicator will do.

> The only way you send information is by encoding it in matter
> and send the matter. You can't send matter at the "speed-of-light"
> because its infinitely expensive to send it at that velocity.

So you don't send it at 0.9999999 c but at 0.95 c, or something.

> You also can't send matter at a speed even close to the speed
> of light because as Spike has pointed out hitting the
> interstellar hydrogen ablates your data "package" so by the time
> it gets to the destination there isn't any information left.

I don't think that's right. If you go fast enough, it only means that
you're bathing in radiation, the harder, the faster your are. This means
you have to have a metabolism, and repair the damage incrementally using
distributed encoding (within the packet, and between the packets).

> The rest of the discussion just gets worse...
>
> This points out that Kurzweil hasn't been reading our archives
> or doing proper net searches (otherwise he certainly would have
> run accross the discussions about the Matrioshka Brains). The
> part that doesn't make any sense is that he was at the start
> of my discussion at the Spring '99 Foresight SA meeting where
> I discussed the fundamental limits on intelligence. So he
> *should* be aware of the concepts and not be going off into
> flights of fantasy regarding worm-hole travel (No MAGIC PHYSICS!).

Why do you expect Kurzweil to be omniscient, or only rational?

> As I believe I've pointed out several times in various forums,
> advanced civilizations don't want to expand their size scales
> they want to shrink them. Moravec's muon-onium or Ander's

Sure they want to go compact, but this doesn't mean they're not going to
spawn off clones (which in their turn will go compact, if that's
possible).

> Neutronium are the way to go (if they can be done without
> magic physics).



This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 08:06:22 MST