>From: Dan Fabulich <daniel.fabulich@yale.edu>
> > >This might not seem like a lot of work, but we don't have a computer
>today
> > >that's capable of detecting anything you list in (2) without a human
> > >pre-analyzing the data.
> >
> > I never said the system should or could be implemented "today." AI
>could
> > very well solve this problem. Even so, I don't think it would really be
> > that difficult to solve even with today's technology. My list of sworn
> > enemies would be finite (quite small actually). You could do it with
>sets
> > of software agents. One agent assigned to follow each of the enemies on
>my
> > list. When any one enemy did one of the trigger events, the agent
>merely
> > reports back. Would that really be so difficult to engineer? I don't
>think
> > so.
>
><sigh> YES. REALLY. Take my word for it for chrissake. :) A list of
>enemies will not suffice: you have to figure out how to write down the
>function that takes grids of numbers and spits out a 1 if it's a picture
>of one of the enemies on your list and a 0 otherwise. This is a very
>complicated function which nobody can write down today.
>
>Using AIs here is the application of a magic genie that will solve all
>your computational problems. A genie like that might not WANT to solve
>all our problems; it might just decide to do its own thing or figure out
>the meaning of life or what have you.
Tell me it is *impossible* and I will shut up. The mere fact that its not
easy does not make a hill of beans difference to me. It wasn't *easy* going
to the moon. If it can be done, at some point or another somebody will throw
enough money and enough programmers at the problem to get it done. The fact
that nobody has done it yet, or that *you* don't see a way to do it does not
matter.
> > I think the problem is that you are thinking of the system as "the
> > computer." Sure it would have to be one *massively* parallel box to
>handle
> > all those requests and all that data. But guess what? Coming soon to a
> > planet under you will be a distributed network of intelligent devices of
>all
> > kinds. Think intelligent, networked, and communicating observation and
> > reporting hardware devices, think innumerable software agents all
>reporting
> > back to you according to your instructions, think distributed networked
> > computing, think "system" as opposed to "the computer" and it should be
> > easier to grasp.
>
>Doesn't matter. Put emphasis on the "massive" and you see what I'm
>getting at.
But even a hugely massive problem becomes manageable when tasks are broken
up and distributed. The bigger the problem, the more hardware and software
you throw at it. If you have every machine on the planet working to break
it there is no such thing yet as an unbreakable code. You have to admit
that at *some* level of computational power the problem can be solved. I
think you'll also have to admit that at some point in the future that level
of computation will be reached.
-Zero
"I like dreams of the future better than the history of the past"
--Thomas Jefferson
______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com
This archive was generated by hypermail 2b29 : Thu Jul 27 2000 - 14:06:46 MDT