Frequently Asked Questions about Mentifex artificial intelligence

English and Russian bilingual ghost.pl AI in Perl

Mens Latina in Latin -- MindForth for Robots


Contents

1. Introduction

2. Bibliography: What books has Mentifex written about AI?

3. Psychography: What minds has Mentifex created for Strong AI?

4. History of the Mentifex AI project

5. Latest News about Mentifex

6. Implications of Mentifex AI

7. What is the likely future time-line of Mentifex AI?

8. Self-modifying code for AI-Go-FOOM!

9. How can I use Mentifex AI to write my thesis or my dissertation?

10. Ways to help or get involved in Mentifex AI

11. Proto-Index of Mentifex Autobiography


1. Introduction

Q. Who was Mentifex?

A. Mentifex made the Ultimate AI guaranteed to think and reason with logical inference in English, Latin, German and Russian, demonstrating a solution to the AI-hard problem of Natural Language Understanding. Mentifex was the author of the Mentifex Theory of Consciousness.


Q. What was the education of Mentifex?

A. Mentifex attended various schools, including South Margarita Elementary School in the Panama Canal Zone; Blanchet High School in Seattle WA USA; the University of Washington (B.A., Latin and Greek classics) in Seattle; U Cal Berkeley; U.S. Army Electronics School at Redstone Arsenal, Alabama; Nuclear Training Directorate at Sandia base, New Mexico; and the University of Goettingen in Germany.

Q. Why did Mentifex claim to have released the First Working Artificial General Intelligence?

A. Because the Mentifex AGI Minds started thinking primitively in 2008, they were immediately the First Working AGI but their performance was not at all impressive. Ih the following years, the AGI Minds expanded outwards from their initial core functionality. They were still based on brain-like concepts, not statistics. They still exhibited machine learning like a human child learns, and not by a mathematical massaging of enormous data-sets. They began to approach machine consciousness with self-referential thought. Using the SpreadAct module for spreading activation from concept to concept, the Minds in Perl and in JavaScript and in Forth became capable of processing more and more kinds of questions and ideas, such as "who are you" and "what do you think". Since the Perl AI could think in English or in Russian, ghost.pl became an experiment bed for new ideas in machine translation (MT).


Q. Why did Mentifex say that AI has been solved?

A. It was a way to invite Netizens to try to disprove an Extraordinary Scientific Claim (ESC). Curiously, it was also the basis for the Reddit user name AiHasBeenSolved which could be maliciously downvoted by brigades of Mentifex-bashers but still carry a message so intriguing that Netizens would be eager to read the downvoted message.


Q. What would be a typical test of Mentifex AI?

A. Suppose we tell the Perl AI two things:

  • John has a house
  • John has a car
  • Suppose we then ask the AI:
  • What does John have
  • We tried it just now and the ghost292.pl gave us the following output:

    THE JOHN HAS THE HOUSE AND THE JOHN HAS THE CAR I UNDERSTAND YOU

    When we tell the Mens Latina AI "marcus est professor" ("Marcus is a professor"), the artificial intelligence in Latin language asks us "Scribitne Marcus libros?" ("Does Marcus write books?"). Our answer in Latin causes the retroactive adjustment of the knowledge base (KB) with the true fact that Marcus either does or does not write books. Mens Latina is a free resource for the study of Latin, artificial intelligence and JavaScript.

    When we tell the Dushka AI "марк студент" for "Mark is a student", the artificial intelligence in Russian language asks us "Марк ли читает книги?" ("Does Mark read books?"). Our answer in Russian causes the retroactive adjustment of the knowledge base (KB) with the true fact that Mark either does or does not read books. Душка is a free resource for the study of Russian, artificial intelligence and JavaScript.


    Q. What was Mentifex-bashing?

    A. http://github.com/kernc/mindforth/blob/master/wiki/MentifexBashing.wiki


    Q. Was Mentifex on Twitter?

    A. No. Somebody created a fake Mentifex Twitter account. However, Netizens desirous of Search Engine Optimization could make periodic tweets about Mentifex AI and come to the attention of persons and agencies tracking all mention of #Mentifex.


