From: Reason (reason@exratio.com)
Date: Tue Apr 30 2002 - 01:20:28 MDT
--> Samantha Atkins
> Eliezer S. Yudkowsky wrote:
>
> > Hal Finney wrote:
> >
> >>I was not present, but it seems to me that Kurzweil has the stronger
> >>argument, that machine technology will continue to advance much faster
> >>than biology can.
> >>
> >
> > You would have thought so. However, Ray Kurzweil ended up arguing that
> > humans and machines would be integrated throughout and AI would
> never pull
> > ahead, while Gregory Stock much more sensibly argued that pure AIs would
> > have a tremendous advantage over biology. The basis for Gregory Stock's
> > argument for a human future is that machines will pull ahead so
> fast as to
> > leave us to our own resources, and that uploading and brain-computer
> > interfacing is a problem so hard as to be infeasible, especially by
> > comparison with machine intelligence - thus leaving *us* with biology.
> >
> It is an interesting twiste in the debate when both agree that
> the AIs will march forward regardless and the debate turns on
> wheter we will at least get augmentation out of it.
One of my game settings ("Artilect Earth") postulated a future in which the
more-or-less accidentally created self-improving uberAIs had their
singularity in nothing flat and got on with their lives, leaving humanity to
make sense of it all and have a more "traditional" scifi future (think
Transhuman Space) as rats in the metaphorical walls. Very large walls, of
course, the whole of potential human experience. If you're in the walls,
you've no idea what the rooms even are. A future in which continential North
America is computronium, the solar system is seeded with trillions of tiny
computronium satellites, Jupiter is being dismantled, the AIs are sending
out von neumann nanoprobes, and human-type intelligences have yet to
reliably communicate with the self-improving AIs due to the differences in,
well, everything important.
The Greg Bear moment for that setting was that the artilects found the
Solipsis, intelligence encoded in the vacuum that was as far beyond them as
they were beyond humanity, and just as different, alien, and
incomprehensible. But I digress.
Anyhow, thought experiment:
1) nanotech is available
2) you are uploaded, with the potential to experiment and self-improve
3) do you hang around for the laggards or see how far the highway goes?
One could argue that all it takes is one person to decide to fill any given
role, from go for broke subliming through to transition guide. Substitute AI
created by person for specific task if necessary. Once we've figured it out,
all possibilities become possible -- I don't think there's any real first
mover advantage to transcendence.
Reason
http://www.exratio.com
This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 09:13:43 MST