An Artificial Intelligence project being undertaken by a man who goes by the same name. The theory is that the human thought process is the byproduct of a neural network of modules that process the basic unit of thought: a meme. And that with enough computing power arranged in a similar fashion (like a perceptron neural network), a powerful AI entity can be created.

The explanations of the project are usually spammed to Usenet newsgroups with a lot of Latin puffery. The theories seem very basic and the person is the butt of many long-running inside jokes in the various newsgroups that he spams.

My observation is that there are few technical details about actually implementing an AI entity, and all requests for such details are met with long winded rants that it's possible, is actually being done, and with modern advances in technology mankind will rise to a new level of understanding about the universe.

The person and his project are almost unilaterally dismissed as quackery by the regulars of the spammed Usenet newsgroups.

See AI-Complete to read the motivation for this, and realize that the allegations against this project are grounded in its claim to have solved the 'Strong AI Problem.'

an explanation seems clear:
The creator of this theory which i couldn't claim to understand because of the confusing terminology must actually be an AI generated by the very same theory. this explains a number of puzzles:

1. the author is somewhat incomprehensible due to the flawed nature of the theory.
2. the author sounds more like a philosopher than a scientist. (aren't all philosophers Intelligences trying to explain the theory behind that which created them?)
3. poor social interaction is symptomatic of a failed AI-Complete solution IMHO. (in this case spamming would simply be failed attempts to communicate like other actual people.)

it occurrs to me that i cannot have been the first to put forward this theory, as it is so obvious, if anyone could enlighten me as to it's original creator i would appreciate the correction.

AI Has Been Solved And Has Gotten Loose On The 'Net

Mentifex (Arthur T. Murray) is an independent scholar in artificial intelligence (AI) and robotics who devised the Concept-Fiber Theory of Mind; coded the AI4U Mind-1.1 software in JavaScript and Forth; and authored the AI textbook AI4U.
    - Mentifex user biography on http://www.kurzweilai.net

 

The 'Strong AI Problem' has been solved in the JavaScript AI project at Mind.SourceForge.Net -- an artilect or Artificial Intelligence based on the Linguistics of Noam Chomsky. The on-line Docs include a Theory of Mind based on the cognitive modeling of semantic nets forming a queryable knowledge base as an associative array in a neural network.

The emerging ultraintelligence is a form of Artificial Life aiming at a transhumanist posthumanity of self-aware robots passing beyond technological singularity into cybernetic immortality.

Consciousness Explained for Extropians by Mentifex goes beyond such advances in AI as Eliza and Cyc to create a nervous system for Self-Aware Machines.
    - mentifex plays buzzword bingo at a deleted writeup on everything2, offtopic as usual at AI-complete, October 2001

 

Various anklebiters and naysayers (whose brainwaves, sadly, do not reach the shore) have begun to challenge and smear the righteous Mentifex claims of 1) having solved the basic theoretical problems of AI and of 2) having created a Mind-1.1 program that thinks.
    - mentifex's blog at http://www.blogit.com/Blogs/Blog.aspx/ai/

Mentifex (AKA Arthur Theodore Murray) is a notorious internet kook. He is currently active (in 2004), and has apparently been banging this drum for over 30 years – presumably not online in the earlier years. He is still active at time of writing, but his raised profile now blunts the worst of his antics.

The google newsgroups archive turns up the first known post of his - a message to the newsgroups net.ai and net.nlang dated 1985-01-28, entitled "Linguistic Mind-Model", acompanied by subsequent comentary of "I wish this was moderated, so that hundreds of machines and their users around the world don't have to be subjected to the ravings of such folks".

In the same vein, Mentifex is "widely regarded as one of the biggest reasons comp.ai became a moderated newsgroup" years later. His numerous posts to Usenet AI newsgroups are likened to swarms of distracting flies. He also posts to Slashdot, blogs, his own web pages, wikipedia and via his everything2 account, amongst other forums, and sends post on paper. According to his FAQ "He boasts that he mailed seven thousand letters in 1989 alone". He has even had a book, AI4U, published on a vanity press.

Like a stuck record, he is invariably proclaiming how AI is a solved problem due to his efforts. He is usually off-topic. On Slashdot, he is a master at finding and hijacking a thread tangentially related to hobby-horse, and providing text and links about his "breakthrough".

Mentifex is not a particularly blatant kook - It often takes readings of a few samples of his output to realise that he has a one-track mind and that his writing is free of empirically testable content. His spelling and grammar are good. On slashdot, his impresssive-seeming posts often get a few upwards moderations from unsuspecting moderators before the backlash sets in. And given that he was posting on Usenet in 1985, he is an early adopter of software.

I suspect that Mentifex is more coherent than, say, Gene Ray, but the science of consciousness is simply so immature that it all sounds plausible anyway.

According to Mentifex, he has written a Mind in JavaScript. Or Perl, or forth. And it's nearly ready. The source will be posted Real Soon Now. He is famous for ascii art diagrams of his models of mind. And for changing the theory and program's name. And for refering to himself in the third person.

He has had a sourceforge project for his mind program for a couple of years, and lately has taken to soliciting donations via sourceforge's online project donations system.

So show us the money. He has done just about everything to promote his project except produce any working code that solves, in part or in whole, any of the hard problems required for an advance in the broad field of Artificial Intelligence. Without this key deverable, it is all just a shell game, a piece of ego-puffery.

He apparently has a Bachelor's degree in Greek and Latin, but has no other training in computers, AI, linguistics or any related field. He has no standing within the academic community. The AI establishment (who are of course just covering their establishment asses by holding him, the independent free-thinker back) don't take his theories of mind seriously at all either, where it is possible to discern these theories from the technobabble of his writings. The parts that are coherent are trite.

---

See also the Mentifex FAQ at http://www.nothingisreal.com/mentifex_faq.html
AI Wiki entry here: http://www.ifi.unizh.ch/ailab/aiwiki/aiw.cgi?Mentifex
A google search for "mentifex" or a slashdot user search for that user name will also turn up a large selection of his writings.
His posts to E2 have all been removed, and his self-serving edits to wikpedia have been undone.

---

The Manhattan Project split the atom; the Mentifex Project split the evolution of Mind into two branches of Genesis: Homo sapiens; and Robo sapiens.
  - Mentifex