Re: False Knowledge

"Eray Ozkural exa" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]
> Jim Balter <[EMAIL PROTECTED]> wrote in message
news:<[EMAIL PROTECTED]>...
> > Eray Ozkural exa wrote:
> > > "Dave Ulmer" <[EMAIL PROTECTED]> wrote in message
news:<[EMAIL PROTECTED]>...
> > >
> > >>Making a computer behave like a human is impossibly complex. Let's say
you
> > >>have a computer running 12 simultaneous applications with 12 open
windows on
> > >>the screen. Now do you think you could figure out its entire software
> > >>structure by simply observing the screen display? What if in the
middle of
> > >>your study of the screen one application shuts down and a new one
starts up?
> > >>If you think you would ever be able to duplicate is operation, you
would be
> > >>wasting your time.
> > >
> > >
> > > That's a very bad analogy for AI research!
> >
> > Perhaps, but not for cognitive research.  Many people
> > expect to be able to figure out "how the brain works"
> > or "what the brain's software (or program) is" -- I'm not
> > at all sure that it's a feasible reverse engineering task.
>
> I think I'm called to defend my statement!
>
> It's a bad analogy not because it says Cog.Sci. is about figuring out
> "how the brain works", but because it claims that we try to do this
> *only* *by* *reference* to observation of extra-neural behavior which
> is not correct!!!!!
>
> There is something called neuroscience that aims to reverse engineer
> the design by other means.
>
> OTOH, AI research tries to figure out how "intelligent systems work",
> so I don't mean something like we're trying to build an
> everything-is-vibrations i-exist-on-a-higher-plane kind of new age
> fluff. (that description is due to Dan  IIRC)
>
> I am simply pointing out that the analogy Dave gives does not show us
> that Cog.Sci. is hopeless because it has a certain constrained
> research methodology. The truth is far from that.
>
> > A Martian who knew everything about physics but nothing
> > about computer technology or the history of our computer
> > design, who was given a computer running a some complex program
> > would, I suspect, be at a total loss to figure out what the
> > program was, and perhaps even that there was a notion of "program"
> > involved.
>
> You may be right about Martians, how can we expect them to understand
> a program if they know nothing about computer science?

Yet the rules of physics execute on processes in Universe. It isnot a
stretch to imagine another society with knowldge of physics to realize
physics can be looked at as Steven Weinberg does.

>From ther - the concept of a program is close.

Note that if Martians knew everything about physics, that also strongly
implies they are capable of learning.  Learning how some machine processes
inputs and delivers outputs (in the parlance of physics of course), wuld be
trivial for them, and thence could easily lead them into learning new
concepts about manipulation of matter/energy in "programming-like" concepts.



>
> Cheers,
>
> --
> Eray