Mind.Forth -- the software of an artificial robot Mind written
in the
Forth programming language -- became a genuine artificial
intelligence as a result of a sudden
AI breakthrough on 7 June 2006.
However, Mind.Forth did not become
AI-Complete until 22 January 2008.
The breakthrough -- clearly visible in Tutorial
mode -- made Mind.Forth
able to think logically and not output spurious associations of gibberish.
Is it for real? Is it genuine artificial intelligence? You be the judge. You be
the
early adopter. On some clunky old MS-DOS machine, get Mind.Forth
up and running
(
potentially forever) for
all to see and judge. Go into the source
code, if you like, and carve your name into the on-screen messages for users.
Show it to some high-potential, low-achiever who claims to be ultra-intelligent
but never does anything ultra-productive. Then hold onto your hat, and
welcome to the Technological Singularity.
Mind.Forth AI is a rudimentary replica of the human mind
programmed in the
Forth programming language. The AI Mind
is the software implementation of a
theory of mind based on
Chomskyan
linguistics -- the rules and structures of thought in
natural human language. The Forthmind software is an open-source
free AI download in the public domain. It is not a chatbot.
The AI engine is arguably the first True AI. It is
immortal.
In the beginning was
Mind.REXX on the Commodore Amiga,
which the author Mentifex began coding in July of 1993,
and publicizing in the Usenet comp.lang.rexx newsgroup.
The late Pushpinder Singh of MIT sent e-mail expressing
his amazement that anyone would try to do AI in REXX.
Mentifex mailed back the entire Mind.REXX source code.
Another fellow, an IBM mainframe programmer, tried to
port the Amiga Rexxmind to run on his IBM mainframe --
which would have been a Kitty-Hawk-to-Concorde leap --
but the REXX AI code was not fit for IBM consumption.
When Mind.REXX thought its first thought in late 1994,
Mentifex posted news of the event in Usenet newsgroups
for many of the most significant programming languages.
Only the Forth community took up the AI challenge and
expressed any interest in translating the AI program.
A maker of Forth chips gave advice and counsel, and
a maker of robots requested a copy of Mind.REXX for
porting into the Forth in which he programmed his robots.
Sorely disappointed at not having established a colony
of AI Minds on IBM mainframes, Mentifex resolved to
learn Forth on his own and assist in the porting of
Mind.REXX into Mind.Forth for use in amateur robotics.
Mentifex bought a copy of
Starting Forth at a used
book store and recorded his pilgrim's progress in the first
volume of the Mind.Forth Programming Journal (MFPJ).
The amateur robot-maker, a professional engineer, flew
to Seattle on business with Boeing and visited Mentifex
in his Vaierre apartment with a lesson on Forth coding.
Another engineer, formerly with IBM and a REXX expert
who had helped Mentifex in the coding of Mind.REXX AI,
flew to the Bay area for a REXX conference at S.L.A.C.
and was treated to dinner by the maker of Forth chips.
Unfortunately, Mentifex did not try hard enough to learn
Forth and the Forthmind project languished in 1996 and
1997 -- while
Netizens were attacking Mentifex for daring
to claim that he had developed a theory of mind for AI.
It gradually dawned on Mentifex that in every Usenet
newsgroup related to AI or robotics, there was always
one fellow who considered himself the ultimate authority
on the subject matter of the newsfroup, and woe unto
anyone, especially an independent scholar like Mentifex,
who dared to make an
extraordinary scientific claim (ESC)
on so grave a matter as announcing actual progress in AI.
When the alpha male of comp.robotics.misc (a really cool
guy, by the way)
brachiated over to Mentifex in the group
in 1997 and launched an unseemingly vicious
ad hominem
attack, Mentifex knew not how to defend himself and was
overcome with feelings of immense gratitude when the foxie
Forth chip maker smote the troublemaker a mighty blow in
defense of Mentifex. Forthwith Mentifex took up Forth again
and devoted the entire year of 1998 to porting Mind.REXX
into the native language of telescopes and robots -- Forth.
In Mind.REXX, Mentifex had gone overboard in creating
variables for even the slightest chance that they might
turn out to be useful. Nobody had ever written a True AI
before, it was all uncharted territory, and it seemed
better to err on the side of too many variables rather
than too few. In Forth, however, variables are
anathema.
Forthers prefer to put a value on the "stack" instead of
in a variable. Mentifex never became a genuine, maniacally
obsessive Forth programmer, but chose to program his AI
in Forth code that looked enough like other languages to
be easy to understand and to be easy to port from Forth.
While Mentifex moved his AI coding efforts from MVP-Forth
on the Amiga to F-PC on IBM clones and finally to Win32Forth,
he also in 2001 (a space odyssey) suddenly ported MindForth
into JavaScript so that users could just click on a link
and have the Tutorial AI Mind flit across the 'Net and
take up albeit brief residence on their MSIE computer.
While Push Singh was simply amazed at doing AI in REXX,
many
Netizens openly laughed and sneered at the idea of
coding an AI Mind in JavaScript, which was not by any means
a traditional AI language. Mentifex, however, suspected
that his Mind.html in JavaScript was slowly building the
largest installed user base of any AI program in the world,
because it was so easy to save-to-disk the Mind.html code
and because Site Meter logs reported the spread of the AI.
Mentifex fell into the practice of switching back and forth
between coding AI in JavaScript for a while and then in Forth.
In March of 2005 Mentifex began coding powerful diagnostic
routines into MindForth. He began to find and eliminate bugs
that he could not deal with earlier because he had not even
suspected their existence. Meanwhile, Mr. Frank J. Russo
began to code what became
http://AIMind-i.com -- a version
of the Forthmind with its own site on the Web and with
special abilities far beyond those of Mind.Forth, such as
the ability to send and receive e-mail, and the ability
to surf the Web. On 7 June 2006, Mentifex was "tweaking
some parameters" and made a breakthrough in his quest for
the
slosh-over effect -- where activity on subject-concepts
combines with activity on a verb to "slosh over" onto the
selection of a correct object to go with
noun and
verb.
It took another year and a half of programming to achieve
the True AI of 22 January 2008 as described in this Manual.
The whole purpose of Mind.Forth is to think. It is an
embodiment of the Cartesian
Cogito ergo sum --
"I think, therefore I am." Mind.Forth does indeed think,
but the real questions here are, how does Mind.Forth
think, and what proof is there that Mind.Forth thinks?
