[MUD-Dev] Virtual worlds as a Society of Mind

Sean Kelly sean at ffwd.cx
Tue Dec 3 08:45:55 CET 2002


Frank Crowell wrote:

> I don't drink or do drugs, I just write like I do.  There is a
> meme in the techno world now to "embed everything" and to "go
> wireless everywhere".  But as I was thinking of alternative to
> FEAR I could see an approach to embed everything with smarts.  So
> instead of a singular brain, you have an augmented brain. I may be
> making this stuff up as I go along, but I believe there is
> material on the Net that suggests creating objects that can
> broadcast their properties.  On paper the idea of smart cars and
> roads have been around.  The designs were based on devices
> broadcasting to cars, such as a road marker saying "slow down
> accident 1.5 miles ahead".  Or the wheel device saying, "chuck
> hole ahead, go left, no no i mean your other left.  argh we are
> dead."  Ok that didn't work out too good.

There is a group of college kids who built a VR interface for
Quake. You wear stereoscopic goggles and carry around some sort of
gun, a wearable computer and integrated GPS system.  The level model
is an exact reproduction of their college campus.  So playing the
game comprises running around and shooting at monsters in the
virtual world using a gun you hold in your hand.  This was covered
on Slashdot about 6 months ago.

The project got me thinking of the potential uses of such
technology, and I came to pretty much the same conclusion that you
did.  Since wireless is becoming all the rage, why not have things
offer information about themselves?  There could be neighborhood
maps, buildings could broadcast building maps and info, etc.  Then a
robot could nearly eliminate complex pathing and calculate its
pathing inside a 3D model of the area it's in.  A HUD projected onto
a car winshield could change colors based on nearby lights, arrows
could be projected when turn lanes were available, robot buses could
navigate preset paths and only have to worry about collision
detection.  The only major downside is it gives a would-be terrorist
the potential to cause massive damage by hacking infrastructure
servers.

> Someone coined "social computing" for when humans are used in a
> computer environment to do a simple memory lookup or to solve a
> problem -- simple or complex.  I don't believe I made up "social
> computing".

There are entire fields of AI that call themselves "social
computing." The most interesting to me is swarm intelligence,
partially because the algorithm is so darn simple.  The conclusion
from such groups is that monolithic AIs complexity outstrips their
skill, and that you can do a lot better modeling some sort of
collaborative comunity.  The general idea is that every knowledge
system can be modeled in an N-dimensional space and solving a
problem is a matter of navigating that space.  If you've got a bunch
of AIs that occupy different points in that space, then they can
communicate with one another and give each other information about
which areas contain more optimal solutions.


_______________________________________________
MUD-Dev mailing list
MUD-Dev at kanga.nu
https://www.kanga.nu/lists/listinfo/mud-dev



More information about the mud-dev-archive mailing list