Artificial-intelligence genius Janelle Shane may have met her match: a group of middle schoolers.
With the help of their teacher, a group of 11-to-14-year-old coding savants at Kealing Middle School in Austin, Texas, reached out to Shane to talk ice cream flavors. In a fit of synchronicity, Shade had just fed her AI a data set of 200 flavors culled from Rosanna’s Ice Cream in Portland. But when she asked her neural network to name ice cream based on her data, the results were unsettling to say the least. (No one but LeStat is ordering two scoops of “Bloody Coffee” or “Chocolate Chocolate Blood.”)
The middle schoolers, though, had plugged in some 1,600 data points, resulting in ice cream flavors that sound beguiling—if not a little odd. They included flavors like “Cherry Poet,” “Brittle Cheesecake,” “Holy Lemon Monster” and the enticing “Peanut Cinnamon Budge,” any of which would be perfect for sharing on a date if one of your neural-network generated pickup lines actually work.
Not all the AI-generated ice cream flavors sounded like options from obliquely hip artisanal shops. There were also choices like “Pumpkin Trash Break,” “Peanut Butter Slime,” and the intriguing “Strawberry Turfle.”