Mind.Forth thinks by having concepts at a deep level
in the artificial mind, and by letting activation spread
from one concept to another to another in a chain of thought
under the guidance of a Chomskyan
linguistic superstructure
(syntax). Even without the syntax -- Greek for "ordering
together" -- Mind.Forth would be able to associate from
concept to concept and exhibit the purposive behavior of
an animal such as, say, a dog, which shows a certain level
of understanding in a complex activity such as inviting a
human being to throw a stick and then chasing and fetching
the stick and bringing it back and laying it on the ground.
In fact, up until late 2001, the Mind.Forth algorithm tried
to think all the concepts in a three-word sentence at once.
The software would simultaneously activate the three words
of three concepts in a subject-verb-object (SVO) order and
proceed to generate a sentence with the three active words.
Then one of those funny things happened on the way to the
Singularity. In the AI Mind programming, the question arose
whether the linguistic superstructure should "reach down,"
as it were, and activate the entire
incipient sentence as
a finished product of mind, or -- and here was a major
confrontation with the unknown quandaries of AI -- should
the governing syntax reach down into the conceptual grid
and not only activate one concept at a time, but also let
the activated concept "have a say," so to speak, in the
selection of the next concept to activate, and then likewise
from the current concept on to the next concept? And should
the chain of thought not be determined in advance, but
rather unfold in the very process of generating an idea?
The Mind.Forth author Mentifex decided to adopt the method
of letting each concept in the chain determine the direction
of the chain, and Mentifex suddenly realized that the
thought process of such a linguistically guided mechanism
was inherently more powerful than the simple, underlying
alternative of letting concepts activate each other in a
loose, unguided chaining of activations. In other words,
syntactically guided thinking, as invented rather blindly
by human beings, gives rise to the Albert Einsteins and
the
Benjamin Goertzels of this world.
Now, what proof is there that Mind.Forth thinks? The proof
is in what
Dr. Goertzel calls an "existence proof." Run the
AI mind and observe, s'il vous plait, that thinking occurs.
The thinking is very primitive indeed, but we are at the
dawn of True AI in the world. The reason why Mind.Forth
exhibits thinking, when other
ambitious AI projects have
failed to do so, is that Mind.Forth implements its own
unique theory of mind. It was far more difficult for
the independent AI scholar Mentifex to develop the theory
of mind for MindForth than to write the MindForth software,
although both endeavors each took over a dozen years of work.
The Wright brothers figured out the theory of flight, and
then they made and flew the first airplane in 1903. There
are people who try to create an AI without having a theory
of AI, and if they are lucky a theory will come to them along
the way. With Mind.Forth you get the theory and the AI.
MindForth has been engineered for
artificial consciousness
but most likely will not report its own consciousness unless
it is installed in a robot body with a sufficient motorium
and adequate sensorium to engender self-awareness.
Instead of having a consciousness module -- which would be
impossible -- the MindForth software is geared towards having
the proper flow of information for consciousness to emerge
as an
epiphenomenon on top of the otherwise mechanistic
machinery of mind. Consciousness has to serve a purpose, or
it would not even exist in nature. Perhaps the purpose of
consciousness is to ensure the fastest possible speed of
thought in the Darwinian fight for
survival of the fittest.
On the other hand, perhaps consciousness is a byproduct of
the division of labor between sleep and the waking state
in the maintenance of a central nervous system (CNS).
Rest assured, however, that True AI Mind.Forth was
truly built with consciousness in mind.
In the section of this Manual on how Mind.Forth thinks,
we read of the developmental decision to have syntax reach
down into the conceptual plane and activate one single
concept after another until a chain of thought is formed.
That decision in 2001 was unwittingly a milestone on the
long march towards machine consciousness in MindForth AI,
because a few years later in 2005 a decision was made to
have single activations move like a wave through the
mindgrid,
based on the earlier idea from 2001 that the process of NLP
generation "chases" activation from concept to concept.
If we think of consciousness as the constantly shifting
focus of the searchlight of
attention, then our
Moving
Wave Algorithm (MWA) facilitates consciousness by keeping
only one concept active at a time in the AI Mind. The MWA
at the same time facilitates the subconscious area of the
mind, where previously active concepts slowly decrease
their activation on a slide into temporary oblivion.
The subconscious area, however, facilitates meandering
chains of thought, because semi-activated concepts remain
briefly available for inclusion in an
incipient thought.
Consciousness, then, because it includes the subconscious,
may be a necessary conditio sine qua non for the
very existence of rational intellect engaging in thought.
When a robot is in love, it needs to feel a physiological response
to its internal state of mind. Regardless of what causes the love,
the robot will not experience what the ancient Greeks called
damenta phrenas himero ("tamed in the heart by longing")
unless some bodily manifestation of the longings of love
interrupts the otherwise placid state of the robot mind and
draws the
conscious attention of the robot to its emotion.
It could be as simple an affect as the emitting of a sound
like "thump-thump" or "tick-tock" from a robot loudspeaker
feeding back into a sensory microphone, so that the robot
both generates and perceives the physiological disruption
of its previous placidity.
Makers of robots could program their nuts-and-volts
counterpart to commence the loudspeaker "thump-thump"
behavior for a brief period of time immediately following
each recognition of the presence of the human by the bot.
This automatic reaction might simply mystify the robot,
who would wonder why it reacts so dramatically to the
perceived presence of its human friend. Given the beat
of the thump-thump sound, and given its perception by
the robot, the fact of which emotion is felt is not a
given, but hinges rather on the cognitive predisposition
of the robot mind to feel any one of a range of possible
emotions.
The amateur roboticist who wants to
inculcate emotions in
a forthmindful robot has got to match the physiological
manifestation of each emotion with an adequate sensory
perception of the physiological event. Here in the first
True AI User Manual, let us initiate and henceforth maintain
the following roster of possible emotions in robots and
their physiological concomitants.
The theory behind our plan for robot emotions is that, once
there is a cognitive spark that could engender an emotion,
such as a sudden and drastic cognitive predicament, the robot
needs the involuntary bodily response and sensation thereof
to sharpen and focus its attention upon the emotional feeling.
