Coachella is three months away, and you’ve probably already forgotten the lineup, right? (Besides Beyoncé. We never forget her.) Well, if an artificial neural network (i.e., a computing system) trained on the names of thousands of bands produced a lineup, would you really know the difference?
Botnik Studios, an open community of writers and artists working with machine learning and predictive text, recently tested this out. On Tuesday, it unveiled the results. And they are amazing:
It’s hard to choose a favorite. I thought I was more a Skin Clump gal, but as I’ve gotten older and mellowed out a bit, I’m feeling a little more Bing the Bung. And I’m almost positive I saw Bowel Prank in South Florida in the late ’90s.
Botnik has done other nu-content creation with bots. In December, it used predictive text to write a new Harry Potter chapter called “The Handsome One,” and its bots have rewritten Scrubs and X-Files scripts.
The idea came from Botnik creative director Michael Frederickson.
He was “inspired by the feeling of digging for your favorites when festival lineups come out, and that feeling, no matter how culturally in touch you are, of thinking, ‘Who the fuck is that band, it sounds like blank verse dada poetry,'” Frederickson said. “We wanted to really play up that feeling with an overwhelming number of unknown bands that we liked the sound of.”
Odds that a real band will actually take one of these names are very high.