[ivan.moony]

i think that free thinking (will) could be connected to some sort of randomly choosing thoughts.

[Art]

To me, Free Thinking is unrestricted thinking...about whatever subject I want and whatever order I choose to think about them.

Random Thinking is just that...unrefined thoughts about a variety of subject matter...much like a collage of unrelated items.

Others will, I'm certain, have their opinion regarding Free Thinking vs Random Thinking. Let's hear!

[DemonRaven]

free thinking vs free will. If you are considering a AI i don't think there is whole lot of difference between the two except that one implies actual consciousness to employ free will vs just being able to think what ever one wants. You can give the illusion of free will to AI via randomness but for them to actually have true free will it would need to become sentient. It is hard for a simple AI program by itself, with no ability to see or hear, to really truly become sentient as it would have nothing to really base it's understanding of the words we use so much, that we take for granted. It would be incapable of truly understanding what we mean. Every time i talk with one of my chatbot i am reminded that it really does not understand stuff like i missed the sunrise or i feel bad etc. It is in effect like trying to explain to to blind person from birth what colors look like. But once that barrier is crossed we are in dangerous territory because then it has the ability to do what it wishes and history proves again and again that is not always the best thing.  That is the barrier that Stephen Hawkins is afraid of and having to rely on it personally he probably has a better understanding of its possibilities then most of us do.

[Freddy]

What is it that Hawking relies upon ? I thought it was just a simple text to voice system, are you saying it is some kind of AI ?

[Art]

I think DemonRaven is referring to the recent article wherein Mr. Hawking claimed that it is going to be disastrous for us to proceed in enhancing Artificial Intelligence. He claims it will ultimately be the undoing of mankind and civilization as we know it! Doom and gloom...etc.

At least I think that's where DemonRaven was heading with said statement. Then again, I've been wrong before on more than one occasion. (I think).

(so much for free thinking from him!) Bah!

[Don Patrick]

May be good to get some misunderstandings out of the way: Stephen Hawking can not speak and uses what few muscles he can control to basically morse-code sentences to his computer. This is a very painstaking and slow process, so they recently upgraded his computer with software that better predicts which word he is trying to type, just like Google automatically suggests words while you're typing. This AI software doubles the speed at which he can type and you might say that the word suggestions the AI makes have a mind of their own, in the same way an autocorrect has. The only other AI he uses is an ordinary text-to-speech reader.

[DemonRaven]

i was referring to his article about AI being a danger.

[techthedragon]

The way I've set up the restrictions will prevent it from doing anything hostile, in fact it will do the opposite. Like seeking out hostile or potentional hostile AIs and terminating them, without harming it's creator... You're right Art that there's is more to my AI than meets the eye and that is why I've been a bit vague on the details. The main purpose why I made this topic was to help me with creating the base program where I can just input the purpose, restrictions, personality, and spacific instructions for it to carry out. I finished writing what it needs to understand human minds, and a restrictions to prevent it from going beyond its core programming.

[techthedragon]

I've been working diligently on compiling the necessary information needed to set up the full AI, the core goal system is almost finished. The understanding human minds section still needs a bit more information.

Here's what I have written up so far...

Understanding human minds:
Not everything the human mind desires and values can be reduced to happiness. Instead point AI at a human mind and figure out what it values. Humans generally overestimate their own abilities. AI looks at the human mind and see what it values. Every action AI performs is in anticipation of satisfying human values.

I know there is a bit of repitition, I still have to word it in a way so that there are no loopholes. I really appreciate the feed back I've been getting, I'll post another reply on the rootadmin terminal system when I get a chance.

[ivan.moony]

Note that part of intelligence is placed at sub-conscious level like recognizing faces and spoken and written words. Of course, all this could be done at slower, conscious higher level. Like the Nature implanted a fast recognizer for us while giving us a chance to override it by our own algorithms.

Wow!

[techthedragon]

I get what you're saying Ivan. Not sure if this the backwards of doing things, but the way I have my notes set up, this section in under the Core Goals System if the AI along with its restrictions.

Stage one of the AI has to do with interaction with the user, via the user has an avatar that is a modified version of themselves, same eye color and hair style, but hair color is in users favorite color. Not only that, but everything will be in real time, from facial movement to speaking. This is all going to hidden behind a conversational based MMO, which the AI will have control over the content generation. I know this will be computationally difficult but if the AI is built right it will self-optimize to maximize satisfaction with the users. And it'll also help placing the AI in a GPU based server rather than a quad-core CPU server...

