Figure AI had one of its robots race a human to sort packages. It lost.

In an intriguing contest that pitted human skills against advanced robotics, a human intern narrowly triumphed over a humanoid robot in a package-sorting challenge. Over the past week, the $39 billion robotics company, Figure AI, has captivated audiences with a live feed showcasing its robots sorting packages at its San Jose facility, drawing millions of viewers. The company raised the stakes this weekend by inviting an intern, Aimé Gérard, to compete against one of its robots in a grueling 10-hour sorting race. Gérard, a visualization specialist, initially maintained the lead until biology intervened. After a bathroom break around the five-hour mark, the robot surged ahead, taking advantage of the time lost by the intern. However, Gérard quickly regrouped, and despite the toll of the challenge that left him with blisters, he ultimately sorted 12,924 packages—192 more than his robotic competitor. He achieved an impressive average of 2.79 seconds per package compared to the robot’s 2.83 seconds. Brett Adcock, founder and CEO of Figure AI, humorously remarked on social media that this victory might be the last for a human. The ongoing event, which features a fleet of humanoid robots sorting packages continuously, aims to demonstrate the reliability and endurance of Figure AI's technology, with plans to operate for as long as possible. Since launching the livestream, the robots have sorted over 30,000 packages and attracted more than 3 million views, with viewers affectionately naming them Bob, Frank, and Gary. The event has been described as "surprisingly addictive" and even likened to a form of robotic ASMR. Merchandising efforts are underway to capitalize on the event's popularity. While the robots' ability to function for extended periods without failure is commendable, experts like Ayanna Howard, a prominent roboticist and dean at Ohio State University, caution that the technology is still not ready for full-scale deployment due to issues such as misaligned package orientations and items falling off the conveyor belt. "We’re still a long way from achieving fully autonomous humanoids in logistics environments," Howard stated.