Newsgroups: comp.ai.nat-lang
Path: cantaloupe.srv.cs.cmu.edu!nntp.club.cc.cmu.edu!goldenapple.srv.cs.cmu.edu!rochester!cornellcs!newsstand.cit.cornell.edu!portc01.blue.aol.com!newsxfer3.itd.umich.edu!su-news-hub1.bbnplanet.com!cpk-news-hub1.bbnplanet.com!news.bbnplanet.com!ix.netcom.com!i17
From: i17@netcom.com (Writer of Wrongs)
Subject: Missing(er) Links
Message-ID: <i17E7y3vx.6y9@netcom.com>
Organization: Lockness Keyster.
X-Newsreader: TIN [version 1.2 PL2]
Date: Tue, 1 Apr 1997 06:11:57 GMT
Lines: 200
Sender: i17@netcom8.netcom.com


I am struck by the absence of hybrid man-machine systems.

Some say that the Japanese certainly have an edge on Americans
in that they are not averse to working with the most intelligent
machinery they can devise.


In America, it seems that we only get a fleeting contact with the idea
of software which is NOT intended to be stand-alone, but instead is
DESIGNED for one or more "pilots", humans, because humans have some
very unique and powerful abilities.

Why isn't there an explosive investigation of "cybrids"?


Secondly, software which predicted the pilot/operator's moves
would be much like living in a giant walking stilt-spider,
dealing with decisional terrains with powerful ease and undo-ability and
flexible metaphors, freeing the human operator from the constraints
of RealTime and RealPlace.

"Presponsive" software would essentially be flowing ahead of the Operators
down alleys and throughout mazes, and would rarely exhibit any sort
of "latency" or sluggishness.  A very "high-G-force" cockpit indeed,
but rather more like being "on high", so that small "angular" changes
in the data-constructs being manipulated would not be disorienting.
My image here is of holding a globe in the hands at arms length:  a slight
twist takes the beholder from Paris to New York.



Thirdly, multiple simulators creating a sort of "Medusa head" or "Writhing
tree" of viewpoints into and around a situation would be similar to a
one-fingered-creature being given an octopus-like manipulator with
which to handle variations on a situation:  timeline desynchronization,
spreadout of parameters, multiple viewpoints on a single scene, the invention
and (guided) evolution of new senses and new disciplines, as well as new
types of situations.


Five minutes ago I frankensteined the neword "SIMULTIMATION"
to combine the concepts:
{ Simultaneous Simulators Multiple Animation Ultimate Stimulation Estimation }


But the crucial idea here is not trying to displace what Humans can do
(which is miraculous and ever-surprising), but instead liberate Machines
to become what they are capable of becoming, which would be quite limited
if it was always being oriented at the kinds of things Humans do.




Fifth, a generalized Goal-creation and Goal-attainment cybrid system.

A Heuman would be strapped into tight interactivity with a MachNine,
and systems and technology would be developed (by the Pilot, by the Machine,
and by "Outsiders" surrounding this in layers and in communities)
which all is directed at the ATTAINMENT OF GOALS AND THEN MOVING ON.

Any problem in "problem-state-space" would be duly noted and analyzed
and metaphorized, and added in to the Terrain Of Future Achievements.

Problems would be metaphorically broken down and used to solve each other.
Learning how to eat an ice cream cone while on a windy (VR-simulated) ride
in Izme-World in 102 degree heat, would immediately help dissolve the
problem-complex also being tracked by the system (and by the mind of the
Human) involving coping with data-structs which mutate or change quickly.



Sixth.  Language.
Some of us are at ease with the fact of telepathy and shared knowledge
and the integration of Intent and System-Response here in this VR called
Earth 1997.  It still remains problematic even for me, since anything can
deny any particular knowledge, and can deny any particular intent behind
any particular action.  So although one can steer complex cybersystems
with the mind, there is the occasional "mishap" which is blamed on
"noise and error".

Having stated that, what I want to say is that we should EXPECT the
"cybrid" system to display a high degree of "handshake" with the INTENT
of the pilot, and we should NOT require all "learning and accomodation"
to be done via extensive explicit "communication".

Every known useful means to achieve this ("minimal magic", I call it)
should be incorporated in such a system, from deep-layered Neural Nets,
to Schroedinger-Ensembled parallelistic "flocking", to branched contextual
LUTs based on past contexts and micro-contexts...  but above all, the
FACT of telepathic "observer effects" *across time and space distances*
should be a GIVEN, and not ghosted (ignored) by "Occams' Razor" mentalities.