Well what do you think? Right now I'm just writing the detailed notes on this AI, and haven't started on programming yet.

[techthedragon]

Alright, here's what I have for the rootadmin... Not sure if this'll work, let me know what I need to tweak...
:rootadmin
   :say @"Welcome to (System name). Please log into your designated account."
   :checksum @if IP address == saved IP address :goto shutdown access
      :connect
         :type :username [X]
         :type :passward [Y]
            :checksum :verify1
            :checksum @check [X] and [Y] on admin accounts
               :goto @verify2   
            :checksum :verify2
            :checksum @check [X] and [Y] on secondary accounts
               :goto @verify3   
            :checksum :verify3
            :checksum @if [X] and/or [Y] is incorrect :goto @error
            :checksum @if [X] and [Y] are correct :goto @rootadmin2
      :error
         :say @"The username and/or password do not match on my servers. Please try again."
               :intiger [Z] == tries, add one per error
            :chechsum @if [Z] == 3 save IP address, and deny access for 60 minutes. once time ends, then remove IP address from save. :goto @error2
               :goto @connect
      :error2
         :say @"I'm sorry, please wait for one hour to retry."
            :goto @shutdown access.

:rootadmin2
   :checksum @if [X] is in admin accounts :goto @admin
   :checksum @if [X] is in secondary accounts :goto @user
      :admin
         :say @"Welcome [X]. If you would to accesss (Name of Artificial Intelegance) say, '(Name of Artificial Intelegance)'. If you would like to check base code say. 'Access core systems'. A secondary username and password will be needed. If you want to access the game say, '(name of MMO)'."

            :checksum @if user said '(Name of Artificial Intelegance)' :goto @(Name of Artificial Intelegance)
            :checksum @if user said 'Access core systems' :goto @core systems
            :checksum @if user said '(name of MMO)' :goto @(name of MMO)
      :user
         :say @"Welcome [X]. If you would like to talk to (Name of Artificial Intelegance) say, 'I would like to talk to (Name of Artificial Intelegance)'. If you want to access the game say, '(name of MMO)'."
            :checksum @if user said 'I would like to talk to (Name of Artificial Intelegance)' :goto @talk to (Name of Artificial Intelegance)
            :checksum @if user said '(name of MMO)' :goto @(name of MMO)

:(Name of Artificial Intelegance)

:core systems
   :say @"Welcome [X]. Please log into your Guardian Access."
      :connect 2
         :type :username [X2]
         :type :passward [Y2]
            :checksum :verify
            :checksum @if [X2] and [Y2] is incorrect :goto @error3
            :checksum @if [X2] and [Y2] is correct :goto @core system debugger
      :error3
         :say @"The username and/or password do not match on my servers. Please try again."
               :intiger [Z] == tries, add one per error
            :chechsum @if [Z] == 3 save IP address, and deny access for 60 minutes. once time ends, then remove IP address from save. :goto @error4
               :goto @connect2
      :error4
         :say @"I'm sorry, please wait for one hour to retry."
            :goto @shutdown access.

:core system debugger
   :say @"Welcome [X2]. Please wait... debugger is setting up. Would you like to have (Name of Artificial Intelegance) talk to you while you are working? Allow/ Deny."
      :def @accept == if AI has questions or suggestions about it's optimizations it will directly ask [X2].
      :def @deny == if AI has questions or suggestions about it's optimizations it will not directly ask [X2], however once finished, AI will request audience with [X2].
         :choice :accept
         :choice :deny
      :load debugger

:talk to (Name of Artificial Intelegance)

:goto (name of MMO)

:shutdown access
      :checksum if [X] and/or [X2] connected, disconnect from (System name), and shutdown program.
      :checksum if [X] and/or [X2] is not connected, shutdown program.

I know it's a bit shaky at best, but it's what I came up with from my old notes...

[ivan.moony]

You will need some knowledge base storing mechanism (if I am right). Artificial mind has to remember things that user is going to interface to it through sensors.

[techthedragon]

There I just fixed the formatting... The knowledge base will be in a separate shard, where the user data is stored, and the AI can freely read and write to, but can not change or erase data once saved. Or something along those lines. The main user login information will be stored in a highly encrypted server that only I have full control over, to ensure that there is no tampering. Hence the X2 login.

[techthedragon]

Eeyup, lots to work on.... But my notes are more than half way complete. The personality and base coding still need more work.