From: Eliezer S. Yudkowsky (sentience@pobox.com)
Date: Tue Jul 10 2001 - 18:12:11 MDT
Reason wrote:
>
> Eliezer S. Yudkowsky wrote:
> >
> > How old am I? Old enough to vote. Anything else is on a need-to-know
> > basis, and you don't need to know. But to my mind, the worry is whether
> > my *grandparents* make it to the Singularity. I seriously doubt that even
> > the old fogeys among us will miss the train. If you're under forty, you
> > essentially have a cast-iron guarantee that you won't die of old age. You
> > may be hit by a truck, die of a premature heart attack at fifty, or have
> > an artery eaten by goo, but you won't die of old age. Hope this cheers
> > y'all up.
>
> I really don't see this; really, really not. It just doesn't sound plausible
> to be so absolute about this; it begins to seem like an issue of faith in
> technology, in everything working out ok. Consider even just the trial and
> error is-this-working timescales...
I also tend to be very nervous about making possibly mistaken
predictions. But the above reflects my current belief that the idea of
the Singularity taking until 2030 is simply too far out for me to assign
it any real probability. If something unexpected leaves us out in the
cold, it'll mean a Singularity delayed until 2020. Not 2030. There's
simply too many hammer-blows due to hit our world in the next ten years,
never mind the next twenty. They may kill us. They may save us. But
it'll be all over, one way or the other, before 2030.
It's not a question of "I predict all these totally different new
technologies will be created sometime in the next fifty years, based on an
extrapolation of exponential trends." Predictions like that sometimes do
go wrong. But when I put my ear to the ground, right now, what I hear is
a world that is already trembling. My feel for these technologies says
that almost all of the race has already been run. The human condition is
drawing to a close.
If I turn out to be wrong, if there's much more distance left to be
covered or if those last few yards are covered very slowly, then that
would mean delaying until 2020, instead of everything being settled over
the next decade. That's a real possibility. But to suppose that things
will not have been settled by 2030 is such a low probability that I don't
feel any worry at all over using the phrase "cast-iron guarantee". I find
myself completely lacking in any emotional anxiety that I will look like
an idiot in thirty years.
> Anyone want to roll out some numerical- and factual- based arguments on one
> side or the other?
What good would those do? If there's enough slack in the numbers and
structure that you can "prove" anything you like by picking your
arguments, then mathematical projection is the wrong way to make
predictions for that question.
-- -- -- -- --
Eliezer S. Yudkowsky http://singinst.org/
Research Fellow, Singularity Institute for Artificial Intelligence
This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 08:08:39 MST