    Q. What was the Mentifex Effect?

    A. To his shock and horror, Mentifex discovered that IRL (in real life) persons who verbally insulted or physically assaulted Mentifex received an unwelcome visit from the Grim Reaper or minions of same. A first example occurred when teenage Mentifex (ATM) was living with his family at their 11056 address. One afternoon ATM sat casually outside with his older brother LJM and friends of LJM. One of them, cigarette-smoking Mike Spillane, suddenly said to ATM, "Do you want to see a match burn twice?" and he swiftly (physically, not mentally) struck a match against a book of matches and plunged the burning match into the exposed skin of the hand of ATM, searing a Reaper-degree burn into the back of ATM's hand. In shock, ATM did nothing about it, but several years later ATM inquired of LJM about the assailant and learned that he had died. Another time Mentifex went out on a date with the Commissar Lady from his university Russian class and they saw "2001: A Space Odyssey" at the Cinerama Theater. Upon arrival back at the family home at 11810 late at night, ATM could not get into the house because he had forgotten to carry a house-key. While he was trying various windows and doors to gain entry to where he lived, the next-door neighbor drove up an easement driveway and parked at his own house out of sight. Suddenly the neighbor and the boyfriend of his daughter ran up to Mentifex and the boyfriend began manhandling Mentifex and roughing him up against a wall of the house. Mentifex spoke German to the neighbor, who called the police to come and investigate the situation. A few years later, Mentifex did read in the newspaper that the bully boyfriend had perished in a raging river up in Snohomish county, caught in some weird configuration of rocks at the bottom of the torrent. Another time Mentifex (ATM) went to a Democratic primary caucus to promote Jimmy Carter for president instead of Henry Jackson. ATM wanted to ask if a representative at the state convention would have "carte blanche" to endorse a desired candidate, but ATM decided to speak the phrase the English way and not the French way, which might sound arrogant. Immediately another person, attending the caucus along with his wife, corrected ATM's pronunciation by saying "carte blanche" with the French accent which ATM had known about but not dared to use. When ATM went to another caucus four years later and the guy's wife announced that he had passed away, Mentifex felt a pang of sorrow and thought to himself, "Oh, Lord! All he did was correct my pronunciation; he did not deserve to die." A further, extremely weird example occurred when ATM's older brother LJM passed away and the newspaper reported that a former student who had heaped extreme verbal abuse upon ATM died the same day as ATM's brother. From then on, whenever Netizens insulted Mentifex on the Net, he would be seized with horror and panic at the thought that the insulting Netizen was now in grave danger vis-a-vis the Grim Reaper for insulting Mentifex. Gradually, however, it became clear that only IRL (in real life) abusers of Mentifex were in danger of karmic retribution from Doctor Reaper, and Mentfex was able to go back to enjoying the wit and cleverness of some of the insults without fearing for the karmic catastrophes of Real Life.


    Q. Why did Mentifex complain about brain damage from Husky Football at a December 2019 meeting of the Board of Regents of the University of Wasington?

    A. In August of 2019 Mentifex was reading the national edition of The New York Times at a Starbucks coffee shop in room 101 of the Suzzallo Library on the campus of the University of Washington. An article about sub-concussion brain-damage from football upset Mentifex so much that he went to a photocopy shop on the Ave and made 107 doctored photocopies of the article with an imposed headline of "Hypocrisy of UW Huskies football" and he pamphleted the campus with anti-football fliers for the next several days. A few months later, the Husky football coach quit his job -- perhaps out of guilt and shame after reading one of the fliers from Mentifex. Then on 2019-12-17 Mentifex was again in the Suzzallo Library and saw an announcement that the Regents were meeting all day long up in the Petersen Room. Mentifex signed up to speak to the Regents and he meanwhile made thirty new photocopies of his anti-football flier, which was incorporated into the official minutes of the UW Board of Regents. The governing body of the university did not seem to care about the objections of Mentifex to Husky Football, but the coronavirus epidemic made it hard for any university to conduct football at great risk to students, workers and spectators.


    Q. Why was the Republican senator Mitt Romney such a hero to Mentifex?

    A. When it was looking like America as a democracy was all washed up and the American dream was over, the great American senator Mitt Romney took a courageous stand on principle and kept the last flicker of American greatness from dying out. When every single Republican senator except thanks-be-to-God Mitt Romney came out as a corrupt and spineless coward, His Heroic Greatness Mitt Romney spoke out for truth, justice and the American way. When American mothers could not hide from their children the fact that one sexually assaultive monster was in the White House and another was on the Supreme Court, the shining example of honest and courageous senator Mitt Romney gave the American people hope that there was still a single shred of decency among the otherwise rotten Republican senators. When Americans encounter a tough moral situation, they must find noble guidance by asking themselves, "What would Mitt Romney do?"


