list of ai underground projects

[raffethefirst]

Hi
I am making in my wikipedia user page a list of ai underground projects - and peoples involved.

Those are the projects that are not allowed in main wikipedia ai projects page.
Please help me improve this list with people and projects you know.

the page is en.wikipedia.org/wiki/User:Raffethefirst

I spot here another candidate: zouze. Can you tell us a few words about you and your project please?

[tkorrovi]

Well the ADS-AC project is not exactly underground, it is listed in artificial consciousness article, well, also in FET AC proposal, but yes it is not listed in wikipedia ai projects page. I think that all open source projects must be referred to in wikipedia, there are not so many of these, and this is important for software development, as every available code can be useful for programming.

Concerning the open source projects, there is of course problem with mentifex, as his project is indeed open source, written in javascript and forth. He once wanted to add a link to his project to the artificial consciousness article, but i deleted it, because his system has nothing to do with artificial consciousness, neither has it anything to do with true AI, in spite that Arthur frequently claims that. That this project has nothing to do with true AI, is all that is important for me, and i therefore am not more interested in that project. Concerning the criticism against Arthur, i would say that it is right what concerns his claims, like "the AI is solved", and recently, when i remember correctly, he claimed something like his system "achieved true AI". These are indeed populist claims, not substantiated anyhow, and it was right to criticize Arthur for that. Otherwise, there is also a lot in criticism against him, what more resembles a personal attack, than criticism by a scientist, these who wrote such criticism should understand that they seriously decrease their credibility by writing such things. Who thinks that their criticism against Arthur was efficient, are mostly wrong, i know people who appreciate Arthur's work, even in spite that i tried to explain them, that it is not what they think it is, and this criticism seems to them so primitive, that it doesn't influence them anyhow. What concerns Arthur's program, then all i know is that it is some natural language processing system, i see there nothing substantially important, like a new model of computation, self-developing system or system which implements a type 0 grammar. Arthur's system can include some unique mechanism, but it has no such importance. Every chatbot can contain some unique mechanism for minor things, like how to answer some particular question, this might be unique but has almost no importance for research. But unfortunately the criticism of Arthur's work is not written competently, it looks like it was written by a person who doesn't know much about programming himself, and this unfortunately makes it not very credible. I would say that Arthur's system has some value as an open source program, some of his code can be useful for people, who for example want to know how to write a program in forth, so it has to be recognized as an open source code, and i see no reason why it cannot be listed as such in wikipedia. But what should not be accepted are Arthur's bragging, populistic and not anyhow substantiated claims about his system, and his description of his system, where he represents his system to be something else, than it really is, and which also doesn't anyhow help to understand what his program does. This is all i have to say about the Arthur's system, but again, i'm not at all interested in that system, because it has nothing to do with true AI, and i'm therefore also not going to study his system more precisely. When anyone should do that, then these should be the people who research the conventional AI, especially these who research the natural language processing. And when they cannot cope with Arthur, then it is their own fault, and they cannot blame me for that.

[raffethefirst]

I never had the patience to read all Arthur`s theory.
I am trying to have a talk with him here: www.ai-forum.org/topic.asp?forum_id=1&topic_id=8385&page=3#end
but he dont reply... maybe he is in vacation .

I want to see first if we can have a normal and constructive talk and if so I will help him with all I can.


You said there are some other open projects. Can you tell me some of them please?

[tkorrovi]

I have talked with Arthur, and he can sometimes be kind of understanding, but then he goes again to talk his "AI has been solved" stuff. He even was not against an unrestricted system, and he even somewhat appreciates my efforts, but said that he had the aim to create an unrestricted system at first, but then abandoned it because he thought that developing it would take a lifetime. Then he started to create a restricted system, but could not have really any success with that.

I said there is modular AI project modularai.messageforums.net I saw how they started that project. The idea was to gather people to make some AI sysem together, so that together they did have more power, and could really create something great. They invited me too, but i refused then, saying that i don't think that a joint project of creating a conventional AI would ever succeed, and there were also other such efforts before, which all failed. But they still started their project. And now it is stalled, exactly because they wanted to develop conventional AI, which is restricted, and it must be decided for what task it should be made, but they could never agree in this, and likely some also were not satisfied with a system for some restricted task. Some people from that project are now in this forum, at least they recently registrated here, and did agree that the only conceivable aim for a joint effort, would be unrestricted system.

There was one earlier similar effort, which also failed because of the similar reasons, when i remember correctly, it was called AI Consortium, but i cannot find their site any more.

Abram Orion Demski also created his Dragonlogic AI project s13.invisionfree.com/dragonlogic_AI which is also dead now. Orion was once a strong supporter of my work, he also wrote in this forum, but now he thinks about some mathematical logic etc, things very far from unrestricted system. He also seemed to prefer the Jeff Hawkins system to mine, though admitted that the Hawkins system is also restricted, also i and Paros here reached the same conclusion in some other thread here, i think that you find it.

[martin]

I was in both the modular AI and the AI consortium, both are dead

The AI consortium fell apart due to a lack of direction, it was simply meant to be a place for people to chat about AI and possibly work on a project but there was never enough interest.
The modular AI fell apart due to work requirements of some of the people, a lack of people and differences in opinion about where we should go with the project.

I'm still around though, and doing a comp-sci degree at the moment, I'll be taking AI modules next year so expect to see some more crazy AI plans from me

[tkorrovi]

Well, nice to know that at least someone is around

What else can i say, well, there are these who made some kind of system, may be natural language processing, artificial vision, whatever, but still like some existing type of system, with some improvements. May be well done, with a lot of good work lasting many years, i know how difficult it is and can appreciate such efforts. And the author sure wants to develop it further, some of them promote their work or even force it to others, trying to find supporters or these who would develop it further. And it may seem that this is exactly the same what i also do, just promote a system which i happened to make. But there is some essential difference which many fail to see. This is not yet another system, this is something much more general. It is about creating and researching unrestricted systems (systems aimed to be unrestricted in every way), a whole field of research which as much as i know is not done anywhere else, something substantially and qualitatively different than the other areas of AI research. A research of something the most general, not just making yet another system. The other efforts may fail, this effort cannot fail, there shouldn't be no doubt that researching something as general and fundamental as unrestricted systems is necessary, and every research produces results, this research already has produced results. Thus dealing with it is dealing with something general, fundamental and important, where one sure learns a lot of new, and creates new tools to research these systems, the systems which likely in a way or another are a part of nature. Thus one can be sure that one does something useful, beyond doubt useful for understanding the most basic things. So, for one who wants to do something certainly general and fundamental, would be a reasonable decision to join the effort. And about crazy plans, well, there possibly cannot be more crazy plan than research of some the most fundamental things