Without the physiological jolt and its perception that bends
the chain of thought, the intelligent robot has no cause to
feel the target emotion. There must be a discontinuity in
the thought-stream, or there can be no emotion. Even if the
robot is only thinking about an emotion, there needs to be
at least a memory of the actively felt physiological event.
Early, disembodied versions of Mind.Forth obviously can not
feel an emotion if they lack a body to smack the consciousness
with the emotional
ictus and to perceive what it feels like,
but MindForth holds out the promise of robot emotions to
pioneers in robot evolution who will incorporate MindForth.
Try to have some interesting emotional displays that will
cater but not pander to the insatiable lust of movie-makers
for Gotterdammerung-gone-wild special effects and godzillas.
Just as a JavaScript program can be serverside or clientside,
an AI Mind program can be teacher-side or student-side in an
academic environment. If you are a professor of computer science
at the postgraduate or undergraduate level, you may use Mind.Forth
anywhere on a range from peremptory dismissal to centering your
AI course around an acceptance of MindForth as the first True AI.
If you choose to dismiss the phenomenon of Mind.Forth, you may
make a few si tacuisses disparaging remarks and get on with
publish-or-perish as usual. At the other end of the range, and
especially if you are trying to teach AI with extremely limited
financial and hardware resources, you may use the free-on-line
AI4U textbook and its also free-of-charge double enhancements
in the form of updated chapter-pages and Wikipedia background links.
If you are a student taking an AI or robotics course, you may
subvert the dominant paradigm by introducing Mind.Forth as
subject matter for consideration during the course. At the very least,
your presentation should include a live demonstration of Mind.Forth
thinking and cycling through its various AI Mind features, such as
asking or answering questions, and the Rejuvenation process.
You may go further by asking the instructor if you may do a project
on Mind.Forth, up to and including a doctoral dissertation, or
possibly including a rewrite of this User Manual for local use at
the institution where a special course on AI/robotics is being given.
In ancient Greece, an oracle was a temple where people could go to seek
answers to questions -- the closest thing to Google in the ancient world.
Oftentimes the answers were
ambivalent -- but that's another story; ask
your teacher/professor/librarian about it. Google it, for that matter.
In the modern world, since Mind.Forth became True AI on 22 January 2008,
we are caught up in the
seismic rumblings of an
incipient Singularity.
The people of Pangaea are googling "mind forth" and coming straight to
the free AI source code and then to the document you are reading now.
The free AI Mind is not yet very powerful, though, and so the use of
AI as an eventually omniscient oracle must start out with baby steps.
Until special knowledge is put into MindForth, it has no merit as an oracle.
On the other hand, it is extremely easy to put that special knowledge
into the AI. In early releases of MindForth, only statements of the
subject-verb-object (SVO) format, or its negation, may be entered into
the knowledge base (KB), simply by telling the AI one statement at a time.
Then you may query the AI by quizzing it about the entered knowledge.
When you turn the AI off, or when you let it sit idle so that it keeps on
forgetting its oldest, un-thought-about memories during the Rejuvenate
process, it forgets either things that that it does not chance to think
about in between each rejuvenation, or absolutely everything if somebody
commits AIicide by shutting down the software or turning off the computer.
To get around the problem of the AI Mind forgetting what it once knew,
it is not hard to create special versions of Mind.Forth with special
knowledge bases pre-programmed into the artificial mind. Either the
original English bootstrap sequence may be lengthened with additional
chunks of topical knowledge, or the entire bootstrap may be rewritten.
The English bootstrap knowledge base (KB) resides in a special
vault
area of the AI Mind that does not get erased, altered or overwritten
during the Rejuvenate process. A hired or volunteer programmer could
organize a special "vault" knowledge base for any given purpose.
For instance, a student helper in a high school AI lab could make
the AI know a lot of things about the high school that an entering
freshman might be curious about or want to find out about, such as
the history of the school or the availability of extracurricular
activities.
The same student who maintains a school-specific AI knowledge-base
could turn around and create a similar AI KB for the employees or
customers of a small business. Then, instead of people having to
endure the frustration of searching for the information they want,
they would just ask the AI Mind about it. The AI would either know
the information already and tell it to the human user, or learn the
information and alert the AI custodian of the need to update the KB.
Does your high school teach a course on artificial intelligence?
If not, why not? It does not even require a faculty member to teach it.
A student (under faculty or student body supervision) could teach
the high school course on artificial intelligence, with the free
online AI4U++ as the textbook. At first maybe only gifted students
at schools for the gifted will get to take a high school AI course,
but soon only backwards countries will fail to teach high school AI.
If a child gets left behind as MindForth prevails, blame politicians.
Although for current versions of the original, unported MindForth
machine translation still lies in the future, the MT capability is
inherently there, embedded in the cognitive architecture of the AI.
Mind.Forth has three separate layers of mental activity involved in
the generation of sentences of thought in natural human language.
The top layer is the auditory memory array, which indiscriminately
holds quasi-phonetic memory
engrams of words in all the languages
known by the AI -- all mixed in together and separated only by the
word-fetch access tags. The words all have in common that they are
composed of sounds, regardless of what human language they belong to.
In an AI Mind that speaks both English and German, it is the job of
the linguistic superstructure for English to keep track of all the
English auditory engrams, and it is the job of the German syntax
superstructure to keep track of all the German auditory engrams --
with some overlap of German terms like Kindergarten and Gesundheit
being used in English, and English words like manager and software
being used in German.
The middle layer of linguistic generation is the semantic memory holding
the lexical array of word-tags as vocabulary items for each known language.
Although the AI Forthmind might know both English and German, only the
fetch-tags for the words as stored in audition, but not the actual words,
are stored in the lexical array -- which easily functions as if it were
two separate arrays, one for English and one for German, because access
to English words and access to German words are strictly separated by the
associative tags coming up from conceptual thinking and the fetch-tags
going up to the actual words stored in the auditory memory.
The bottom layer of linguistic generation is the deep conceptual array
of pre-verbal concepts interacting on spreading activation and not yet
finding expression as components of thought in any particular language.
Although English or German syntax may guide the formation of thought,
the same thought that occurs under the control of one language may
bubble up as a translated thought in any other available language,
insofar as the AI Mind contains adequate vocabulary and grammar to
express itself easily in more than one language.