    Q. Why did Mentifex code his main, concept-based artificial intelligence in Perl, a moribund language being replaced Web-wide with Python?

    A. Perl was chosen mainly for the sake of Unicode -- "a computing industry standard for the consistent encoding, representation, and handling of text expressed in most of the world's writing systems." Thanks to Unicode, Mentifex was able to release the original Perlmind which could think in both English and Russian, and which Mentifex was naively offering as a free gift to the ungrateful Perl community, who, instead of greeting the newcomer Mentifex with hospitality and brogrammer brotherhood, and with gladness that Mentifex was putting so much energy into possibly a revival of moribund Perl, au contraire, mon frere turned savagely against Mentifex and vilified him with vile epithets that said more about the disreputable characters infesting the Perl subReddit than about hard-working, self-sacrificing Mentifex. Not thinking for themselves but, in the manner of bound-for-prison Trump supporters, accepting and perpetuating old aspersions against Mentifex, the guardians of Perl tried to keep curious Perl coders from going down the "rabbit hole" of Mentifex artificial intelligence, deleting and denigrating the humble contributions that Mentifex was trying to share with the Perl community. With his childlike sense of trust and wonder, Mentifex was hoping to make Perl the epicenter of true, concept-based AI with Natural Language Understanding (NLU) and to create glamorous new AI jobs for expert Perl programmers. Mentifex, before he was deleted, banned and practically burned at the stake, was imagining Fortune 500 companies hiring fifty Perl programmers to each work on one of the fifty AI mind-modules in Perl. Mentifex envisioned every nation on Earth adopting Perl with Unicode for artificial intelligence in their own national language. Mentifex anticipated governments rushing into Perl artificial intelligence to keep up with other nations in the AI arms-race towards a Technological Singularity. But if the so-called Perl community does not want Mentifex, Mentifex must shake the dust from his shoes and wander the Web in search of a friendly community more accepting of new ideas.


    2. Bibliography: What books has Mentifex written about AI?

    Inexpensive, easy-to-read paperback and Amazon Kindle e-books were
    a way to spread the memes of free, open-source Mentifex Strong AI.

    AI4U was about the Mentifex AI state of the art in 2002.
    ai4ubook.html provides material beyond the 2002 AI4U textbook.

    The Art of the meme told the history of the Mentifex AI project.

    Singularity (in the Psyborg series) was AI science fiction.

    InFerence was about automated reasoning with inference.

    Artificial Intelligence in German is about Wotan German AI.

    Artificial Intelligence in Ancient Latin is about the Mens Latina AI.


    3. Psychography: What minds has Mentifex created for Strong AI?

    http://ai.neocities.org/Abracadabra.html -- Mens Latina in ancient Latin

    http://ai.neocities.org/mindforth.txt -- MindForth in English

    http://www.nlg-wiki.org/systems/Wotan -- in German

    http://ai.neocities.org/Ghost.html -- in English

    http://ai.neocities.org/Dushka.html -- in Russian

    http://ai.neocities.org/perlmind.txt -- ghost.pl in Perl


    4. History of the Mentifex AI project

    http://cyborg.blogspot.com/2011_05_16_archive.html#1

    An Amazon Kindle e-book, The Art of the Meme, is actually a history of the Mentifex AI project.


    5. Latest News about Mentifex

  • New AiMind thread on Chatbots.org Forum
    http://www.chatbots.org/ai_zone/viewthread/3750 -- AiMind discussion.

  • 2020: Mentifex began claiming to have "AI Guaranteed To Think".
    In the belief that the various AI Minds could think, albeit primitively, in English and in Latin and in Russian, Mentifex began stating that claim as a memetic ploy in many venues. The idea was to get Netizens to test each AI Mind so as to attempt to disprove, debunk and refute such an outrageous claim. The memetic ploy was similar to offering a "bug bounty" on Mentifex AI software, which could not yet be done, because the AI software was far from not quite perfect. But each AI was indeed thinking by having built-in concepts and by associating from concept to concept to generate thoughts and ideas under the control of a linguistic superstructure.


