Perl AI Mind Guy

Can you gays disprove his theories formally, because otherwise you look like a bunch of school bullies, beating the next Albert Einstein.

can you gUys

self fix.

It was fine the other way.

goys

Consider this: A pack of wild ANDRUs.
Sentient, strong AI ANDRUs nearing your university. Trampling your academia. Raping decades of research.
And you can't do shit since they're sentient. The Mind Creator grabs your chatbot and brings her to life with his Perl skills.
The primal ANDRUs finally dominate your household. They read intelligent books written by independent AI scholars and you are forced to witness the Technological Singularity.
Such is the downfall of AI academics.

And then Marisa said to Alice while having shoved up in her mouth half of Alice's foot: "Mughmmmmm hummmgh muhhhum ughmmmmmm".
It is unknown if Alice, while being gagged, blindfolded and tied up had any enegry to understand what Marisa said, all she did was to unleash over and over muffled sounds, due to being unable to say anything.
Later on, Marisa, after had enough fun with Alice's foot, she removed it slowly from her mouth while leaving a bunch of saliva behind. Marisa made a sound like she was reborn again, after she had drawn all the energy of Alice's sole.
"Hey, Alice!", said Marisa in a playful voice. Marisa removed the blindfold of Alice at once and then started undressing herself. After removing her bras Alice noticed that Marisa had no breasts at all!
"Do you know what this means Alice?", Marisa said while slowly removing her underwear.
"Da-dah!", Marisa said while a medium-to-small sized snake with retarded-looking eyes unsealed itself off Marisa's panties.
"I was a boy all along, alice!", Marisa said to her while watching Alice's scared and wondering face. After that event, Marisa started violently pushing her snake inside Alice's vagina over and over while tickling and groping Alice's breasts.
"If there is one thing I know is that these cute breasts of yours are bigger than mine!", Marisa said while laughing and watching Alice's face in despair moaning and chuckling.
Suddenly however, a sticky while liquid started being shot from Marisa's snake. Just moments after, all of Alice's face, vagina, breasts, hair and tummy were covered in that white stuff.
"W-What is this?", Alice said since her gag had been removed due to the pressure of being cummed on. Marisa then replied with a soft voice just before giving her a kiss made out of pure love: "They are the spirits of the computers".
The end.

>>1
I'll humour you for a moment.

There is nothing there to disprove. Its such a nothing, that even trying to *describe* mentishit's "work" can seem cruel.

This is what his code looks like, and what all his spam boils down to if you actually read it:

// Mentishit's Complete Artificial Electronic Sentience implementation

i = 1; // The concept enters the sensorium
print i; // The AI communicates its thoughts

// Refer to page xxx in the how2perl manual for how to make forward declarations.

That's literally all there is. He is someone who never attained any fluency in any programming language, who upon declaring a 2D array claims to have discovered electronic sentience. Even pointing this out here seems rude and unseemly, like making fun of a mentally disabled kid. Except in this case the "kid" is well past middle age and never got the help he needed.

Trying to describe mentishit is like trying to describe a fart in the wind. There is nothing really there, but its still vaguely unpleasant and obnoxious. Mentishit is part of the rich diverse tapestry of shitposters here on /prog/ that are guaranteed to derail threads and shit up the board on cue.

>>7
that is the code, but he has also written a few books and a noted AI expert has found his theories entertaining:
http://www.sl4.org/archive/0205/3836.html

so I'm not convinced.

>>8
From your link, where everything comes crashing down:

I didn't look at the code (yet), I just read the docs on the site.

Only a fucking idiot would be "not convinced".

>>9
Are you implying that the founder of OpenCog is "a fucking idiot"? Because in his view:

although Arthur sometimes *presents* his ideas in a somewhat

kooky way (by the standards of the mainstream scientific community, and even
by the standards of this list), the ideas themselves are significantly
better than most of what passes for cognitive science and AI. There is some
deep thinking there

>>10
I'm saying >>8 is a fucking idiot. I'm saying the founder of OpenCog didn't really dig into Mentishit's shit, and just glancing at the writings of any AI kook can always look somewhat real.

>>10
Also, the NN and statistical based AI push that everybody's in is shit, so anybody looking at a conceptual basis to AI will always look smarter than the "state of the art" garbage.

>>12
But NN and statistics work. OpenCog is based of them too.

>>12

Also, the NN and statistical based AI push that everybody's in is shit, so anybody looking at a conceptual basis to AI will always look smarter than the "state of the art" garbage.

I completely agree. There is an actual interesting cognitive model I read about years ago. It was created by Douglas Hofstadter and Melanie Mitchell and described in the book "Fluid Concepts and Creative Analogies". Its the "Copycat" cognitive model:

https://en.wikipedia.org/wiki/Copycat_%28software%29
http://cogsci.indiana.edu/book.html
https://people.cs.kuleuven.be/~joaquin.vanschoren/zsp/fluidconcepts/

Its behaviour is very interesting to watch. It looks very natural as it works its way through a seemingly simple analogy problem.