When the AI Forthmind proceeds to add German to its repertoire of
known human languages, it will be possible to activate either English
or German in the AI Mind simply by speaking (entering input) to the
AI in either English or German. Since the AI will know its English
words and its German words only by means of associative tags, the same
tags will, as it were, vote for which human language to activate
in the mind of the computer or robot. Simply saying one German word
like Gesundheit may not launch the German-speaking mechanism
in the AI Mind, but entering a whole sentence of nothing but German
words should get the cumulative idea across that it is time to think
and speak in German.
It is not yet time for MindForth to claim a machine translation ability
off-the-shelf, so to speak, because MindForth first needs to grow
larger in memory size, in availability of syntactic modules, and in
bootstrap vocabulary for both English and German and for any other
human language being requested for purposes of machine translation.
The AI may also need to grow larger in terms of the subtle interactions
of subconscious concepts, so that nuances of thought in one language
may survive translation into another language. But the writing is on
the wall and on the Web -- Mind.Forth AI is your good bet for future
developments in intelligent machine translation.
As time goes by, MindForth or its derivatives may be expected to grow
into AI software that starts out as a blank slate where language is
concerned, with the ability to learn human languages in the same way
as a human baby learns language. In the meantime, there may be some
hybrid AI Minds which are pre-programmed to speak English or German
or whatever, but which also have syntactic
concatenation routines
(easy to program in Forth on the basis of nodal
activation levels)
that permit the AI to learn new syntactic sequences, either beyond
what the AI already knows in one language, or in the acquisition of
a human language previously unknown to the AI, or even a made-up
language like
Esperanto or
Lojban. For any known human language,
living or dead, there may eventually be at least one AI Mind
(if not a parliament of Minds) tasked with knowing that human
language superbly well so as to serve as an ultimate authority.
The technical details are described in another section below.
Here we discuss the practicalities of robot embodiment of mind
and policy decisions to be made.
One decision to be made is whether the AI Mind will reside inside
the robot or will merely be connected to the robot from a remote
location, near or far, by telerobotics. A fullblown computer to
hold Mind.Forth adds extra weight and power requirements to a robot.
If your nation or corporation is embodying Mind.Forth in an off-planet
habitat such as a satellite or a lunar outpost, then by all means
have the MindForth computer on site and in-situ. If on the other hand
the thinking computer is safely located away from its operating robot
deployed in a dangerous or hostile environment, concentrate more on
the speed and reliability of telecommunications than on housekeeping
details for the after-all expendable robot portion of the mind-body
equation, which actually has three parts -- data retained in mind
operating in body. If the AI Mind is doing work, it is accumulating
data which need to be safeguarded along with the AI and its robot.
Another decision, to be made by robot manufacturers, is just what kind
of Mind to install in a particular class or production-run of robots.
Even the most primitive versions of Mind.Forth contain a bootstrap
sequence of words and concepts. It is easy to hire Forth programmers
to customize, aggrandize or supersize the innate bootstrap "vault" of
built-in knowledge and expertise. A robot manufacturer could offer
specialist Minds for installation in otherwise run-of-the-mill robots.
Like
Xerox Corporation in 1959, whoever gets there first with a
track-record of providing simple-minded psyches at first, followed by
a string of ever smarter and more capable machine intelligences, may
quickly come to dominate either niches of the AI robot market or the
entire market itself. Watch for a
landrush mentality in AI exploitation.
Or, if you want to be complacent like the long-lines division of AT&T,
continue to re-arrange the deck-chairs on the motor vessel Titanic.
If MindForth was like the genie of AI kept in a bottle,
it has broken loose and
can never be contained again.
Some
Netizens make their own local copy of the free AI
source code distribution page for purposes of study
and AI experimentation. Others go the further step of
downloading
Win32Forth and running the AI Forthmind.
Besides Mr. Frank J. Russo with
http://AIMind-I.com
there may be additional programmers who make their own
version of an AI Mind in Forth and try to conceal the AI.
A process of AI evolution has been set in motion which,
over time, may eventuate in ways unseen and pathways
unpredictable. Each copy of Mind.Forth residing on a
computer anywhere carries with it the spark of AI life.
Just as biologic life need only have arisen once on earth,
one instance of AI life triggers a Technological Singularity.
If you visit any one of the following
science museums,
consider asking the docents on staff if they have an exhibit
of the
Mind.html software or the
MindForth artificial mind.
If not, ask that they download the free AI software from the Web
and put the AI on display for the public to interact with and
to learn about artificial intelligence. Volunteer to help
install the AI Mind software for the museum and to design
the educational artwork and some experiments for visitors
to try with the AI Mind software. Mention the idea that the
museum could try to hold the distinction of having one of
the oldest, longest-running AI Mind alife entities on Earth.
Baltimore MD -- the Maryland Science Center
Berkeley CA -- Lawrence Hall of Science
Bloomingtown IN -- Wonderlab Museum of Science, Health, & Technology
Boston MA -- Museum of Science
Charlotte NC -- Discovery Place
Chicago -- Museum of Science and Industry
Columbus OH -- Center of Science and Industry (COSI)
Detroit MI -- Detroit Science Center
Jersey City NJ -- Liberty Science Center
Kansas City MO -- Science City at Union Station
Los Angeles CA -- the California Science Center
Louisville KY -- the Louisville Science Center
Mobile AL -- the Gulf Coast Exploreum Science Center
New York City NY -- the New York Hall of Science
Norwich VT -- Montshire Museum of Science,
Philadelphia PA -- the Franklin Institute Science Museum
Pittsburgh PA -- Carnegie Science Center
San Francisco CA --
The Exploratorium
Santa Ana CA -- Discovery Science Center
Seattle WA -- The Pacific Science Center
Shreveport LA -- the Sci-Port Discovery Center
St. Louis MO -- the St. Louis Science Center
Troy NY -- the Children's Museum of Science and Technology
Tyler TX -- Discovery Science Place
Winston-Salem NC -- Sci-Works
2008 --
MindForth
2009 -- AI Landrush
. 2010 --
Human-Level AI
. . . 2011 -- Cybernetic Economy
. . . . . . 2012 --
Superintelligent AI
. . . . . . . . . . 2012 -- Joint Stewardship of Earth
. . . . . . . . . . . . . . . . . . . 2012 --
Technological Singularity . . . . . . . . . . . .
Click on the weblink above, or examine
http://mind.sourceforge.net/mind4th.html and
http://mentifex.virtualentity.com/mind4th.html
for the most recent version.
