On the way to Human-Level-AI

[menti...@gmail.com]

"Trollheaven Magazin Januar 2016"
on page 26 has a section in German
"1.9.4 Auf dem Weg zu Human-Level-AI"
or, "On the Way to Human-Level-AI".

The author Manuel Rodriguez says in German that
"AGI oder auch Human-Level-AI kann auf zwei Wegen
realisiert werden. Einmal klassisch durch einen
selbstlernenden Algorithmus. Bestes Beispiel hierfür
ist sicherlich der Ansatz von DeepMind (General Game Playing),
oder AIXI (Kolmogorov Induction). Beide Ansätze basieren darauf,
dass man sich vom Servereine Datei herunterlädt die rund
100 kb groß ist, diese auf seinem Rechner startet
und aus diesem Programm dann eine Human-Level-AI heranwächst.
Auch Mindforth gehört in diese Kategorie und zwar deshalb
weil der Source-Code vom Umfang her sehr klein ist."

My rough translation of the above paragraph into English:
"One can realize Artificial General Intelligence (AGI)
by two routes. For one, classically through a self-learning
algortihm. The best example for this is surely the attempt
of DeepMind (General Game Playing) or AIXI (Kolmogorov Induction).
Both attempts are based on the idea that from the server
one downloads a database that is roughly 100 KB in size,
starts it on a computer and then a Human-Level-AI grows up
from this program. Even Mindforth belongs in this category
since the source code is very small in size."

As the author of the above-mentioned Mindforth, I hasten
to point out that the Mindforth AI is available from

http://www.nlg-wiki.org/systems/Mind.Forth

by ensuing links, and a German version from

http://www.nlg-wiki.org/systems/Wotan also by reference.

Apparently the author of the "Trollheaven Magazin Januar 2016"
first became aware of the Mindforth and JavaScript AI Minds
around September of 2011 as shown in a Usenet exchange at

https://groups.google.com/d/msg/de.comp.lang.javascript/JgTCdm3bu5c/KrTH9EUxeeEJ

and the same author four years later in 2015 made these blog-posts:

http://aa5new.wordpress.com/2015/02/22/mindforth-oder-warum-forth-heute-ausgedient-hat

http://aa5new.wordpress.com/2015/02/24/mindforth-genial-oder-absurd

http://aa5new.wordpress.com/2015/02/24/bug-mindforth-ist-schlecht-dokumentiert

http://aa5new.wordpress.com/2015/02/24/mindforth-dbpedia

https://aa5new.wordpress.com/2015/05/13/was-ist-true-ai

which collectively show a transition from an initial reluctance
to ascribe any merit to the AI Mind programs and a gradual
acceptance of the possibility that Mindforth may have merit.

I wish to thank the author of the German blog material
for his voluminous AI writings that show some long overdue
respect for Mentifex open-source independent-scholar AI
and I wish to greet AI enthusiasts in this New Year 2016.

Arthur T. Murray/Mentifex
--
http://ai.neocities.org/AiSteps.html
http://www.google.com/patents/US8620890
http://dl.acm.org/citation.cfm?doid=307824.307853
http://ars.ict.tuwien.ac.at/new/wp-content/uploads/2012/06/Diss-Deutsch.pdf