2-D diagrams LO4579 and Knowledge Distribution

GaltJohn22@aol.com
Wed, 3 Jan 1996 11:05:23 -0500

Replying to LO4550 --

In a message dated 96-01-03 00:10:27 EST, you write:

>So my answer is I don't believe it will work in the "real world"... not
>because of the the non-linear, creative aspects of many 'real' tasks...
>but because of the _human_ aspects.

This response is highly typical of the AI "uninitiate" (no insult intended
at all). At to of the "Big Three" auto makers my company makes as much
money from the "human element" (United Auto Workers) *SUGGESTING OUR
SYSTEMS* by brand name to their management for continuous improvement
reasons.

Imagine that, the *workers* recommending real time AI to help them do
their job! Why? Because the systems, mine or others', absorb the
complexity-driven information volume common to the job, and make the
"easy* calls leaving the worker to grow the more creative aspects of
her/his job. Further, since the knowledge bases are cooperating,
multi-disciplinary knowledge holders, when the worker asks for an
explanation the worker gets a *multi-disciplinary, detailed,
situation-specific answer*. This would be called multi-disciplinary
training. The worker, after a short while, understands far more than
before about the *role* of their job in the overall *system* of the
organization.

Most of the *encoding* work is long over. These systems learn from direct
experience, from cases in the past, and from discrete event models that
they use as a "Rule Discovery Playground".

So, none of these objections are based on facts but, rather, on what would
to me appear to be a superficial and dated knowledge of the technology.

--
Hal Popplewell
Chairman, IntelliSys Companies
GaltJohn22@aol.com