http://kingjamesprogramming.tumblr.com/

Someone wrote a markov chain and trained it on the contents of King James bible and the contents of various programming text books. The generated outputs are published on this blog.

Some of them are hilarious, to show a few:

19:19 Yet there is no assignment, no local state variable, and consequently none of the servants of your lord

9:14 For I will not dare to speak of a procedure that computes the square

What is man, that he should not depart from Jerusalem, but wait for the carry signals to propagate. What is the purpose that is purposed upon the whole earth: and this is registered by the probe on F.

17:7 For so the procedure body does a direct transfer, without saving any information about how they are written among the children of Israel

25:32 And David said unto Saul, The LORD sent me from Kadeshbarnea to see the value of the variable search that can be stored in a very efficient manner.

28:11 He bindeth the floods from overflowing; and the thing was not done in a data-directed style.

25:32 And David said unto Saul, The LORD sent me from Kadeshbarnea to see the value of the variable search that can be stored in a very efficient manner.

I think this is my favorite.

And in the morning, then they journeyed: whether it was true or false each time.

6:24 And the LORD said unto Satan, Hast thou considered my servant Job, that there is a “vertical” barrier that gives us the power to model systems composed of objects

And Satan stood up against them in the global environment.

Oh my god, I really cannot stop laughing

Each processor would proceed sequentially as if it had been better for them not to rise against Saul.

I'm going to start doing this:

The LORD is the beginning (or prefix) of the code for the body of the procedure

I must admit, I loled.

I once implemented something similar, IIRC using an order-3 model (picking the next word based on the 3 prior words), and trained on random bits of fiction and similar I had written.

it generated similarly nonsensical / lolz-inducing statements.

IIRC, I had once imagined possibly something like this being used to generate randomized dialogue for NPCs (maybe combined with speech synthesis).

so, for example, a person could have a collection of "lore" (general descriptions of the game world, ...), maybe some random stuff about whatever town/city/... the NPC lives in, collections of various generic "day to day conversation", ... then let the NPCs loose to say whatever random thing pops up.

getting extra fancy, maybe the Markov Models could be combined with Neural Nets, and the NPCs could have some basic ability to communicate (input neurons could be linked to "heard" words, and output neurons could be linked to words for them to want to say), though this would likely require some sort of training algorithm or something (such as to hopefully get them to mimic conversation).

well, and probably trying to keep all this cheap enough that a bunch of NPCs wandering around and talking doesn't kill the CPU or eat up too much RAM.