George wrote:

> I was disappointed at Stan Franklin's book when I first read it,

> why? Because it was not to the point about how to build AI in

> terms of software implementation.

> 

> I don't think any software engineer who turns to the question

> of what AI is about is curious about the nature of feelings or

> religions or the meaning of psychosis and schizophrenia.



Indeed. I'll tell you about software engineers. Their good point is that

they don't simply hook up on any philosophical issue, but will hardcode

if the situation requires them to and their bad point is that they don't

mind if they've just created a new nuclear bomb which will take away

half the planet, the important thing is that they did what they were

told to do.



I am a coder and I have been (and still am) working on some very

practical issues of AI, atho I have started up in this very newsgroup,

being a philosopher naturaly. When I came in contact with a pro AI

software engineer, he already had a fixed idea that good software (that

would solve the current problems) was impossible to make... Well, closed

minds stay closed!



I have checked the issues and have cooperated with a friend (also

originaly from this NG) to create a piece of software that can solve the

integration problem that was on the table at the time... YOUR TURN!

Where is everybody? With your genious little programs, capable of doing

so many interesting things in their area? We need those now, to make

them work, ok? We need to integrate the programs with our new

technology!



> Not philosophy about Chinese rooms

> and consciousness and Searle, etc.,  but software that performs

> what they expect to see. Programmers want to see an object called

> CGeneralIntelligence with programming examples and relies

> simply on software/hardware configuration tools to make it work.

> Would it use neural networks or other implementation means in the

> background, not as important (I prefer distributed objects and

> massive parallel hardware support over ANNs.)



Oh, sorry, but you're still a bit "up in the clouds", you might not

realize it. You see, CGeneralIntelligence does not exist anywhere; not

even in the human brain.



> Such a general intelligence system needs to be able to learn (to

> collect knowledge) about it's environment and to pursue it's goals

> within that environmental familiarity, for example to be able to drive

> a car from one city to another on it's own. Any learning system needs

> to make sure that everything that was learned is configured toward

> safety, and the formed expert system [artificial character] does not

> represent any harm to it's environment -- unless the technology is

> used for the military of course.



I'm sure you know you've just described a learning program (better known

as a databank creation and backtracking utillity), not a

CGeneralIntelligence. But that's ok; if you can make one (it just takes

a little skill and work, basicaly), we can work on integrating it into

our system and your AI will be ready to rock.



We are also working on a NLU feature right now, so if anybody has any

code up to those Alice-like chatterboxes, we are ready to give that

NL-warping code a link to a brain! Any programming language that can I/O

to ASCII text files will be ok, atho the practical interface modules are

all written in VB.



C'ya!



--

Cellphone: 0038640809676 (SMS enabled)



Don't feel bad about asking/telling me anything, I will always gladly

reply.

[AC/HFA(AS) -- no suprize]



Digging for info? Try AI Meta Search:

Http://WWW.AIMetaSearch.Com



MesonAI -- If nobody else wants to do it, why shouldn't we?(TM)

Http://WWW.MesonAI.Com

