[yotamarker]

short story :
"For sale: baby shoes, never worn"
what would make the AGI continue the story ?

[yotamarker]

battle programming burst mode !

if person Xn tells person X(n-1)  a story and so on till N = 1 then
person's X(n = 1) story will most probably be much shorter then person's Xn in the vast majority
of D cases => stimulation of the story teller is key to puff up D story.

how do you like them apples ?

[yotamarker]

a story is built upon :
1 universal emotional experiences
2 details leading to those events
3 mutation based on new data (scenes)
 

[yotamarker]

a story must have continuity a connection between it's elements
the brain probably achieves that by inflating the stories associations
(the stimulated parts)   

[keghn]

 For sale baby shoes?
 A AGI would handle it in the way a human would. With that said, it depends on the age of the AGI.
 A young AGI would play with it. To get to know it and then grow up and put it to the side and move on to something
new. 
 A old AGI would make long term plans and would make make quick change if need be. If the shoe was in the long
term plan?

[Freddy]

I'd sadly wonder why the shoes were never used.

[yotamarker]

many attribute all sorts of meanings to the story of : "red riding hood"
all sorts of lessons are supposedly derived loosely.
but looking at the varies versions of the story including perraults it is pretty
obvious it's key stimulated elements are RUNNING and EATING nothing more !
at any rate the story has been so mutated over time that is it ridiculous to even say
it contains truths.

[yotamarker]

it appears writing instruments enable the writing of long scripts like paper or notepad.
a long 12 episode OVA is not written in a day or an hour => it is possible
to assume the brain "waits" for parts of the story to light or get stimulated.

[Art]

The AGI would need data and a proposed path from the programmer (or the story could proceed at a random selection).

This is only conjecture in line with the question and is not mean to upset anyone of a sensitive nature or easily offended by baby topics.

The data can only contain a few possible scenarios or outcomes in your example:

1. The baby died before it was born but the shoes had already been purchased.
2. The baby died during birth ...
3. The baby died after it was born...
4. The baby was born with no feet / legs...
5. Someone stole the baby's shoes and posted them for sale.
6. The AGI has an abundance of new, baby shoes for sale.

perhaps a few others but the point is there.

[yotamarker]

but what would make the AGI want to continue the story
she must be stimulated.
for example if the AGI was raised by a mgtow she wouldn't discuss
the diary of bridget jones

once the AGI has been stimulated she can continue the story with :

A a parallel algorithm story like the examples you GAVE ALL of which lead to "never been worn"
B a spra story ( a sub scene ) some call those fillers.

[yotamarker]

FYI I challenge any of you to prove me wrong 

[kei10]

Rather trying to solve a problem about why an AGI has trouble doing a story, let's think outside the box --

Let's focus on three things; How to do any story, and what starts a story, and what ends a story?

Just like Art's given example; To do any story, one would require relevant data. To acquire data, one would require looking for new information.

To look for new information, there are a two days; Look for information from sources, or try to form of new possible information alone by using existing data.

Like you mentioned, what starts any story is emotion. Acquiring any data, telling any story, as all of them are goals, they gives you something to thrill about.

But what ends a story, is when the information starts to get cut off and stop, it can get tiring, exhaustion, and possibly other relevant reasons. This is known as "writer's block". When this happens, our emotion gets shifted to the opposite. If the AGI experiences this, they would stop doing a story.

However, the story can be continued if one waits for new information. As you mentioned, by giving some time to one's mind, new idea can come alight if we try slowly. Although idea do not come by itself without trying. So, yeah it has to be simulated.

It can be either filler or continued cannon story depending on the idea. One would attempt to create a filler just to continue telling the story, if there is a time limit. Animes usually follows ones from Mangas. When a manga's chapters didn't deliver fast enough, the anime studios have to either put the anime on hiatus -- but that won't do well for their company, thus they create fillers to continue the story (that is why now anime studios usually wait for enough chapters, and only deliver 25 episodes nowadays before another season). But that's a poor example.

Although an AI may not get tired of telling a story, depends on the way it operates upon. But that may bring problems without setting a limit for the AI.

There. 

[Art]

Every story usually has an opening, middle and end, much like a chess game.

What would make the AGI want to continue telling the story? What would it have stopped in the first place unless it had specific instructions to do so?

The AGI should follow its programming and in this scenario it's ultimate goal is to tell a story. (I think).

Every story should have an ending (unless you're writing a Daytime Soap or similar non-ending saga).
The End is brought about by winding down...examining that the know or reasoned data has been found, delivered as being told in the story format and concluded by having the AGI's initial goal(s) met. End of story.

If you come back with some silly anime or video game explanation then it is game over for me. You throwing bits and pieces of these made up "puzzles" you contrive, rarely meet anyone's answer(s) or your expectations. You need to define the terms of the program of the AGI or the purpose behind it's programming instead of these random bits.

The fun / edutainment value of this, if any, has escaped me / us.

[yotamarker]

many times the story end is open, the brain has no stimulation or
maybe it has a length limit.
think about it man, most songs are 4 minutes long, most tv shows are 30 minutes long
most movies are 1.5 hours long, most speeches are 10 minutes long, novels are 12 episodes
or 100 pages long.

it could also be because we need to go to the toilet or eat every set amount of time.

and yes my current view of A.I is connect puzzles solutions to power up the AGI, more puzzles = more power
like connecting zords.

[keghn]

 AGi mostly compares one object from another or one storey from another from another and finds what is the same and then what is different.
Not really good about photographic memory. Just enough information to tell them apart.
 Then within the story it finds thing that are the same or similar and then what different.
 A AGI brian is more interested in using a transformation algorithm then remembering something.
 Like remembering the bare minimal and then decode it with a transformation algorithm, decompression.
 Or screws remembering it and use a trained internal transformation algorithm on written words.:-)
 Neural Networks do this extremely well.