Re: and what if Manhatten is nuked?

From: Anders Sandberg (asa@nada.kth.se)
Date: Sat Aug 17 2002 - 05:49:20 MDT


On Sat, Aug 17, 2002 at 06:25:57PM +1000, Avatar Polymorph wrote:
>
> It is better to encourage mechanisms for protective shielding in the
> fashion of Singularitarian sysop rather than destructive punishment. In
> this context, despite Mutual Assured Destruction arguments, it would be
> better to concentrate on ABM systems and unmanned fighter aircraft along
> with laser systems etc.

Which are useless against suitcase nukes and anthrax. The problem with
"sysop shields" is that it is a single point of failure: somehow the
system has to prevent threats, but if it fails there is no other
defense. I think the interview with Bruce Schneier in The Atlantic
Monthly makes a few very salient points about this:
http://www.theatlantic.com/issues/2002/09/mann.htm
It is better to have systems that fail smartly than foolproof systems.

This is roughly my position on AI (as I stated at Extro 5): better to
have several imperfect layers of protection than spend all effort on one
"perfect" protection.

Deterrence is one such imperfect protection: if you are known to
impassionately wipe out anybody who attacks you (according to clear
rules that you can be observed to obey) the value of an attack becomes
lower (but again, an irrational attacker doesn't care). It will decrease
the likeliehood of one category of attacks, leaving resources for the
other kinds.

> Does anyone think the US has been working on the theory of anti-matter
> bombs? The thought struck me the other day... [Scientists at the European
> Particle Physics Laboratory, on the Franco-Swiss border, have now set up
> the world's first anti-matter factory, one of them, Professor Frank Close,
> told the annual science festival in Wales. Nine atoms of anti-hydrogen were
> produced just over a year ago. Now, the new factory will produce them at a
> rate of more than 2,000 atoms per hour, Close said.... 1998]

I would be surprised if no Sandia or DoE engineers have ever thought
about it, maybe even written some memos about it, but it is unlikely
there has been any serious work. Antimatter is enormously more costly
than plutonium, but the bang is not that different from a weapons point
of view.

> Also, conentrating on forgiveness and fixing up the structural problem is
> always a good idea. Also remembering that all human beings are living,
> breathing, thinking, feeling entities is not bad. Remembering two wrongs
> don't make a right is excellent. Maintaining a higher moral and ethical
> stance when you don't have to is a sure sign of advanced ethical mechanics.

Yes, and also useful realpolitik: if your behavior is ethical according
to some system, then you can be a predictable partner for your friends.
They might even disagree on your ethics, but your consistency makes
building trust easier. An arbitrary and opportunistic person or nation is
hard to trust, and hence will have to pay higher prices.

-- 
-----------------------------------------------------------------------
Anders Sandberg                                      Towards Ascension!
asa@nada.kth.se                            http://www.nada.kth.se/~asa/
GCS/M/S/O d++ -p+ c++++ !l u+ e++ m++ s+/+ n--- h+/* f+ g+ w++ t+ r+ !y


This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 09:16:11 MST