Loading ...
Sorry, an error occurred while loading the content.
 

Re: On consilience and literary Darwinism (weblog post) - the `10-st

Expand Messages
  • tintner michael
    Hi JT, You re trying to blind me (and yourself) with a lot of bibliography, which actually isn t relevant. The issue is this : is there a formula/algo for any
    Message 1 of 1 , Aug 25, 2013
      Hi JT,

      You're trying to blind me (and yourself) with a lot of bibliography, which actually isn't relevant.

      The issue is this : is there a formula/algo for any genre of art  (or cooking - or indeed any human activity whatsoever).?

      Literature on creativity is irrelevant. Any serious discussion on creativity will have to acknowledge that the literature is endless, and there are no accepted definitions of creativity - and there is certainly no such thing as a "scientific" definition - it's actually a philosophical matter. You talk about the Singularity field (and presumably by implication the field of AGI) - the discussions in this field as to the nature of "intelligence" are endless - see Hutter's article on the matter - wh. quotes endless definitions and redefinitions of "intelligence". The field has no accepted definition of intelligence or creativity)

      The key sub-issue is whether you understand - and are giving proper examples of - an *algorithm*.

      With respect, you flat out don't - and aren't.

      You think, as I indicated, that an algo is equivalent to vague general directions for a course of action. Hence you say: "Michael you are still conflating the small, `granular detail' -- where randomness emerges - with the `big algorithmic structure'...  ("big algorithmic structure for you, being vague general directions).

      An algo has to define the "small granular detail" - it has to define IOW every single action on every single object a machine (or human) takes within a given course of action. So a storytelling algo has to - and does - define every single word as well as sentence of the stories it writes. There are indeed storytelling algos, which can say produce variations on stories about a prince, a frog and a princess - and passable stories as end-products. And they define and produce every word of those stories - otherwise they wouldn't be algos. (They can't do anything else but produce variations on their recipes - ask them to introduce a *corrupt banker* and they won't understand you or be able to do anything).

      Any higher-level instructions in an algo *must* be automatically, deterministically convertible into precise specific instructions/actions.

      You see algos, as I indicated, basically as "verbal formulae" - purely general instructions. Hence your story "algo" :

       "a /story algorithm /is such as this:

      */Start:/*
      */(1) A [Character/s] - has
      (2) a [Problem/s] - and
      (3) experiences [Conflict/Obstacles] - in solving it/them (the 
      Problem/s)...
      (4) A [sacrifice] must be made by the [character/s], that (either):
      (5) finally, resolves (or - does not resolve)/**/: the Problem/s.
      - End.
      /**//*"

      That is a set of general concepts/ideas which are in no way instantiable by any machine (or human) - or deterministically convertible into any actions.

      What is a "character" - oh just an infinity of possible human beings. Oh take any, you say. Great, I'll have a "corrupt banker, Jamie Dimon."

      What is a "problem" - oh just an infinity of possible human problems. Oh take any, you say. Great, "=I'll take "is pregnant, and doesn't know what to do."

      Great story huh?  "Jamie Dimon is pregnant and doesn't know what to do"

      Your algo - which is no such thing - will produce nonsense.

      I suggest you read up about concepts, because they are a completely unsolved problem in AI/AGI. Language use is also a completely unsolved problem in AI/AGI.

      And GO TO THE KITCHEN



    Your message has been successfully submitted and would be delivered to recipients shortly.