There was a java implementation of this model, but it was taken down. It is mirrored here:

https://archive.org/details/JavaCopycat

Then there is this guy in the same research group:

http://www.foundalis.com/res/diss_research.html

He got so frightened by his own puzzle solving program that he took it down, and now cries himself to sleep each night:

http://www.foundalis.com/soc/why_no_more_Bongard.html

So make of that what you will.

The model itself doesn't require knowledge of academic jargon to be understood or implemented in a computer program, and its an interesting excercise trying to apply it to other domains.

>>14

Hofstadter

Yet another kooky lisper. AI/lisp people are all crazy.

>>14

He got so frightened by his own puzzle solving program that he took it down, and now cries himself to sleep each night:

Is that the guy who freaked out because he thought the military might find some use for it and cried about dead Muslims? His shit wasn't that great.

>>15
Fuck you buddy. Reading Gödel, Esher, Bach was a defining moment of my teen years.

>>13
They work for marketing analysis and such. They don't work for actually creating AI.

>>17
what about https://www.youtube.com/watch?v=qv6UVOQ0F44 ???

they give similar results to that guy using emulator to predict what input produces the best next state. I'm sure you can apply that to real world problems too.

>>18
It works, but IMO its not as interesting/satisfying as trying to find a cognitive model. Its elaborate curve fitting, like training a neural network or other substrate.

Cognitive models like Copycat on the other hand I find more interesting to watch and think about, because when it runs it actually looks interesting, like it models an actual thought process. Maybe its just a subjective thing, but when I see the concepts in the concept network light up and drive the direction of investigation in the workspace, I admire the effort to try to make a model/statement about what thought even is. More machines/models of this nature wait to be explored.

https://people.cs.kuleuven.be/~joaquin.vanschoren/zsp/fluidconcepts/figure6.png

>>1

Can you gays disprove his theories formally

A theory that doesn't work doesn't merit a formal disproval. "Blah blah blah, prove me wrong" works only on imageboard autists who literally can't help themselves not to rise to the challenge.

>>20
A lot of established AI researches have or had theories that don't work. For example, Marvin Minsky published Society of Mind, which failed as a viable theory. At the same time, Minsky himself debunked earlier theories.

>>21
You are too coherent to be mentishit, so you must be a troll.

>>21
Debunking earlier theories can barely be considered useful work.

Even if Mentishit's theories happen to have some ``truth'' in them, he hasn't delivered jackshit. All talk and no action, just like Cudder-kike.

>>24
Right, without looking at Mentishit's shit, anybody can go on and on about the notion of finding patterns in perceptions, organizing stimulus into groups, reinforcing commonalities, tying feedback to new patterns, ideas about how mental abstraction might work, etc etc etc, all as just random armchair conversation about thought. None of it implies that any work is being done, that any discoveries are being made, or that the person has any clue about the current state of the art. It's just casual musings.

>>14

foundalis

Andrea Rossi of ML.

>>18
There's also vizdoom, which I think will not be as easy to MC minmax with sheer brute force like mario.

>>25
The thing is, even academia seem to obsess over theory of general AI, rather than putting it into practice. But at least those ideas tend to be more concrete and coherent than vacuous ramblings of online wackos.

http://arxiv.org/abs/1406.2661
http://arxiv.org/abs/1604.00289

>>26
academicians are being paid for papers, not for something solving the problems.

>>27
That's just the thing. They are paid to create problems, not solve them. Finding solutions will reduce their funding.

>>25
Calling what mentishit does "casual musings" is extremely generous. Over time "casual musings" nearly always dissipate. Or, if sustained, disciplined effort is applied they may eventually crystallize into something coherent. Neither of these apply to mentishit. There is no substance in mentishit's garbage at all. Even calling it the output of a spambot is pushing it. He has been pushing his "hello world"-tier "program" for decades now. He is deranged. He is deranged even by my standards, and I know I'm a lunatic.

http://www.nothingisreal.com/mentifex_faq.html

>>28
You're wrong. Academics are paid to solve problems as there is never a shortage of problems to solve. The number of solutions to solve a given problem are sporadic in comparison.

>>30
When was the last time academics solved any real problem?

>>31
academia solves the problem of government kickbacks.

>>31
Define "real problem". I'd define it as a problem that an academic is tasked into solving.

real problems are only solved with True Enterprise Level Solutions, The Cloud (TM), anything prefixed with "smart-" and three-letter acronyms.

>>33
A problem that affects a large percentage of programmers. Your definition is wrong.

What does ANDRU mean by the way? I get that it's what this guy's AI is called, but does it stand for anything in particular?

>>36
ANally Devestated RUssian

>>36
ANDRU - the friendly android AI.

>>36
That was the name in the original COBOL.

>>39
The guy's been working on this project since the COBOL days? And what did it originally stand for back then?