Every wave of impressive robot demos brings the same fear back: that humanoid machines are about to walk into factories and warehouses and take the jobs. News this week that Unitree priced a landmark humanoid-robot IPO will sharpen it. The worry is understandable. The evidence for a jobs apocalypse, so far, is thin.
Start with what is actually deployed. Real humanoids are on real floors: Figure's machines at a BMW plant in South Carolina, and Agility Robotics' Digit units working shifts for Amazon, Toyota and the logistics firm GXO. But the verified Western fleet still numbers in the low hundreds of machines, not the millions the panic implies. These are pilots and early deployments, not a workforce.
What the robots are actually doing
So far the machines are filling gaps rather than emptying payrolls. There is no documented case of a company laying off large numbers of workers and replacing them with humanoids. Instead they are being pointed at jobs employers already struggle to fill. Warehouse work carries high turnover and hundreds of thousands of unfilled US positions, and the first robots are being slotted into those holes rather than into occupied chairs.
There is also a quieter effect running the other way. A humanoid on a floor today needs human minders: people to handle exceptions, correct its mistakes and keep it running. At some sites the workers standing next to the robots have been promoted to oversee them. The machines are creating jobs even as they threaten others.
The hard problem is hands
The deeper reason to doubt the apocalypse is technical. Walking is largely solved. Manipulation is not. The dexterity and on-the-spot judgment that most physical jobs quietly depend on remain well beyond current robots. The roboticist Rodney Brooks, who has spent a career building them, argues that robot hands will lag human ones well past 2036. Jobs that need fine motor skill or improvisation are safe for years, not months.
None of this means the technology is harmless. Displacement is real for narrow, repetitive tasks, and the picture will shift as fleets grow and costs fall. But that is a slow, uneven change, closer to the gradual pressure on entry-level office work than a sudden replacement of the labor force. The honest answer to whether humanoid robots will take everyone's job is this: not soon, and not the way the demos suggest.
Commentarii · 0