[ExI] online resources for identifying symbols
Anders Sandberg
anders at aleph.se
Thu Dec 12 12:46:17 UTC 2013
On 2013-12-12 06:48, spike wrote:
>
>
> This is analogous to computer chess in a way. A long time ago,
> computers could play a good game at the really fast speeds and in open
> positions. So a human could beat a computer for a long time if the
> human would play into super-closed positions with lots of blocked
> pawns and long pawn chains. So computers gradually got better, but
> for a long time, you could beat one if you knew how to play its
> weaknesses, but if it managed to tear open the position, forget it,
> you might as well climb in the coffin and pull the lid shut behind you.
>
This is exactly what Kasparov said too. So after Deep Blue he very much
took a "if you can't beat em', join em'" view and started freeform
chess. He is currently interested in constructing smarter ways of
pooling human and machine intelligence for solving important problems
(like political decisionmaking).
> This whole thing again has me thinking about singularity scenarios
> where software surpasses human abilities in a lot of areas but the big
> S Singularity still doesn't happen.
>
As Damien wrote, besides the Spike there is the Swell. We might get a
very broad singularity.
If you can use the software to amplify your capabilities or a group can
do it, then you might get a big S singularity even without AI. Many
technologies seem to be mutually amplifying: if you can build nanotech
you can make a lot of computer power to brute-force some classes of
problem, with lower-case ai you can do a lot of automated engineering
and research, a bit of intelligence amplification may allow better
strategizing, and so on. It will be fun to see how it works out.
--
Dr Anders Sandberg
Future of Humanity Institute
Oxford Martin School
Oxford University
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20131212/9ff557ce/attachment.html>
More information about the extropy-chat
mailing list