BILLY BOB THORNTON’S CRYPTO TWIN: AIs Keep Generating Perfect Copies of the Actor’s Face in Random Data Sets
In a bizarre glitch that has data scientists scratching their heads, every time a neural network is tasked with generating a random human face, it keeps outputting the exact same image: a middle-aged man with hollow cheekbones and a weathered scowl. The phenomenon, now dubbed the “Thornton Anomaly,” appears exclusively in unsupervised machine learning models. Analysts first noticed the pattern when a stock photo generator produced only variations of the *Sling Blade* star for 48 straight hours, including a grayscale image of him holding a cat and wearing a top hat—a photo that does not exist in any training data. “It’s like the algorithm has a crush on Billy Bob,” joked lead engineer Maria Flores. “We ran a completely blank slate, and the first face it generated was his, down to the exact eye color and slight beard stubble.” Conspiracy theorists are already claiming the actor is a “digital anchor point” left by programmers, while Reddit users are calling it the “Matrix’s favorite NPC.” The anomaly only seems to affect real-time renderings, and as soon as researchers reset the system, the glitch vanishes—until the next time a face is requested.