From: Rafal Smigrodzki (rafal@smigrodzki.org)
Date: Tue Jul 22 2003 - 14:30:59 MDT
Bryan wrote:
> Rafal Smigrodzki wrote:
>
>>> The idea that we can simply sit down and program "thought" is
>>> absurd.
>>
>> To make such categorical pronouncements, one has to be able to
>> explain the meaning of terms used. So what is this "thought" that
>> you can't program?
>>
>> Be as specific as you can.
>
> That is actually the point I'm making: What is this "thought" these
> AI researchers are programming? They don't know, that's why it's
> absurd.
>
### AFAIK, the AI researchers who are actually writing code are currently
working mainly on the lower levels of the information processing hierarchy.
They code perception, parsing of data, simple inferences (simple in
comparison to the calculations in our brains), the levels that form the
basis for conscious deliberation which we sometimes call thought. Of course,
some of them do know what they are programming, and you can tell that by
their results (visual recognition, motor programming for robots, some
results from the Cyc). We do not know what is thought at higher levels yet -
but we will, and then we will program it, perhaps using the lower level
routines developed today.
That we don't know the technical meaning of thought means we can't make
immodestly confident pronouncements about future prospects, we can only use
less reliable means to advance conjectures. For example, since for us
materialists "thought" is just a function of certain objects, and it does
exist, it's reasonable to claim that "thought" can be replicated in other,
similar objects, unless there are some special considerations or features
that would prevent it. What special considerations are you claiming in
support of your "it's absurd" thesis?
Rafal
This archive was generated by hypermail 2.1.5 : Tue Jul 22 2003 - 11:38:00 MDT