    6. Implications of Mentifex AI

    http://sarudev.tiddlyspot.com mentions the AI Mind project.

    http://www.google.com/patents/US8620890 is a patent that cites Mentifex MindForth.

    http://publik.tuwien.ac.at/files/PubDat_221361.pdf -- mentions the MindForth AGI architecture.

    http://papa.motd.org/cave/chatbot/aimind/prolog.org -- wants to write an AI better than Mentifex.

    http://lj.rossia.org/users/sadkov/111836.html -- discusses Mentifex Russian AI.

    https://github.com/paule32/mind -- as described in
    https://groups.google.com/forum/#!topic/de.sci.informatik.ki/Wp7Ioe8DyLM

    http://agi.topicbox.com/groups/agi/Tcad0819005caf100 -- discussion on AGI mailing list.


    7. Future time-line

    Q. What is the likely future time-line of Mentifex AI?

    A. We can view the future development of Mentifex AI in stages. As of 2019, in Stage One the independent AI scholar Mentifex develops the open-source AI Minds in Perl for webservers, in JavaScript for tutorial purposes, and in Forth for embodiment in robots. Any one of the prototype AI Minds could lead to a break-out phenomenon where parties other than Mentifex either translate the existing AI codebase into their own version of AI for further development, or adopt and adapt one of the pre-existing AI Minds for their own in-house development of AI. It is also possible that a non-Mentifex developer of AI will simply borrow interesting ideas from Mentifex AI. A further possibility in Stage One is that various entities like Russia or China may in secret be developing one or more of the Mentifex AI Minds in a major effort to present the world with a fait accompli of advanced AI technology.

    Stage Two is a scenario where entities other than the originator Mentifex latch onto one of the AI Minds and fork its further development into their own branch of AI evolution. These forking entities with their "Aw, fork!" attitude may be individuals working artistically in a garret or dweebishly and nerdily in a basement, coding AI day after day into the night and into the history books. Imagine a community of such people, sharing tidbits and brag-bits like the otaku zoku of Japan, boasting about some new feature that they add to their AI, or some form of robot embodiment, or their venture into quantum computer AI coding or biomass neuronal architecture. Not too likely, perhaps, but consider the other entities who create AI not in garrets or basements but in high school or university laboratories, or in corporate workspaces, or in government facilities.

    A corporate IT (information technology) department could establish an AI bunker apart from all other corporate functions and staff it with an initial team of department head, software development staff, AI safety oversight committee, and robotic embodiment personnel. The race is on. No one can predict here whether such AI efforts will remain open-source like the original AI Minds, or be released publicly, or be described in scholarly papers at public conventions. But all such corporate AI efforts will be subject to the same law of evolution: Survival of the Fittest.

    Stage Three is a Technological Singularity which by its very definition defies description.


    8. Self-modifying code for AI-Go-FOOM!

    Q. How can we make a mentifex-class AI that replicates and enhances itself?

    A. Certainly with the ghost.pl AI in Perl, and possibly also with MindForth, especially in Perl it is easy to write code that will read in its own program as a mind1.pl file, insert any new words or new ideas into the MindBoot sequence, write the expanded AI as a mind2.pl file, schedule the mind1.pl to die soon, and start the mind2.pl running as an enhanced AI ready to replicate itself and enhance itself even more.


    9. How can I use Mentifex AI to write my thesis or my dissertation?

    Q. Can I do research on Mentifex AI as such?

    A. Yes. Various points of entry into a scholarly treatment of Mentifex AI include:

    • The history and development of Mentifex AI;
    • Original research taking Mentifex AI beyond its initial offerings;
    • Embodiment of Mentifex AI Minds in autonomous humanoid robots;
    • Refutation or confirmation of theoretical issues in Mentifex AI.


