[ExI] Unfrendly AI is a mistaken idea.

Lee Corbin lcorbin at rawbw.com
Thu May 31 00:52:57 UTC 2007


Stathis writes

> > I think that men did try to have as many children as possible, for a while, until
> > women wised up. (I mean that quite seriously;  it's a theory of evolutionary
> > history that an "arms-race" developed between women who can have only 
> > relatively few children, and men who can have many).
> 
> With modern reproductive technologies, it would be possible for women in
> power to have thousands of children, and men to have many thousands,
> perhaps even millions.

Sounds great!  I've always believed that as many people should be saved from
non-existence as possible.

> And that's without even considering the possibilities raised by cloning.
> Therefore, given enough time, a human will arise who will take advantage
> of these opportunities, and that human will dominate the world.

Well, it won't be just *one* such person who will dominate the world.
Gradually, those who have more children (viable offspring) will come
to be highly represented, and those who don't will gradually go 
extinct in the sense that there will be fewer and fewer people like
them.

>  Do you think this is going to happen soon?

Oh, a singularity or a Moslem dictatorship or Fall of Civilzation or
a thousand other things could occur first. Two of the most interesting
things that will occur first are Age Reversal and Genetic Engineering.

Still, right now the environment favors---in strictly population biologic
terms---those who have as many children as they can and who allow
the state to raise them (e.g., they pay exactly zero for medical care,
relying only on ER services).  That's the path of the immediate future.

So, keeping a Darwinian view in mind...

> If not, why assume it is any more likely that AI's will take over the
> world, especially in view of the fact that they won't start out with
> the desire to survive and reproduce which is basic to every naturally
> evolved organism? 

Because one way or another, one or some of them will stumble upon
the behavior of trying to take over.  One way it could happen is that
a human participating in an early AI design will want "his baby" to
forestall other developments, or to be supreme.  Another way is
that an AI breeding experiment will produce a "winner" who'll just
"naturally" want to dominate everyone and everything, just the
way a typical two-year-old does.

Lee




More information about the extropy-chat mailing list