[extropy-chat] pest-devouring automaton

[Adrian Tymes]

--- Eugen Leitl <eugen at leitl.org> wrote:
> On Fri, Dec 31, 2004 at 07:23:25PM -0800, Adrian
> Tymes wrote:
> > As has been pointed out, these robots don't hunt -
> 
> What, they can't hunt? A remote exploit there, a 
> code patch here, here ya go. 

One needs certain hardware to be a good hunter.  If
you're a 10 cm wide toy robot on 1 kmph wheels with no
manipulators, it doesn't matter what software you run.

Which doesn't mean that there can't be robots with the
requisite hardware.  But even human intelligence can
see the potential for that, and is likely to guard
against it when designing robots that could hunt well.

> > and besides, their food sources aren't quite the
> same
> > as ours.
> 
> Oh yeah, that makes me feel completely safe already.

It should.  When was the last time, outside of
fiction, you saw a cow (or any other herbivore) eat a
person?  Or, for a closer analogy, a vulture (or any
other carrion feeder) eat a living person?  (They may
circle, but they stay away until you're dead.)

> > But putting those aside for the moment...if they
> did
> > evolve to "compete" with us for food, then most
> likely
> 
> How about you being their food?

Self-optimization would argue against that.  Humans
don't make good food animals, compared to the food
animals that humans have bred.  Any robot running its
own directives enough to want to eat humans would
realize that.  Any other robot wanting to eat humans
would be doing so at the directive of other humans -
and then it just becomes a new way for people to kill
people.  We've plenty of those, yet we're still around
today.

> > what would happen is the same thing that's
> happened as
> > more and more humans have "competed" for food over
> 
> But robots are not animals.

Do you mean to emphasize that they would not react
like animals competing for food?  True - but,
likewise, the ability to meet needs with technological
development has been one of the things setting mankind
apart from other animals.