Originality’s Never Too Novel

Run plots through a baffle box
The way backgammon addicts
Roll their dice. What comes out then,

Events randomized, winnow
By offering sequences
In text screenings to random

Selections of anyone
You can drag in to read them.
Rate and repeat and repeat.

Here’s the question—do you think
You’ll get same or similar
Plots back from such processing

As the plots you started with?
Do you think they will converge
On stories you’ve never seen?

Every algorithm tests
Skill at manipulation,
But the best would be to get

Whatever’s not been gotten
Yet, but seems perfectly kin
To what’s liked best of what’s been.