I posted this generative text project on Facebook but it's something I'll want to come back to again. These kinds of project are just so neat in every sense of the word. I know they aren't that complex in their implementation but it's all a matter of presentation and how context deeply influences our understanding of even perfectly uninspiring text. The source is all here on Github. A selection (anything this project generates is a selection, anyway, as with the Twitter implementation it's nearly infinite, though the possible maximum number of tweets comes to mind):

...I saw stretched upon his back, and gazing up straight at the terrible sun, the man I was seeking. I saw about Miss Vivian's death to-day, and I was afraid Hal would be all alone fretting. I saw that William Oke, in his heart, thoroughly looked down upon all his neighbours. I saw only a snapshot of her, which showed her to be beautiful. I saw smoke in the chimney this evening. I saw a man talking with a woman there, at your door. I saw some beauties there a while ago. I saw tears in their eyes, tears of joy for the honors paid me; and especially, said they, for the manner in which I had received them. I saw her in the prisoners' dock, the Katusha betrayed by me, in a prisoner's cloak, condemned to penal servitude through a strange mistake, and my own fault. I saw them! The damage amounted to seventeen rupees, eight annas. I saw my Indian fall to the ground. I saw it in his desperate face. I saw my daughter at the theatre in London. I saw her only to love her; nor was it a common passion she inspired in me. I saw absolutely nothing else on the floor...

And so on. The Twitter version obviously doesn't benefit from being studded with excess punctuation but it does feel more like an authentic stream of consciousness, or maybe a stream of collective unconsciousness.