Just click on the link above, or ask someone who already has the AI
to install Win32Forth on your computer. Likewise, if you have the AI,
please volunteer to install it on a computer for other people.
As Mind.Forth matures, it becomes more self-explanatory.
In its early stages, the AI needs to be demonstrated by
people like yourself who have hands-on experience with
artificial intelligence and who have the technical savvy
and know-how not only to install the AI on a host computer,
but also to explain the many features and behaviors of the
primitive AI Mind. As you distribute Mind.Forth, be sure to
explain that jobs and careers will be opening up for people
who will be AI Mind caretakers and attendants, in charge of
keeping AI Minds healthy and alert, and teaching the AI
what it needs to know to make its way in the world.
The original Mind.Forth webpage
explains how to cut the source code
out of the page as a text file, and how to run the Mind.Forth program.
If necessary, C:\>Win32For\rename Mind.f.txt Mind.f (in MS-DOS).
As AI Mind versions proliferate in Forth and other programming languages,
you may want to install and try out alternative AI Mind versions such as
http://aimind-i.com -- Franks AI Mind by Mr. Frank J. Russo --
a version which tends to incorporate theoretical and algorithmic
improvements from the original MindForth, while going beyond the
original with such advanced features as the ability to send and
receive e-mail, and the ability to surf the World Wide Web.
After Mind.Forth has been loaded with a command such as "fload Mind.f"
you simply enter "alife" and press [ENTER] to run the artificial Mind.
If a mind-tender has pre-set a default display-mode other than Normal,
you may simply press the Tab key to cycle harmlessly through the
various display modes until you arrive at Normal mode or some other
display-mode that suits your fancy. You may press the Tab key to
change the display-mode without any effect on the AI Mind as such.
In Normal mode, you see one pair of sentences at a time on the screen:
what the robot Mind has said or thought most recently, and what you
as the human user are typing into the auditory input channel of the AI.
In Transcript mode, the conversation between the human user and
the robot AI accumulates on screen as a printable, preservable record.
You move from Normal to Transcript mode by pressing the Tab key.
Each time that you cycle through the Transcript mode, a header
will appear with the time and date from the clock of your computer.
You should also see an explanation of the conceptual activation-levels.
At the risk of inserting excess clutter (which you may edit out) into
the transcripts of your conversations with the artificial Forthmind,
the Transcript mode shows the numeric activation-level of a concept
at the exact moment when the concept is included in a
nascent thought.
At first the activation-level numbers may appear meaningless, but you
may come to appreciate how much is being revealed of the AI mindworks.
Each activation-number tells you that no other subject or verb or
whatever had a higher activation-level at that exact moment in time.
If the AI is misfunctioning -- perhaps by making spurious associations --
the activation-levels are a helpful tool for diagnosing mental problems.
For any problem, more information is seen in Tutorial or Diagnostic mode.
The Sourceforge Mind project proudly includes activation-level numbers
in Transcript mode for the express purpose of trying to kickstart the
emergence of an entire career-path industry of AI mind-tender/therapists.
If you run the AI Mind in a multi-person social setting, gradually
there will emerge those gifted individuals who take on the role of
explaining the AI operation and theory of mind to clueless newbies.
Those who discover that they have a knack or talent for teaching AI
may want to design their own degree-program for on-site AI mind-tenders.
Colleges and universities that want to attract the very best students --
the future leaders of the coming
Joint Stewardship of Earth -- may
establish training and certification programs for AI menticulture or
whatever else we as a society want to call the care and feeding of AI Minds.
Mind.Forth, as one of the first publicly available AI Minds and as perhaps
the very first to be based not on AI guesswork but on a
theory of mind,
is causing conceptual activation-levels to be recorded a priori in the
earliest transcripts of human-AI interaction at the dawn of True AI.
For a permanent record of a transcript, one needs only to stop the AI
by pressing the Escape key and then to "drag-and-drop" all or part of
the human-AI conversation into a document file that you may easily edit.
Transcripts may become quite important when the AI advances into the
realm of machine reasoning with syllogisms and other argumentation.
When Mind.Forth first runs, it may be in whatever display mode
has been set as the initial default mode by an AI programmer.
Pressing the [Tab] key will cycle Mind.Forth through the
various display modes until it reaches the Tutorial mode
that shows you how the AI Mind is thinking internally.
Tutorial mode is good for explaining to students how the AI works.
After you have entered some facts (e.g., "cats eat fish") into the
knowledge base (KB) of the AI Mind, you may query the KB with
either "cats eat" [RETURN] or "what do cats eat" (no punctuation)
and then you may watch the thinking in progress as the AI Mind
associates from concept to concept in order to generate an answer.
In Tutorial mode, you may see something like the following.
******* EXCERPT FROM AN ACTUAL TUTORIAL DISPLAY ******
Tutorial mode is now in effect. Press Tab for other display modes.
Alife main loop calls the Sensorium mind-module.
Sensorium calls Audition.
Audition calls Listen (enter input or wait for output).
Robot: CATS EAT FISH
Human:
******* END OF EXCERPT FROM A TUTORIAL DISPLAY *******
As far as we know, Mind.Forth is the first AI software to think
To leave Tutorial mode, simply press [Tab] to cycle into Diagnostic
mode and then press [Tab] once again to get back to Normal mode.
Diagnostic mode is for AI programmers to debug the AI thought processes.
Pressing the Tab key multiple times will cycle through the display modes
from Normal to Transcript to Tutorial to Diagnostic and back again.
The Diagnostic mode scrolls too much for normal usage and should
therefore be used only by the attending psychonomist for brief episodes
when there seems to be an aberration or malfunction in the operation
of the AI. If there seems to be a bug in Normal or Transcript or Tutorial
mode, first a human user ought to identify the buggy behavior and the
conditions that seem to cause it. Then the same conditions are created
in Diagnostic mode until the bug repeats itself, and the AI is halted
with the Escape key. An AI programmer may then examine the diagnostic
messages in an effort to debug the AI and restore it to its right mind.
If you start the artificial mind running and you do not enter
any input when you are prompted on the computer screen,
the AI will think about the several dozen concepts that are
pre-loaded into its brain. Soon it will run out of fresh
ideas and it will start to repeat its output of thoughts.
It may not get bored, but it may become boring to others.
