For today’s readers, 20th-century recipes involving Jell-O molds—where fruit, vegetables, or even seafood is suspended in a mass of gelatin—can be creepy enough. But AI researcher Janelle Shane, known for her AI Weirdness blog, announced on Twitter Friday that she trained a bot to develop new recipes in that style, creating what she says may be “the worst recipe-generating algorithm in existence.”
WHAT HAS OCCURRED CANNOT BE UNDONE
I have trained a neural net on a crowdsourced set of vintage jello-centric recipes
I believe this to possibly be the worst recipe-generating algorithm in existence pic.twitter.com/cwQwOpUNDv
— Janelle Shane (@JanelleCShane) February 7, 2020
The recipes look like harmless vintage weirdness until you look closer. Then, you’ll notice that one calls for twice removing “internal rinds” from crab meat and another flavors lemon Jell-O with a mix of cheese, peppermint, and pickle relish. The bot also designed an innovative blended cocktail involving Splenda, Kentucky Bourbon, and a block of chocolate, which might be served to accompany these meals.
NO neural net
that is too much extract pic.twitter.com/BCRxKiMVtC
— Janelle Shane (@JanelleCShane) February 7, 2020
Yet another recipe involves “bunny’s food,” along with walnuts and pimiento, and has the appetizing name “Potty Training for a Bunny.” That’s not to be confused with a previous name dreamed up by Shane’s AI systems, the Madland Cookie, which wasn’t for a food product at all but for a racehorse.
Recognize your brand’s excellence by applying to this year’s Brands That Matter Awards before the early-rate deadline, May 3.
