1143Views18Replies

Author Options:

HAL 9000 Open Source Community Project - Announcement and Invitation Answered

Goal:  Build a HAL 9000 Series Computer
Specification:  "2001: A Space Odyssey"
Website:  www.hal9000project.org
Instructable coming soon

I watching "2001: A Space Odyssey" the other night and I decided it would be cool to build one.  I've never managed a project this big, and don't plan to start doing it now, so hopefully people will step up and assume various roles and responsibilities.  A project like this needs a program manager, deputy program manager, project managers, systems engineers, software engineers, programmers, graphic designers, web developers,etc.  I'll assume the role of chief engineer and chief software architect.  Also needed are an admin for the sourceforge project, and a webmaster for hal9000project.org.  I don't plan on having any special rules or regulations, I'm just throwing this out there to see what comes of it.  It could end up in a hairball of chaos, or it could evolve into a smooth running machine, although I suspect nobody will join in and it'll be just me with nobody watching.

This is not a far fetched pipe dream.  The IBM Watson has already beaten the reigning champion on Jeopardy.

I'll be defining all the individual subprojects such as the speech recognition function, lip reading function, speech synthesizer (it will have the exact voice used in the movie,) facial recognition function, voice recognition function, object recognition, various other artificial intelligence functions, neural network, physical mockup, physical prototype, software simulation, etc.

There is a distinct possibility of getting government funding and along with commercial grants, therefore, some people could theoretically carve a full time job out of this.  That would work for me.

18 Replies

user
BOT4 (author)2011-11-17

Well this is great and all, but you do realize i have already done this. now my Hal 9000 may not be all polished up with metal and all the original stuff, but it is fully functional and you can have a full conversation with it.

Select as Best AnswerUndo Best Answer

user
Goodhart (author)2011-05-26

There is a tremendous difference between answering questions "that need to be formed just so" and cognizant determination of what the the other "entity" being spoken to's intention, meaning, and such, and then giving an intelligent response.

True enough, we have come a LONG way since ELIZA (web based version of the old PC program)   but we are still miles off from truly intuitive conversation.

Select as Best AnswerUndo Best Answer

user
el_roboto_loco (author)Goodhart2011-05-26

Actually, there's something called the Turing test, in which a computer has to interact with a human such that the human doesn't realize it's talking to a computer, (or having an online chat with a computer.) The IBM Watson is the Jeopardy all time champion, and I'm pretty sure it won the first and only Turing prize, but there's also another guy, from Harvard, who won the Turing prize, which might be the same award with IBM.
http://goo.gl/K2rP0

And check out my latest indestructable.
http://goo.gl/fkiLc

Select as Best AnswerUndo Best Answer

user
lemonie (author)el_roboto_loco2011-05-26


Be careful not to confuse Turing tests with awards named after Turing (awards given annually since 1966)

As far as I'm aware the only prize Watson has won was the Jeopardy $1,000,000.

L

Select as Best AnswerUndo Best Answer

user
lemonie (author)el_roboto_loco2011-06-01


It'd be that small word "like" in "like IBM’s ‘Watson’ computing system" I guess.
Why did you use that short URL?

L

Select as Best AnswerUndo Best Answer

user
Goodhart (author)el_roboto_loco2011-05-26

Indeed, that "test" has been around for a long time (updated obviously since it's inception in 1950). :-)

Select as Best AnswerUndo Best Answer

user
lemonie (author)2011-05-25

Have you an existing design in mind as a starting point, or is the project mostly in your mind at the moment?
Also, why would (potential funding) be interested in this over any other powerful computers - what are the key advantages that would make this worthwhile?

L

Select as Best AnswerUndo Best Answer

user
el_roboto_loco (author)lemonie2011-05-26

Fair question. Why fund something like this? Well, what's it worth to hold the patent on the most powerful computer in the world? But, the real and tangible thing is the business you would get from the Department of Defense. Billions and billions. The way it works, for example, I'm submitting a proposal for an Air Force solicitation called "Countering Future Cyber Threats to Air Force Weapon System." Now, making the assumption that a HAL 9000 represents the software universe (it can do anything,) then if a computer system, which is a subset of the HAL 9000, meets that Air Force requirement, then you get a phase I study contract, followed by a Phase II contract to build a prototype, then a Phase III to go into production. Then, since you have a patent on the most attack proof system in the world, the Air Force is only going to buy your computers, and in fact, your system will become the new military specification for military cyber systems. Also, the rest of the Department of Defense needs your computer systems now, because they are the only ones which meets their security specifications. Then the rest of the government only wants your computer systems. Then the commercial world only wants your computer systems. And so on.

Is that clear enough? Do ya see buck being made somewhere along the way? BTW, I fully expect to win that Air Force contract. I've got a 100% lock on it because I did a proposal on the same thing 3 years ago, and I would have won then, but something came up and I missed the cutoff date. Also, I did a vulnerability study a couple years earlier, on a large Air Force system at the cape, and I found some big problems that affect all their systems. Unfortunately for them, they never saw my technical report. So they will be reading about all of their security vulnerabilities for the first time when they read my proposal, which will have a case study based on my vulnerability technical report, which I found the other day on an old flash drive. Perfectly good question though, because it gave me a chance to explain why the project is worth something. And, I just submitted a provisional patent application for "Threat Detection and Response to Insider Attack against a large-scale Military Cyber Weapons System."

Oh, and the answer to the first question is yes!

Select as Best AnswerUndo Best Answer

user
lemonie (author)el_roboto_loco2011-05-26


Thanks for the reply, but please respond to what I say rather than giving me what may be mistaken for a delusional-rant.

