Maybe the final straw is when they start ruining our childhoods, as the There’s Waldo robot seems determined to do.
Built by the creative ad agency Redpepper, the Google machine learning-driven robot utilizes facial recognition to absolutely dominate the classic, analog hidden image game. And honestly, it’s kind of rude.
In a video that seems to mock the pathetic ineptitude of all humanity, There’s Waldo finds the sneaky bespectacled cartoon man in a whopping 4.5 seconds. When it finishes scanning each face on the page, its silicon hand points to Waldo with an implied attitude of, “You morons.”
Or maybe we’re just projecting that last part.
There’s Waldo robot consists of a robot arm that’s powered by a Raspberry Pi computer chip. But the key element to its puzzle-solving mastery is the camera enables the robot’s facial recognition abilities.
The process for how There’s Waldo works is actually really simple.
The camera takes a snapshot of the chaotic mess that is a Where’s Waldo? scene. Then, each face is captured, analyzed by the machine learning algorithm, and compared to the Waldo photos it was trained to recognize. Finally, the A.I. picks the face that best matches what its looking for, in this case Waldo.
It does all of this in mere seconds — with over 95 percent confidence, apparently. When the job is done, the silicon hand fitted to the end of the robot arm promptly goes back into resting position, as if daring us to challenge it again.
The video shows just how easy it was to create this fun-ruining robot. Redpepper technologist Matt Reed told The Verge that they only had to train There’s Waldo with “62 distinct Waldo heads and 45 Waldo heads plus body” Google images in order to make it this accurate.
Sure, using machine-precise facial recognition on a Where’s Waldo? puzzle feels kind of like bringing a homing missile to a knife fight. But if it keeps actual knives out of the hands of robots, let There’s Waldo have its fun.