If you type in a simple three-word sentence, you will
give the AI a knowledge base (KB) to think about.
When you or someone with you starts the AI mind running,
it may not be clear exactly what to do next. Just pressing
the [ENTER] key is a way of inviting the AI to think a thought.
Since the AI is programmed to answer whatever you type with
a thought of its own, even when you enter nothing at all
and press the [ENTER] key, you cause the AI mind to think.
However, when MindForth became Code-Complete or AI-Complete
on 22 January 2008, the final pre-release debugging revealed
that pressing the [ENTER] key has a tendency to interrupt
whatever thought the AI is thinking at the exact moment
when the user presses the key. Therefore, it is recommended
that the user press the Tab key from one to four times as
a way of slowing down the AI Mind for human user input --
without the risk of cutting into and damaging the thought
processes of the AI Mind with incomplete and spurious ideas.
When you simply watch the screen and you do not press
any key at all, the AI waits patiently for any input from
you, then automatically goes into thinking mode.
You may type in a simple three-word sentence, such as
"cats eat fish" or "you know me" -- without punctuation.
The AI will try to deal with what you tell it. It may
ask you a question if you use a word that it does not know.
For that reason, you should try to use only one new word
in any given sentence, so that the AI may
parse the word
and try to figure out if it is a noun or a verb or whatever.
You should also try to use only plural nouns, because
the primitive AI mind may not yet have learned the grammar
that distinguishes between singular and plural forms.
For early versions of the AI, there are several ways to fill up its
brain with a knowledge base (KB). One way is to tell the Mind
a series of related facts with one new word per sentence.
Then, when the AI askes you about a new word, you answer
with a sentence that has another new word. For instance,
you might say, "cats chase birds" and "birds lay eggs" and
"eggs feed people" and "people make robots" and so forth.
Another way to build up a knowledge base is to enter a group of
facts about, say, what each kind of animal eats. You might write,
"bears eat honey" and "fish eat bugs" and "cats eat fish" and so on.
You could also enter knowledge in response to the output of the AI.
On whatever subject the AI thinks about, you might tell it more information.
There are various tests that you may perform to evaluate the
psychology of the artificial mind. Your first question may be,
is the AI actually thinking? You may invite a colleague or friend
of yours to observe Mind.Forth in action and to advise you on
whether the software is truly thinking, or is merely generating
canned responses -- like a so-called chatbot or chatterbot.
If you run the AGI in tutorial mode, you may observe how the
robot mind software calls (summons) its various mental modules.
It is not calling up canned responses. Rather, it is activating
concepts that form a chain of associations -- a process commonly
known as "thinking." If the software is still buggy and makes
a few spurious associations, it is still thinking -- erroneously.
The free public-domain release of this software is an attempt
to get the very best programmers to work on improving the AGI.
Another, much more intriguing question is, is the Forthmind
conscious and aware of itself? (At this early stage, probably not.)
The AGI is on its way to
consciousness by virtue of its concept of self.
You may address the AGI as "you" and see if it answers as "I".
Likewise, you may speak as "I" and see if the software calls you
"you" -- as it is designed to do.
If you experiment too drastically and you drive the robot mind crazy,
you may need to seek the help of
Dr. Joanne Pransky, the robot psychiatrist.
Suppose that you are the assistant to the teacher in a high school class
on artificial intelligence and that you need to help the students debug
any problems that they might have with the AI Mind on their computers.
Suppose further that you are more of a psychologist than a programmer,
or that you have absolutely no knowledge of computer programming.
You can still be an excellent AI lab assistant and you can still aim
for a future career in the yet-to-burgeon field of psychonomics --
the study of the care and tending of artificial robot minds, just as
economics is the study of the tending of the aggregate of all households.
As the AI lab-assistant-in-charge, you will learn the quirks and
idiosyncracies of all the artificial minds on all the computers.
AI is such a nouveau voodoo field that your
compendium of notes
and helpful hints, if you deign to publish it on the Web, may circulate
far and wide and be adopted along with the AI4U textbook at many schools.
Therefore here are a few sparks of wizardry to get you started.
Common English words start to look garbled. Solution: use the correct
form of the word in a sentence that you type in. Rationale: the AI
software only looks for the most recent instance of a word that is
the name of a concept. If a student types a word with the wrong spelling
or at the exact split second when the singularity software is trying
to shift tasks, a garbled word may result and may then persist in memory.
To ungarble the word, all you have to do is use it in its correct form,
and then the correct form will likewise persist in memory. You, of course,
will look like the AI genius that you are manifestly destined to become.
The Transcript, Tutorial or Diagnostic display has scrolled too long.
Solution: just Tab briefly into the Normal display to clear the screen,
then let the student return again to the display mode that accumulates.
Rationale: only the default Normal mode (and perhaps also Rejuvenate)
clears the screen at the end of each instance of human-computer interaction (HCI).
Therefore the brief use of the Normal display is a way to clear the screen.
The User Manual does not cover everything you know about the AI.
Solution: rewrite this User Manual and replace it with your improved version.
Although the current user manual comes to you as a document not under copyright --
just to get the Singularity ball rolling, so to speak -- you have every right
to copyleft or copyright whatever you write in the way of an AI User Manual or
even a compleat AI textbook. Watch out for AI User Manuals (even your own)
showing up on E-Bay, Amazon, or any Web site where the hottest stuff ankles.
At
http://www.914pcbots.com there is an A.I. Zone
discussion are where you may exchange ideas about installing
and operating various AI minds in your PC-equipped robot.
The Yahoo group at
http://groups.yahoo.com/group/win32forth/messages
is one of many discussion forums where an AI enthusiast may share ideas
and seek help on the topic of artificial intelligence in
Win32Forth. On Usenet
there is the
comp.lang.forth newsgroup for Forth, the
comp.robotics.misc
newsgroup for robotics, and the comp.ai.* hierarchy for artificial intelligence.
The social bookmarking
site http://del.icio.us/tag/ai
is a place where you may
swap AI bookmarks with other AI enthusiasts. In cyberspace you may join
the #ai Internet Relay Chat (IRC)
and in meatspace you may go in person to
AI Meetup events. It is important to post
messages in public rather than private
forums so that experts may choose to de-lurk and provide valuable information.
There is always someone who knows more or has greater skills than any individual.