Oh, and the answer to the first question is yes!
The question requires a one or the other answer, not "yes".

To repeat the second a bit differently: what key differences will there be between this and any other existing computer, that are actually proven and realisable? Why are you the only person who can do this, and why are you telling us about it?

L


Select as Best AnswerUndo Best Answer

user
iceng (author)2011-05-25

Iv never even done a chess game :þ
Saw the movie twice.

A

Select as Best AnswerUndo Best Answer

user
jeff-o (author)2011-05-24

Wow. Well, good luck with your project. It's a few miles above my head...

Select as Best AnswerUndo Best Answer

user
FoolishSage (author)2011-05-24

It is quite ambitious, I'll give you that. Great things have been accomplished by dedicated ad-hoc groups so I wish you the best of luck.
I have some project management experience but not with software development so I doubt I could be of much help. Should you manage to pull it off put me on the list of buyers ;)

Select as Best AnswerUndo Best Answer

user
FoolishSage (author)FoolishSage2011-05-24

One thing I forgot to mention:
If you get a proper development plan set up you could try crowd sourcing the project (like with kickstarter or something)

Select as Best AnswerUndo Best Answer

user

Cool. I hadn't thought about that. Is that free?

And I already have a complete development plan. It's all in my head. I just have to type it in. And it's going into an instructable, and it will have all of the functions, 1 per step, and the requirements and algorithms. I've been doing this for 30 years so I don't even have to think about it.

BTW, about the program manager, the biggest part of the job is simply writing letters to people, like starting with President Obama, and right on down the line to congressmen and senators. Also, I need somebody to help me get a proposal written to Homeland Security. In fact, I forgot to mention that I am going after an Air Force contract, and I need help with it. Stuff like proofreading it, and making sure I don't forget to submit some stupid one page thing that's required. And this Air Force project is a 100% exact version of the HAL 9000. Furthermore, I'm 99.9% certain to win that contract because I have inside knowledge that the Air Force doesn't even know. That stems from a vulnerability study I did years ago and found major problems that were systemic, which means all the computers in the Department of Defense probably have the same vulnerability. The title of that solicitation is "Countering Future Cyber Threats to Air Force Weapon Systems"
http://alanmollick.com/airforce

Anybody can be program manager, they never do any of the actual hands-on work. What kind of work do you do? Do you have a website? Do you want to work with me on getting that Air Force contract? It's not a lot of work, I'm just doing 50 things in parallel. And I swear, I've got a lock on that contract. I've got to get it in probably by the end of next month. That contract would lead to a huge contract at Phase III. God, if I can get one person to help me in any way, big things will happen quickly. I'm just one guy, but with one person, things would start happening immediately. So, that's 2 contracts I'm going after, and either one means this HAL 9000 is fully funded, and I'd be hiring 1 or 2 people immediately. And, I'll probably be doing one or more proposals for DARPA. And there's a bunch of government agencies that also give out funding and grants. One place I applied to for a job is Idaho National Laboratory. Anyway, I need somebody to do something, anything, and big things will happen. Unless I have a heart attack or a stroke, as hard as I'm working right now...because I'm unemployed!!!

Select as Best AnswerUndo Best Answer

user

I'm not exactly sure how kickstarter (http://www.kickstarter.com) works on the end of those proposing a project but it boils down to asking people to help fund your project in exchange for something once you get the ball rolling. (like a "thank you" mention on the website for a small donation or one of your products for a larger donation)

As for my own contribution, I had not realised that you intend to do this with/for the US government. That being the case I no longer have interest in participating.

Select as Best AnswerUndo Best Answer

user

I'm not doing it for the government. Nobody is paying me a penny to do anything. However, taking into consideration that a HAL 9000 does everything imaginable, obviously any other system in the world has to be a subset of what a HAL 9000 can do. I'm not pitching anything to the government about a "HAL 9000."

Note that, because I have a very small mind, the only way I can fit a lot of complicated information in my head, is to simplify everything down to the lowest common denominator. When the denominator reaches 1, as in this case, I know I'm done. So, with three projects in mind, I combine them into 1 project which covers all 3. What the govern might pay me for, is just a matter of book keeping to me. And they would only pay me for either the border system, or the Air Force system. And, in my mind, a computer is just a computer, and software is just software.

Select as Best AnswerUndo Best Answer

user

Well, I don't think a program manager necessarily needs software experience. I was thinking mainly keeping track of the status, and mainly be the POC. And, I know I can pull it off. If you look at everything that's been done before. Plus a few things I've done that nobody knows about, like voice cloning. And this project is a follow-on to the border security article I wrote, and I would use the HAL 9000 to operate the border system, or the software anyway (which is exactly what HAL 9000 was anyway, the red lens didn't even modulate when he talked) as I am currently lobbying Homeland Security to let me take over that border project. Furthermore, most of the software I'm proposing in my border system proposal (that I'm working on) would cover 90% of what would go into the HAL 9000, which means if I can get a contract from Homeland Security, then the HAL 9000 project is also funded. I'm drafting a letter right now to Regina Dugan, director of DARPA, letting her know about the HAL 9000, and other things I'm working on. I'm hoping she might nudge Janet Napolitano towards giving me a contract. I sent a letter to Janet Napolitano only just recently, so I don't even know what her response is yet. In a perfect world I would get a contract because I can salvage the whole system with no additional hardware, and give the government back something that exceeds all expectations by a mile. I'd have to hire a couple people of course, I don't plan on doing it all by myself, in fact, I'd rather not even write any software, just write the requirements and algorithms. Border Security instructable: http://goo.gl/mFgox DARPA (Defense Advanced Projects Agency) http://www.darpa.mil/

Select as Best AnswerUndo Best Answer