RE: "Heart"

["Brandon J. Van Every" <vanevery@3DProgrammer.com>]

> Anyway, that's how I see the issue of "heart".  I'm not saying it's
> impossible for a computer to express "heart", by any means.. but I do
> think that the issues involved and so complex.. and so far away from
> formal understanding, that, as of yet, only people can even dream of
> doing it.

Pretty much my understanding of it as well.  I agree with your basic
framework of intuition as computation.  I would only add that there are some
problems which are not impossible to address by formal means, but it's
expedient to use artistic intuition to solve them anyways.  Do I want to
spend 5 years developing a formal technology for 1 minor area of authorship
that a skilled human being can achieve in 5 minutes?  What about all those
other minor areas of authorship that will also need 5 years of brow
scratching to come up with decent results for?  Considering how short life
is, I think it's a waste of time.  I could produce 100 times more work the
old fashioned way.  Even if you grant that the skilled human being will have
to take 2 years minimum to learn their skill, the technologist is going to
have to take those 2 years as well, plus whatever is needed to implement the
technology on top of that.  2 years vs. 7 years.

Also there's the problem of selection.  To evaluate artwork, one has to
actually view it.  In this way art is selected for its relevance to life
experience.  Let's say someone creates this great automated system that can
spew infinite stories until the cows come home.  Who cares?  Nobody is going
to have time to read all those stories.  Moreover, they will have no import,
as they did not come from a human being organizing specific life experience.
The machine will inevitably become a mirror of the viewer.  We already have
many such mirrors, at a much lower tech level, so why do we need more of
them?  Consider all the books gathering dust in libraries around the world.
Consider all the paperbacks being returned to their publishers for lack of
readership.  We have a billion stories already, we could select endlessly
from what we already have.

Thus, I think we could stand some cold calculation as to why we pursue the
automated storytelling objectives.  Are we trying to do something practical,
like write a story more quickly?  Or create a massively multiuser online
service that tells acceptable stories?  Aren't both of these things more
about business model, productivity rates, and techno-kewl than about Art?
Are we pursuing automation in the name of pure research, in the name of
Science?  Fine, but Science is not Art.  Art can be made out of sticks of
wood, it doesn't need a new technology.

Do we envision a technology so radical that it changes the very substrate of
our perception and existence?  If so, I think we need to contemplate
technologies much more radical than any possible variant of Interactive
Fiction.  You'd have to start toying with reality itself and try to envision
the consequences.  Brain plugins, drugs, dismemberment, rebirth from death,
matter/energy conversion, the existence of a soul or lack thereof, communal
consciousness.  We can write about these things now, but we cannot really
know what it would be like to live in a universe where such concepts are
continuously applied.

Anyways, in the absence of a compelling justification for super-automation,
I think the goal should be to provide intuitive human authors with good
tools.  We shouldn't try to replace human intuition, we should try to
leverage and extend it, in terms of sheer cold hearted calculating
productivity.  It's a better tool if the job gets done in 5 hours rather
than 8.


Cheers,                                     www.3DProgrammer.com
Brandon Van Every                           Seattle, WA

For plot and pace, writers use words; game designers use numbers.
Anything understood over time has plot and pace.