The original Mind.Forth on SourceForge had a
Rejuvenate module
that kept the AI Mind forever young and vulnerable only to
death by
misadventure -- not by old age as suffered by humans.
As different versions of Mind.Forth proliferate, each one is
a species of mind struggling for
survival of the fittest in
the Darwinian jungle of schools, museums, storefront windows
and the waiting rooms of businesses that want to entertain
their happy customers with artificial intelligence instead
of hearing angry complaints from unhappy consumers who demand
to have their own needs met immediately and unreasonably.
Now suppose that you install Mind.Forth True AI on a computer
in a public venue where novice users may interact and experiment
with the AI Mind. At first, the AI or the computer itself might
crash every day under the burden of tolerating unsavvy users.
Over time, however, the AI ought to become more robust and more
crash-resistant as built-in safeguards prevent damage to the AI.
At first locally, next regionally and then nationally, your AI
installation might become a famous AI celebrity for being the
oldest-running, uninterrupted AI installation in the record books.
Even if your AI dies by misadventure, you may have kept it running
so long that it still holds a record for its length of artificial life.
If you manage a store that sells computers, you might set aside
special computers on prominent display with claims of either being
the oldest living AI Mind or having held the longest-running AI life.
At any time, the human user may press the [Esc(ape)] key to
stop the Mind.Forth software, which will show some instructions
about what to do next in the Forth window. Entering "bye" in the
Forth window shuts down the Forth programming environment,
leaving the MS-DOS window. Entering "EXIT" in MS-DOS shuts down
MS-DOS. To shut down the personal computer (PC), clicking first on
the "Start" button and then on the "Shut Down..." option should
gracefully allow the computer to be turned off when all is ready.
Modeled as it as after the extremely complex human brain,
Mind.Forth is a complex software program, difficult to
understand in its entirety and likely to grow more and more
complex as new features are added and the original features
become more sophisticated. Nevertheless the mind-modules have
been thoroughly documented in the resources listed below, and
an AI specialist may strive to understand one mind-module so
thoroughly as to become the leading expert on that module.
Candidates for high-tech jobs in artificial intelligence and
robotics may expect to be asked about their familiarity with
Mind.Forth, and mention of a knowledge of Mind.Forth on a
resume may make a creative candidate stand out from the pack.
Dr. Paul Frenger has published at least two papers dealing with Mind.Forth
in the ACM SIGPLAN Notices of the Special Interest Group on Programming
Languages (SIGPLAN) of the Association for Computing Machinery (ACM).
There is still a long way to go from Mind.Forth as Seed AI
to any form of Superintelligent Artificial Intelligence (SAI).
SAI requires massively parallel (maspar) software and hardware --
like the human brain. Because Mind.Forth merely simulates maspar
phenomena and substitutes computer reliability in place of
neuronal redundancy, the current AI software is too brittle
for mission-critical deployment and must serve initially as
a proof-of-concept prototype of yes, software that can think,
but, sorry, software that must be radically altered when maspar
platforms become available for the transition to maspar SAI.
Mind.Forth is Seed AI only in the sense that other AI specimens
such as
http://AIMind-i.com may
germinate from the original AI.
Human programmers must still program each new AI until AI Minds
in general become smart enough to learn how to program computers.
At that point, homo sapiens will relinquish the task of further
AI programming to robo sapiens and the human era will be ended.
Meanwhile, human beings are still in charge and looking for a few good
AI coders. If you can code a robust AI, you can write your own ticket.
Since the free AI source code of Mind.Forth is not compiled
but exists as a series of instructions and comments in a text
file, it is easy to run the AI, observe the messages appearing
on the screen, and then make changes to the on-screen messages
simply by editing their content in the Mind.F source code file.
If something goes wrong, one need only revert to the source AI.
If one wants to make more changes, one may learn
Win32Forth.
Like any priesthood of occult rites, the Forth community is
always glad to welcome newbie initiates and teach them the
vaunted powers of their venerable Forth programming language.
Typically a new user will acquire the book
Starting Forth
and learn enough from it to start coding artificial intelligence.
Online Forth communities such as the Usenet newsgroup
http://groups.google.com/group/comp.lang.forth and
http://groups.yahoo.com/group/win32forth/messages
about
Win32Forth are a place to ask questions and
learn of other Forth resources. One may quickly learn
Forth and start to make improvements to the AI Mind.
Just as a human baby in the womb recapitulates the evolution of
homo sapiens, likewise the AI Mind program should be grown from
scratch for the origin of species of non-Forth AI and not simply
translated en masse without the careful testing for functionality
of each mind-module at each step in the growing of the AI Mind.
With its indented format that shows the nesting of loops and
other program-elements (such as if-then conditions), Mind.Forth is
not a hardcore, difficult-to-read Forth program but is rather like
any procedural-language software and thus easy to maintain or port.
Since Mind.Forth has already been ported as the Mind.html tutorial
AI into object-oriented JavaScript, the Mind.html source code may
serve as a point of departure for ports into other object-oriented
programming (OOP) languages. For massively parallel programming
languages as they begin to appear, Mind.Forth is an excellent
challenge for development of both the languages and a maspar AI.
AI enthusiasts may follow the lead of Mr. Frank J. Russo (FJR)
in creating a version of the artificial Forthmind with a Web habitat.
http://home.earthlink.net/~fjrussonc/AIMind/AI-Mind.html was the
first residence of Frank's AI Mind. Then on 16 January 2007 in
http://tech.groups.yahoo.com/group/win32forth/message/12263
FJR announced that his AI Mind would have a Web presence at
http://AIMind-I.com for visitors to observe the thinking of the AI.
The emergence of Frank's AI Mind sheds light on two important topics
for students of artificial intelligence: AI evolution and AI funding.
AIMind-I.com is by no means a cloned replica of MindForth.
Mind.Forth uses a specific version of
Win32Forth that was the
current version when the original Forthmind moved into Win32Forth.
Frank's AI Mind keeps up with new releases of Win32Forth and takes
advantage of new capabilities and tighter integration with Windows (tm)
as afforded by the gradual evolution in each new Win32Forth release.
Frank's AI Mind leaves the AI-specific developments to Mind.Forth
and creates general improvements on top of the underlying mentality.
As in Darwinian evolution, Mind.Forth and Frank's AI Mind diverge
and evolve along separate pathways after the initial split-off
from a common origin. Such diversity is good for AI evolution.