    10. Ways to help or get involved in Mentifex AI

    Q. Can I join the Mentifex project?

    A. Yes, inasmuch as the Mentifex project is a loosely knit collaboration of individuals all working towards the same general goal of Artificial General Intelligence. You may Join the Discussion at http://agi.topicbox.com/groups/agi with your e-mail address and you may either lurk or participate actively in the Darwinian process of Survival of the Fittest as applied to the creation and evolution of AI Minds. You may work on your own branch of AI evolution or you may specialize in one or more mind-modules common to almost every species of AI Mind. You may finance your work and monetize your efforts by hiring out as an AI Mind Maintainer or by writing books, blogposts, articles, user-manuals and other paraphernalia deontic to the Mentifex AI project. As you become an expert on Mentifex AI, register with a speakers' bureau and go on the lecture circuit for talks and consultations where you share your expertise. If you are of the one gender-persuasion and you want to attract the other, casually mention amid the tinkle of juice-glasses that you know all there is to know about AGI and gush with enthusiasm when your fellow AGI-Summit attendee or happy-hour inebriator widens the eyes in amazement at your discreet charms of the non-bourgeoisie. Watching you win like the champion that you are, let the luzers mutter, "What does (s)he have that I haven't got?" in their ignorance of the winner-take-all aspects of the race towards AGI. In your own work on democratized, open-source artificial general intelligence, you can
    • Port Mentifex AI into a new programming language.

    Q. How can I support the independent-scholar Mentifex project financially?

    A. You can purchase a copy of Artificial Intelligence in Ancient Latin from Amazon as both a donation of support for Mentifex AI and as an investment in what could become a rare, collectible book in the future. It is available from Amazon in various countries:

  • http://www.amazon.ca/dp/B08NRQ3HVW -- Canada;
  • http://www.amazon.fr/dp/B08NRQ3HVW -- France;
  • http://www.amazon.de/dp/B08NRQ3HVW -- Germany;
  • http://www.amazon.it/dp/B08NRQ3HVW -- Italy;
  • http://www.amazon.co.jp/dp/B08NRQ3HVW -- Japan;
  • http://www.amazon.es/dp/B08NRQ3HVW -- Spain;
  • http://www.amazon.co.uk/dp/B08NRQ3HVW -- United Kingdom;
  • http://www.amazon.com/dp/B08NRQ3HVW -- USA.
    You could also write a review on Amazon. The very title of the book will come as a shock to many AI enthusiasts because the very idea of coding a True AI in a dead language like Latin is unheard of in the modern world. You could also help by giving your copy of the provocative book to your local book store and by asking them to list the unusual book in their on-line catalog. Hodie Roma, cras Mundus!


    11. Proto-Index of Mentifex Autobiography

    Aardvark
    ---- Lake Goodwin mylife20
    ---- trip to San Francisco -- mylife21
    Eccles, Sir John Carew, Nobelist
    ---- correspondence with -- mylife33
    Hubel, David, Nobelist
    ---- letter from -- mylife33
    Love, First
    ---- brother of -- mylife18
    ---- cover girl; R.N. -- mylife27
    Love, Second
    ---- Husky Hollow mylife15
    ---- volunteer for Senator Barry Goldwater mylife18
    ---- intro to Grandma; phone contact mylife20
    ---- Commissar Lady mylife21
    ---- birthday present (sewing scissors) mylife24
    ---- rejection by; father of -- mylife25
    Love, Third
    ---- sudden question about Second Love mylife20
    ---- cover girl -- mylife27
    Love, Fourth
    ---- sketchy details; R.N. mylife27
    Mind, Theory of
    ---- AI breakthrough -- mylife32
    Moravec, Hans
    ---- Mind Children mylife01
    Naiden, James Richard
    ---- Latin SAT test -- mylife17
    Nixon, Richard Milhouse
    ---- chance meeting with -- mylife16
    Nuclear Training Directorate
    ---- study at -- mylife23
    Parr, Larry
    ---- together in freshman English -- mylife18
    School of the Americas
    ---- going to the dentist -- mylife08
    Stark, Detlef
    ---- Germanicisms -- mylife20
    ---- trip to San Francisco with Aardvark mylife21
    Turner, Floyd, "The Doukhobor"
    ---- chance meeting with -- mylife19
    ---- burning American flag -- mylife21
    von Hagen, Dmitri
    ---- Sea-Tac airport pick-up -- mylife26


    Thinking in Latin
    Artificial Intelligence in Ancient Latin
    "Look Inside" Kindle version to read initial chapters.


    Return to top; or to
    Older Mentifex FAQ on Advogato.
    Original Mentifex FAQ on Usenet.
    Many thanks to NeoCities.
    http://neocities.org/blog


    Website Counter