Re: [silk] More from Bill Joy

[Eugene.Leitl@xxxxxxxxxxxxxxxxxxx]

Arsalan Zaidi wrote:

> Slightly OT. I sometimes wonder at the excitement the concept of
> self-replicating nano machines generates. Wouldn't it be far easier to use
> existing micro-organism to do the work you want? Better to use existing
> 'code' rather than re-inventing the wheel.

You may observe that biology did not invent the macroscopic wheel, nor 
the electromotor, nor the plasma drive, nor the atmospheric reentry heat 
shield, the superconductor, radio communication nor the terawatt pulse 
laser. Critters do not operate in a vacuum; they're a part of the 
biosphere and hence subject to its usual direct or indirect feedback and 
regulations mechanisms, as well as design limitations. They can't cross
extended patches of bad fitness landscape in person, and mutation-driven
tunnelling is highly limited in extent. The reason we're wrecking so 
much without even really trying is because we've stepped outside of the 
domain of biology with our technology. Nothing can touch us but ourselves.

This does not apply to an artificial design, using utterly alien means of
an omnivorous, encapsulated metabolism, methods of replication and usage of
structural materials. Gray goo can literally eat the entire biosphere for 
breakfast, in the space of a few scant weeks, and nothing can check it it 
but ubiquitous presence of an equally micron-sized countermeasure -- nevermind
that the conflict itself would kill you as surely. In case it's feasible
it's really really bad shit, a 24 ct genuine Ragnarok event. You'd hardly
fare worse if two solitary neutron star wanderers collided and fused a 
megameter over your head. FLASH! -- you're history. Only, in this case
you're slowly, painfully lysed into nanoslime instead of being suddenly 
ionized by the gamma flash. Be afraid. Be really really afraid.
 
> Why bother?  An engineered organism and one generated using a more brute
> force technque will both kill your target. Of course, you'll have to do more
> work if you want it to do something specifically nasty.

The raw kill efficiency of the agent is irrelevant for a successful 
bioterrorism act. Take Ebola: it hits its host so fast you'll get 
terminally sick by the end of a domestic flight, while you board it 
with no visible symptoms. People naturally avoid you if you look sick 
(haemorrhagic viruses will make you look very sick indeed very soon), 
and you'll be almost instantly quarantined. Here's a self-terminating 
transmission chain for you.

A successfull terror act will try to contaminate as many people as possible 
with a high-latency high-infectivity high-mortality agent. Think of the 
properties of HIV and influenza or rhinoviruses combined. It will take a 
decade before you even realize you're sick. By the time you're doomed, and 
have ten thousands of random people infected. Because there are rare instances 
of natural immunity, you would want to create some ten independent strains of 
pathogens with above properties, to that the overall kill is quantitative, i.e. 
well in excess of 99.9% of world population. For practical reasons, you will 
probable set the kill timeout (delayed expression of a neurotoxin, apoptosis 
or autoimmune trigger) for a year or two, and limit the mutation hot spots to 
exclude the timer and the kill genes. You'd probably want to make the delay 
intrinsic to the host using an emergent mechanism, so that it will be immune 
to mutation altogether.
 
> Maybe it's doable in the some parts of the world. Here the health services
> have a tough time controling age old cureable and preventable diseases.

True. Part of it is lack of infrastructure, part of it is supportive
climatic environment, and presence of sufficiently rich number of
alternative hosts. It really helps to live in an infrastructurally
advanced arctic or desert society (move to Canada or Arizona, folks 
;), and keep good control of cohabitants.
 
> Another great vector is the rat (or the parasites it carries). They're
> present in just about every major city and are impossible to eradicate.

Good example. I would also target dogs, cats, mice and cockroaches (ravens,
foxes, and a few more exotic species in other parts of the world), apart 
from rats.

> Besides, if you use something like bubonic plague then the more rats you
> kill, the greater the number of homeless fleas and the greater the risk of
> infection!

What we need is a variant of hoof-mouth which also targets people. Maybe
something like this already exists in some critter in some rapidly eroding
jungle pocket somewhere. With some luck we'll find it in time so it can
jump hosts.
 
> Acutally, if you're looking to knock out a city, but don't want the disease
> to hang around forever, something like Ebola might do the trick. It spreads
> fast and kills quickly. Too quickly in fact, making it self-limiting. If you
> want to kill large portions of the world population, something like HIV
> (only airborne!), might be more useful.

People who want to kill using fundamentally uncontainable and unconstrainable
pathogens are psychopaths, or people engaged in acts of postmortem retaliation 
(and haven't heard of Dr. Strangelove). They probably wouldn't want to limit 
themselves to a city or even a continent. As you mentioned, Ebola is a particularly 
poor choice, and can be only spread in a kind of a viral firestorm, requiring a 
critical host density and interaction rate.
 
> What's a pathogen umbrella. Sounds fashinable! :-^

A neologism, a set of blanketing countermeasures against acts of 
bioterrorism and flareups of highly lethal natural pathogens. We don't
have anything like that in place yet. (Wink, wink, nudge, nudge).
 
> What kind of 'quite, pragmatic solutions' could there possibly be?

In case of bioterrorism: A network of biological sniffers in urban areas 
with wireless realtime reporting. Blanket PCR screen for nonhuman DNA of 
all blood and tissue samples submitted. Infrastructure to rapidly order 
and enforce a stringent curfew and quarantine cordoning. Ubiquitous 
presence of personal sterile seals and training in their use. And quite 
a few other things I forgot.

It's doable technically, and not even prohibitively expensive. It's a 
political and social problem, mostly.