Quote: Original post by saintdark
i just have this outer plan 'woods'. I need more volunteers to take decisions and sm brainstorming sessions...
R u interested in the project..
thought abt lisp.. but dunno much abt its stability with other low-level langs...
regards
saint dark

I didn't mean to say "use lisp" as an answer, but rather a means to an end. With existing technology, Lisp-based OS's, and the relative ease of adding at least a portion of your proposed solution, the 'intelligence', be it with GP or otherwise, should be easily prototyped in Lisp for various reasons. You could have a 'proof-of-concept' in an acceptable time period.

I'm just throwing out raw ideas here. Perhaps there are other languages/tools that would enable you to prototype something like this effectively. I think that's what is important though. A prototype. Something that works.

Quote: Original post by Sagar_Indurkhya
I mean, we can't even build a bot that can past the turing test,

Please join the 21st century; the Turing test was passed / proved a waste of time LAST century. A mildly entertaining way to pass the time, yes, but not of any scientific value (beyond PR stunts).

Quote: Original post by saintdark
--As for code efficiency, the OS will have to learn to make its code more efficient on its own..
The Operating system will decide over time wht to run and wht not to run.

I'm sorry, but in the interests of sparing some of the precious few hours of your life, and speaking from a position of knowledge and experience (someone who's been working in EP for years, and has good friends and colleagues who are still active in the research side of EP):

Get a life, do something meaningful. Don't waste it on pointless silliness like this

You will not achieve this in A MILLION YEARS of processor time. If you used every single internet-connected machine on the planet then - maybe - you'd get something vaguely interesting after 200 or 300 years. Probably not. You certainly wouldn't have a working OS, that's for sure!

Quote:
YOU CAN REALLY MAKE A CHANGE.

No. Not by doing something like this you can't. Might as well start a project "The Human Bird: we need volunteers who we'll bombard with radiation then see if any of them produce children who can fly. It may take a lot of people, but we'll get there eventually!"

"The human bird" idea is really kewl, why dont you start the project anonymous user("why hiding").
Every technology during its research time always seemed partially impossible, if evry ppl who were on research gave up like this, we wuldnt have been here, or wuldnt been enjoying these technology..
AI is the future, We will have to slog and spend our lives on this for the future generation to enjoy the fruits of knowledge

I attempted some ridiculously ambitious things in my early 20's... this project makes my own past foibles look quite reasonable. Is there any way to let someone down gently? Probably not, but I'll at least try logic.

Point one: I have no idea, looking at the website, what you actually expect this OS to do. In particular, how much you expect the OS to bootstrap through 'intelligence' and 'learning', and how much you intend to code by hand.

Point two: have you considered the number of times you'd have to reboot to query a random piece of hardware, to reverse engineer its interface? Have you added up the seconds for all those reboots?

Point three: have you considered the budget you're going to need for destroyed hardware, querying it in such a random fashion?

Point four: have you considered the rate at which the HW industry changes its interfaces, vs. the rate at which you can produce results for your OS project?

I think you'd be a lot better off trying to write querying intelligences for interfaces that are actually stable. Like, for instance, the C++ language standard. Random probes might actually get something done in computationally tractable time. Automagically bridging between programming languages is a very hard problem, but it would be terribly useful if it was solved.

Trying to query and reverse engineer the disposable PC industry? Forget it. You're strategically DOA.

Quote: Nex will be a completely independent intelligent entity with the ability to self evolve its code, except its primary objectives.

It won't run on longhorn then ;)

Quote:
The Operating system will decide over time wht to run and wht not to run.
THE PLAN IS NOT TO MAKE AN OS THAT IS INTELLIGENT IN CERTAIN AREAS.. IT HAS TO LEARN EVERYTHING OWN ITS OWN.. WE WILL JUST PROVIDE THE BASIC NECCESSARY DETAILS.. AND THE FIRST VERSION OF ITS CODE ;)

This sounds a lot like trial and error to me. Why trial and error when you yourself (should) know what the OS does and doesn't run.

I don't think code evolution is feasible, at first I thought you meant an OS which recognizes patterns in what the user is doing, etc, and then maybe creates a macro for it, and maybe it will fine tune itself depending on what programs are being run (eg, shut off desktop rendering (or take it down in priority) if a game is being played, etc). That sounds like a fine plan, and hard enough (and still not invented). Tell me something, do you know all about how your standard OS works, or indeed how you would code this OS to evolve itself.

Are you planning on using genetic algorithms?

Now think about this:

Given enough monkeys a typewriter, the theory goes, and they will eventually produce the works of Shakespeare.

yes.. give the monkeys the typerwritier and they will be shakesphere "well he is not GOD, is he"..
Stupidical referances is not neccessary here, guys i needed help from you.. and thx for your suggestions and criticisms.

--The OS work will not be done by the AI development team..
The OS will be developed and will provide the framework for Nex to run..

--Your suggestions are warmly welcome.. I Need more ideas

if the os totally independent from the ai, then you just should conecentrate at writing the ai and use an existing os

Yeh anonymous i even thought abt it, but much of the ppl are interested in scratch, so that this will be a great learning experiances.
Then the problem od modifying an existing OS which ofcourse has to be Opensource, at the moment i really dont support the idea of Opensource.
I really want(if the project finishes) the volunteers and me to benefit from it some way or the other, so i am not thinking of modifying an existing OS.