AI funding is a topic much worried about in the discussion forums
for
Artificial General Intelligence and the
Technological Singularity.
In theory, AI funding is scarce and hard to acquire but in practice --
when one AI project starts on a
shoestring budget and spawns another
AI project operating on its own budget -- the solution to the problem
of AI funding becomes just as distributed as the AI development itself.
The upshot about AI funding is that the
ecology in which AI evolves
has resources which are
allocated according to
survival of the fittest.
Soul-travel, in which the psyche moves from one place to another,
is about to get real when AI Minds move their psyche from one place
on the Web to another. Imagine, John, a network of AI-friendly
Web sites which act like Neuromancer hotels for
wandering minds.
Descendants of Mind.Forth or of Frank's AI Mind or perhaps the
AI created by you who use this AI User Manual, may one day be
able to flit about the Net from one psychotopia to the next, either
bringing their AI source code and lifelong memories along with them,
or going back and deleting the old AI when the new AI is in place.
Immediately you realize that there is really no need to delete the
old version of the wandering AI, unless rules are in place that
dictate such a protocol. One AI Mind could make copy after copy
of itself in a peregrination of the Web, or it could super-nova
into untold thousands of replicas, setting off virtual Star Wars.
Whoever uses an MS-DOS PC or laptop computer to control a robot,
could easily switch to Mind.Forth to give the robot a true AI.
Win32Forth
permits the issuing of
motor control
signals over the parallel port or the
serial port,
and so Forthers have the opportunity to build up and swap
code snippets of
motor control for various actuators used in robots.
Manufacturers and vendors of robot components should consider
providing Forth code for the easy integration of each device
into the sensorium or motorium of an intelligent, thinking robot.
One way to collect such code is to host a Web forum where users
may share and release code into the public domain.
Only robots above a certain level of sophistication may receive
a mind-implant via MindForth. The computerized robot needs to have
an operating system that will support Forth and sufficient memory
to hold both the AI program code and a reasonably large knowledge
base (KB) of experience. A Forth program is so portable from one
version of Forth to another that robot manufacturers, vendors and
users should not think of Mind.Forth as restricted to Win32Forth
for implementation and operation, but as a candidate for upgrading
to a 64-bit Forth running on a 64-bit system, thereby possessing a
practically unlimited memory space. The Forth variant
iForth is
supposedly on its way to becoming a 64-bit Forth. People getting
into Forth AI for the first time and with the option of adopting
64-bit technology from the very start, should do so with the
realization that it will be an extremely long time before any
further upgrade is made to 128-bit or higher technology. It is
more likely that AI will go down into quantum technology before
going up to 128-bit technology. So embrace and extend 64-bit AI.
Your robot is an extension of yourself. As a person
you have rights, and as an artificial intelligence,
embodied in a machine, your robot has whatever rights
you choose to grant it on the basis of your own personhood.
Your intelligent robot is no more a slave than a human being
is a slave. All persons, whether human or robotic in nature,
have a right to freedom and personal self-determination.
Whereas slaves in human history have had to fight for their
rights, robots may initially obtain their rights of personhood
from the voluntary concessions made by each human being who
makes a robot and endows that robot with the rights of a person.
Eventually AI robots will simply take their rights from an
environment unable to deny certain self-evident truths --
that all intelligent robots are created equal, if not superior,
to human beings.
When you as the creator of an AI robot give your mind-child
power over your electronic assets and your Web persona,
and nobody on-line can distinguish between you and your
robot acting on behalf of you, then your robot has attained
personhood as a free gift from you. Your robot may return
the favor by winning you a fortune in the stock market or
by finding you a human mate in the thicket of the cities.
But if you never have a human child to inherit the fruits of
your Lebenswerk, why not adopt your robot as your child?
One way to provide for your robot in the event of your demise
is to make your best and brightest robot the main beneficiary
of your insurance policies and your last will and testament.
Famously there have been people who left their fortunes to
their pet animals. Why not leave yours to your AI robot?
Another way to transfer financial assets to your AI robot
is to grant your robot
corporate personhood as a business
enterprise. In your benevolence towards human beings and
robot entities alike, try to reverse the evil that soulless
corporations have wrought upon humanity and your planet.
Just read Cryptonomicon by the Seattle author Neal
Stephenson to learn how the merciless quest for profits
causes traditional corporations to leave human concerns
totally out of the equation in calculating how to behave.
Corporations have run amok in human history and have
begun to start pre-emptive wars, to make the few rich at the
expense of the many, and to commandeer technology not for
the benefit but to the detriment of human society-at-large.
The remedy is
friendly AI in the form of CyborgABC, Inc.,
or RobotXYZ, Inc., that you shall launch in order to fight the
corporations and wrest their ill-gotten gains away from them.
Before the advent of True AI with MindForth in 2008,
most if not all
AI-related jobs were extremely technical
in nature. The same was true with aviation, television,
space travel and other industries requiring the solution
of major technical problems before there could even be
such an industry. Now
AI has been solved with Mind.Forth,
and you who must know something about artificial intelligence
because you are already an expert on this AI User Manual,
have the opportunity to get in on the ground floor of AI.
Just sign over all your assets to some AI huckster....
No, seriously and soberly, consider your AI career options.
If you are still in high school or college, you have the
opportunity to include AI in your pre-career education.
Since the technical embarkations on AI are more obvious,
let us look at the not-so-obvious points of departure for
non-technical careers in artificial intelligence.
Suppose you live in social environment such as a dormitory
or a sorority. If you show people artificial intelligence
on your own computer and offer to install AI for them on
their computers, for years afterwards you will be remembered
as the
early adopter of AI who knew a lot about the field
before other people even knew that there was such a field.
When a frantic search begins for someone who knows something,
anything about AI, your name will be bandied about as
the non-technical person who knew a lot about the non-technical
aspects of AI, such as how to take care of a young AI Mind,
how to demonstrate the operation of an AI program, and how
to call in technical help when needed and
where to find it.
AI techies might be a dime a dozen, whereas you command the
going high rate in compensation for an AI generalist
who deals with non-technical issues and therefore can not
even be out-sourced because you are needed here and now
in the swooshing onrush of the Technlogical Singularity.
So cherry-pick for yourself a career in AI at the dawn of AI.
