A professor at Georgia Tech decided to source all the movie, book, TV and video game titles and summaries on Wikipedia to power a neural network. Then, he asked it to name the next blockbuster movie or hit TV show. The results were … a bit mixed.
Prof. Mark Reidl used 112,936 titles and synopses to power his little project, and it cranked out some positively hilarious (and a few semi-plausible) titles for films literally built piece by piece by its components. As areas of media are made by committee more and more, this is a fascinating (and pretty funny) peek into the world we could be living in once Skynet starts making the programming decisions. You know it's only a matter of time before we get that Star Wars Christmas movie, right? The special was just the beginning.
Check out a longer list of highlights here, and we’ve cherry picked a few of our favorites below for good measure:
Pirates: A Fight Dance Story
Cannibal Spy II
American Hero: Fire of Crusty
Legend of the Experience of Scarlet Freedom Damageboo
Under the Daleks
Batman and Flancles: The Fun Tree
The Star Wars: The Santa Contact
Swords and Batman: Summer Party ?
Barney’s The Devil’s Treachery
Merry Scroobers: Crown of Evil
The Steel-Pounted Murder King