If a goal is “clearly definable”, it’s specific.

What we want for AGI is a conceptual system that functions like or mirrors 
language -  no concept in language is “clearly definable” – hence philosophers 
can go on for ever about the definition of any concept.

We want a robot that can “GO TO THE KITCHEN” – those are general, not clearly 
definable, specific concepts/goals – and that is how an AGI like you can use 
them to GO via any mode of travel to any kind of KITCHEN.

“Clearly definable” and “specific” are dirty words for AGI.

From: Jim Bromer 
Sent: Saturday, February 09, 2013 11:31 PM
To: AGI 
Subject: Re: [agi] Could Algorithm Generators be a Feasible and Effective AGI 
Method?

No, I am not thinking about specific goals.  So you are wrong.  I believe that 
every concept is relatively general. So perhaps this is where your 
misunderstanding is coming from.


On Sat, Feb 9, 2013 at 5:47 PM, Mike Tintner <[email protected]> wrote:

  It’s suicidal –in the real world. You won’t be able to adapt to new 
situations  (“sorry Chinese is off tonight”).

  You’re thinking in terms of specific goals.AGI is about general goals -  all 
living creatures are driven in the first instance by general drives and general 
goals- not specific ones.

  Narrow AI is always specific – and your & every other AGI-er’s urges are 
always to be specific – whereas AGI is the opposite.

  From: Jim Bromer 
  Sent: Saturday, February 09, 2013 10:36 PM
  To: AGI 
  Subject: Re: [agi] Could Algorithm Generators be a Feasible and Effective AGI 
Method?

  It is not suicidal to set goals and a remark like that shows that you are on 
the wrong track.  Just because the goals that we are most interested in may be 
elusive that does not mean that we cannot use goal strategies to help us learn 
what is achievable and acceptable.  The irony of course is that once we 
discover more realizable goals we usually find that our previous goals were not 
only not realizable but they were really not that desirable.

  Pursuing unclear goals are not the path you choose.  Evidently, you have to 
decide whether that is a path you would like to take.


   
  On Fri, Feb 8, 2013 at 10:35 AM, Mike Tintner <[email protected]> 
wrote:

    Jim: So if a goal is clearly definable (for an AGi program)...

    No goals are clearly definable in AGI – at the program level as opposed to 
in particular situations. What are the goals of real world activities – like 
conversation/ reading [science/lit..] books..? You can set goals in particular 
situations ... “I want Chinese tonight”  But in general, it would be suicidal 
and isn’t possible.  AGI is about creative machines dealing with a creative 
world – in which both the machine and the world keep changing, and there is no 
way of predicting what food (or other goal instantiations) will be available, 
or what of the vast untried number of cuisines might be worth sampling, or what 
condition/state your body will be in and therefore what foods it will be 
attracted to .  The goals of an AGI have to be general, vague and capable of 
continuous refinement - and new combinations with other goals.
          AGI | Archives  | Modify Your Subscription   


        AGI | Archives  | Modify Your Subscription   

        AGI | Archives  | Modify Your Subscription   


      AGI | Archives  | Modify Your Subscription   



-------------------------------------------
AGI
Archives: https://www.listbox.com/member/archive/303/=now
RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-f452e424
Modify Your Subscription: 
https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-58d57657
Powered by Listbox: http://www.listbox.com

Reply via email to