So:  Language.  If we cannot quite get machines out of Searle's Chinese
Room, and prove that "They know what they are talking about", WHO CARES?
My car doesn't know a pothole from a stopsign, a little old lady from a
traffic cop... but when I buckle myself in and do a 1000-mile drive, *I*
have a rather over-intense "meaning" for EVERYTHING that goes on, and
so it doesn't matter that the car has a "hollow mind".

Similarly, my TV and telephone and the paper my news is printed on haven't
the slightest idea what is going on.  And we could chase the angeldemon
of "meaning" all the way up the stairs into my pineal gland and beyond,
and none of this "hollowness" affects my ability to write hilarious netposts
on occasion.



Seventh:  allocation of manpower.

I kind of imagine it like this.  You take anywhere from 1 to 25 "pilots",
and "plum-pudding" them into your immersive "cybrid" system, and let them
start talking about what they would like this new environment to be doing.

Your pilots form multiple overlapping and transitory sub-clusters, and
some pilots tend to be more isolated than others.

"Outside" of this "inner environment", you have what I call a "Hollywood
Surround", which consists of a large community of support crews and "ninjas
in black" who are troubleshooters and fixit-teams, and who communicate
with some of the pilots less directly than with others.

Some of the pilots are on drugs, and have been able to hypnotize themselves
into creating a world inside the System.  Others are Liason Officers and
Point Men who try not to disturb a Pilot's Virtual Reality unless necessary.

The system itself creates Entities who might be very Pilot-like, or might
be more like Corporations, with some degree of Personality or lack thereof.


Eight: Subcombinatoric Isomorphism of Sub-Problems and Sub-Solutions.

(SCIOSPASS??)  In other words:  the System and the Pilots are constantly
re-metaphorizing any new "tool" or new ( constraint + solutions )
satisfaction to be useful globally.

The Global Intent is to create an Ultra-World of deeply "significant"
symbolsystems and formulations, which would tend in large part to be
rather concentrical and hierarchical, as well as decentralized and
pioneering, and also have both highly temporary (fast-evolving) aspects
and majorly-"eternalistic" aspects.

Motivation might be used in some domains of endeavour, where other areas
might be entirely "pressureless".  Games of "rescue the dog" and "heal
the wound" could coexist as focusing techniques with more aetherial
satisfactions like "improve this MonaLisa mutation" or "remix this techno-
trance symphony".  ANY problem is tractable if dealt with through metaphorics
and combinatorics.  A differential equation might be instantly manipulable
once translated into a multiple-version Aesops Fable about "The Bird Who
Couldn't Remember How To Meet Chicks", or maybe the latter problem is easier
to deal with when broken down into overlapping physics problems involving
optimal golfclub selection on damp fairways.


At first, the Cybrid System might be largely text-based, or radio-drama-like,
instead of bogging down in crude cartoon worlds and rubber body suits and
electrodes in the noggin and goggle-gizmos... relying on the Human Pilot's
own imaginations to create entire planetary systems and ringworlds around
Black Holes, or creatures better "suited" to certain problem-domains.

Nanotech, for example, should be deeply, even UTTERLY, *simulated*
LONG BEFORE any fools go and try to subvert the Dominant Paradigm
of What We Already Got Here and "actually" make that stuff.


Appropriate regions of the histories of particular worlds would be used
to deal with one problem-aspect or another.  Coming up with a good recipe
for making rocks edible might be achieved by cave-men better than by
space-men, and even better, by BOTH these approaches in cooperation.
(HA: A big celebration ROCK EATING CONTEST with spacemen and cavemen vying
for Blue Ribbons.  "Best Imitation Apple Pie Using Gravel And Mica Schist".)

The Hollywood Surround would actually prepare these recipes and feed them
to the Pilots, who might imagine that some kitchen gadgetry in their
mobile home did all the work.  But note that the Pilots are left to do
the Tasting, where the Cybrid System might just "talk about it
enthusiastically" and keep track of the Pilots from a behaviourist
standpoint.  "I recommend the Quartz Sundae, sir, it is surprisingly
piquant and flavorful", says a NPC waiter to a Pilot scanning the menu.



IAn. 7:17 31mar97. 187lines 8888chars.
-- 

------------------------------------------------------------------------------

               My god is AMNISCIENT.  He forgets EVERYTHING.

         He even forgot about Me, and I'm his coolest creation ever.

                        Now... What was I saying?

------------------------------------------------------------------------------
22:11 31mar97.
-- 

------------------------------------------------------------------------------

               My god is AMNISCIENT.  He forgets EVERYTHING.

         He even forgot about Me, and I'm his coolest creation ever.

                        Now... What was I saying?

------------------------------------------------------------------------------
