I’ve been a bit obsessed over machine-made art, not art that machines help to make, but machines learning how to make art. I’m currently very much in love with Ross Goodwin’s work in this field(see: Adventures in Narrated Reality, Part 1) . A bunch of monkeys in a room typing may eventually produce the complete works of Shakespeare, but a bunch of robots being trained to deconstruct the linguistics of poetry and create new work…that’s already here, and it is far more interesting.
(Goodwin’s work is much more compelling, but here is the inevitable link to a shitty robot-written Christmas carol.)
Sadly, my first entry into this field is as someone who can’t quite code well enough to parse language, so I am currently dependent on the sheer flukes of phrase randomization. @PythiaBot is a fortune telling bot, seeded with a series of odd phrases that are all syntax compatible. It is a bit repetitive, but it has generated some surprisingly lovely sentences.
@trashhaiku is vaguely political art – it remixes Donald Trump’s incoherent twitter feed into similar incoherent haiku. It is currently hand fed, by me reading the original twitter feed to find phrases that match(more or less) a syllable count. I have to admit that I am not sure I understood what “suffer for your art” meant, but I do now.
Both are built on Zach Whalen’s Google Spreadsheet twitter bot code, which is remarkably simple to follow.