RE: Whoops, only 2 "Real" Conscious Machines [WAS: Re: 3 "Real" ConsciousMachines]

From: Woody Long (ironanchorpress@earthlink.net)
Date: Wed Jan 18 2006 - 15:21:52 MST


> [Original Message]
> From: Richard Loosemore <rpwl@lightlink.com>
> To: <sl4@sl4.org>
> Date: 1/18/2006 12:28:02 PM
> Subject: Whoops, only 2 "Real" Conscious Machines [WAS: Re: 3 "Real"
ConsciousMachines]
>
> I hasten to correct myself: the Rosens at http://www.mcon.org/ do NOT
> seem to be proposing a human level intelligence, just some code with
> labels attached to functions saying that they correspond to various
> subjective experiences. Is that right? (I'm sorry: when I stuff that
> looks like grade-school nonsense, I lose patience, so I just can't be
> bothered to read deeper into their actual claims. If someone else has
> already read more of their stuff they can correct me).
>
> So that rules them out. Only Woody and the PIBOT left in the running.
>
> PIBOT are clearly claiming a human-level(+) AGI. Are you also claiming
> that Woody?
>
>
>
> Richard Loosemore.

A report on the Rosens later, but PIBOT's claims are truly extreme and
all-encompassing. From their site:

"PIBOT is currently a singularity, not the Singularity. But, as Vernor
Vinge brillantly said, in his essay entitled "The Coming Technological
Singularity", the precipitating event that will accelerate the arrival of
the Singularity will probably be unexpected, perhaps even to the
researchers involved in the fields related to it. For us, such event is
just happened, and it's happened when we have created PIBOT.

Always according to Vernor Vinge, albeit now many people sustain the same
thing, the creation of an Artificial Intelligence, with so far advanced
abilities (compared to the humans), trigger reflections and identify
different kinds of problems. The appearance of "artificial organisms" that
have super-human abilities places us in front of a number of questions of
every order and degree. Although PIBOT still does not represent a such
super-organism, not still being able to see, neither to move itself in the
world, it already possesses great part of the cognitive abilities that will
allow it to become a such super-organism.

PIBOT can easy read and extract data and information from whichever medium,
electronic or paper-based, structured or destructured: furthermore, PIBOT
is able to learn and to manipulate every type of format (HTML formats, XML
formats, Text formats, Graphics formats, GIS formats, DB formats, etc).

PIBOT can assimilate, in its cultural baggage, the entire Web, so creating
an ontological and semantical representation of each information contained
within (in very few time, with the advent of PIBOT, all the currently
search engines are destined to become obsolete and prehistoric).

Using the natural language, and with its multilingual features, PIBOT is
able to converse with humans on each argument pertaining to the human
knowledge. PIBOT possesses conversational abilities that are infinitely
advanced compared to humans, both in terms of width of the expressions,
both for its contexts manipulation abilities, both for its mnemonic
abilities: PIBOT will too be able to entertain millions of contemporary
conversations with millions of persons in every part of the world."

So to claim to have built such an all-encompassing singularity in the FIRST
VERSION is truly extreme. My claim is not nearly so grand. My MC droids in
the first version are like human intelligent, clever five year olds. They
have very little knowledge, only a couple skills, and only the fundamental
human reasoning capability. All else must be version 2/3 ... developments
that ultimately allow it to graduate from a high school, and then from a
university. At that point (version 3/4) maybe it will be ready to enter
into specialized singularity training and self-modification so as to maybe
(with the help of a 1000 scientists) grow ultimately into the technological
Singularity. So even after the birth of a V1 MC droid, there will be much
work to be done by the AI field to continually add on new math skills,
simulation skills, natural language processing skills, knowledge bases,
sensor capabilities, actuator capabilities, creative skills, etc., etc..

Right now I an detirmining how hard it would be for my V1 MC droids to use
their fundamental human reasoning skills to perform the level 2 textual
entailment recognition task called for by the Pascal Challenge. Will it be
right out of the box? After a "semester of specialized training"? Not at
all? Stay tuned ... For those following this debate, check out the Pascal
TER Challenge at
http://www.pascal-network.org/Challenges/RTE/Introduction and see how
your own pet AI system would fair against it. Does it prove anything about
the consciousness of the system, as opposed to a Turing passer? Is that
system starting to understand language the way human level consciousness
does? Would you need to know that it is doing it in the same way as well?
Would you then say it is exhibiting at least some machine consciousness?
You be the judge.

Ken Woody Long
http://www.artificial-lifeforms-lab.blogspot.com/



This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:55 MDT