Selected Newsgroup Message
From: "Sergio Navega" <snavega@ibm.net>
Subject: Re: Letter to Another Editor
Date: 01 Apr 1999 00:00:00 GMT
Message-ID: <3703e9b3@news3.us.ibm.net>
References: <3703753f@news3.us.ibm.net>
<19990401113324.19774.00001000@ng33.aol.com>
X-Notice: should be reported to postmaster@ibm.net
X-MimeOLE: Produced By Microsoft MimeOLE V4.71.1712.3
X-Complaints-To: postmaster@ibm.net
X-Trace: 1 Apr 1999 21:48:35 GMT, 166.72.29.98
Organization: SilWis
Newsgroups: comp.ai.philosophy
MMS1979 wrote in message <19990401113324.19774.00001000@ng33.aol.com>...
>> You're assessing
>>not the "humanity" of that machine, but the fact that it responds
>>*intelligently* to your queries. This means that we should forget
>>about "human-like intelligence" and think only about
"intelligence".
>
>
>Dear Sir... If you should forget abour "human-like intelligence" then why
>bother with the question of strong AI in the first place?
>In that case, you would belive in McCarthy's thinking thermometer for
example..
>A thinking thermometer? you might ask... yes, since it is correctly stating
the
>temperature, then that must mean it *believes* the room temperature to be
what
>it is... and I find that line of thinking absurd.
McCarthy's thermometer is not intelligent because it can't grasp anything
from its environment and use that information to *perform better*.
Wanna see what I call an intelligent thermometer in action?
How about the one that perceives the setting its "owner" does in relation
to external temperature? Say that it is winter, and the thermometer
perceives that its master adjusts the temperature so that it stays
around 19 C (centigrades), when the the external temperature is
around 10 C. Now, during summer, the thermometer will perceive that
its master is calibrating its knob to maintain the temperature
around 24 C, when external temperature is around 30. The intelligent
thermometer will "conclude" that it could do better if it adjusts
automatically its temperature according to the external temperature,
in a way that seems reasonable, given its previous "experiences".
It will "guess" its owner's desire and will act "intelligently".
An intelligent thermometer will not only solve this case but, if
*the same equipment* is installed in the house of another man who
wants temperatures *always* around 20 C, it will recognize that
in this case its adjustment should be different. It will act
differently, according to another set of experiences. It is the
very same hardware, but with different behaviors, given different
"living" situations. That's perception, the first step toward
acting intelligently.
Now think if we apply this same strategy in televisions, refrigerators,
lights, doors, telephones (refusing to sound loud if rang at 3am,
unless it "perceives" that it should be an "emergency", because
of previous situations), etc, etc.
>We must set a standard for determining strong AI, if we use the word
>*intelligence* too looslely, then almost everything would have some sort of
>belief...(e.g. a glass believes it must hold the liquid!) how ridiculous!
But a glass that could intelligently change its transparency according
to the liquid it carries can be the ultimate gift for the christmas
of 2030's. I would buy one.
Regards,
Sergio Navega.
From: "Sergio Navega" <snavega@ibm.net>
Subject: Re: Letter to Another Editor
Date: 02 Apr 1999 00:00:00 GMT
Message-ID: <3704ce69@news3.us.ibm.net>
References: <3703753f@news3.us.ibm.net>
<19990401113324.19774.00001000@ng33.aol.com> <3703e9b3@news3.us.ibm.net>
<7e27vd$b8b@idefix.eunet.fi>
X-Notice: should be reported to postmaster@ibm.net
X-MimeOLE: Produced By Microsoft MimeOLE V4.71.1712.3
X-Complaints-To: postmaster@ibm.net
X-Trace: 2 Apr 1999 14:04:25 GMT, 166.72.21.55
Organization: SilWis
Newsgroups: comp.ai.philosophy
Vesa Monisto wrote in message <7e27vd$b8b@idefix.eunet.fi>...
>MMS1979 wrote:
>
>> How about the one that perceives the setting its "owner" does in
relation
>> to external temperature? Say that it is winter, and the thermometer
>> perceives that its master adjusts the temperature so that it stays
>> around 19 C (centigrades), when the the external temperature is
>> around 10 C. Now, during summer, the thermometer will perceive that
>> its master is calibrating its knob to maintain the temperature
>> around 24 C, when external temperature is around 30. The intelligent
>> thermometer will "conclude" that it could do better if it adjusts
>> automatically its temperature according to the external temperature,
>> in a way that seems reasonable, given its previous "experiences".
>> It will "guess" its owner's desire and will act
"intelligently".
>
>That is exactly what my integrating 'thermometer' ("Tassu") does.
>It "sees" the trends in curves of parameter-values (in its memory) and
>adapts now to future states of weather (seen in trends). I have
>sensors and adjusting potentiometers in every room and so the heating
>and temperature is what I want it to be (e.g., lower in bedroom).
>It lowers the temperature in working rooms on Saturdays/Sundays for
>to save electricity, etc.. "Tassu" is not an exception, it's a 'must'.
>
Ha! That's it. I was hoping that such a thing could exist in the future,
but now I know that some clever engineers (VM included!) devised
such a thing. It will be easier for me, from now on, to exemplify
what I think is a simple kind of intelligence. Vesa, is there any
reference on the web about such device? You may do any merchandizing,
I wouldn't mind :-) I could use it as reference in a paper I'll write
soon.
>To say that "it concludes and guesses" is animation, of course,
>therefore your quotation marks, I guess. This animation (transference
>of agenthood) might be a problem. There are no ghosts in my thermometer
>(I know because I was in the team which made it), but are fishers
>fishing with or *by* nets. Are nets fishing for fishers even when
>fishers are just sitting and waiting in their armchairs?
>
Yes, I used quotation marks because I wanted to specify that
"acts intelligently" was a little bit different than the traditional,
"human" view. But I really consider this as an example of a lower level
intelligence (whatever that might mean).
>There are lots of natural phenomena which can be 'captured' by human
>facilities to make work/gain 'autonomously', 'games giving gains', even
>I'm just sitting in my chair.
>
>>...
>> But a glass that could intelligently change its transparency according
>> to the liquid it carries can be the ultimate gift for the christmas
>> of 2030's. I would buy one.
>
>So would I. - I actually bought some months ago a new welding helmet,
>which has a glass (4x9 cm) with liquid crystals. It is transparent but
>some microseconds after the arc flashes the glass is nearly opaque.
>
>The glass "feels" ultraviolet radiation, "guesses" that I want to
protect
>my eyes, "concludes/knows" that this can be done by opaqueness, and
>"shows its consciousness and intelligence" ...
>
>Absurd! ... but isn't 'absurdism' a style of art, a language game?
>(Epistemic games are at least short, simple and pragmatic ones.
>What scientists/engineers had to do *actually*, that's boring ... :)
>
Your welding helmet actually is another example of intelligent device,
although due to its purely reactive behavior, I would classify it as
less intelligent than the thermometer.
Regards,
Sergio Navega.
Back to Menu of Messages Sergio's Homepage
Any Comments? snavega@attglobal.net