Remote-controlled demonstrations, low success rates, sceptical experts: Jamie Condliffe puts the promises of a humanoid robot in every home into perspective.
In an analysis published on October 8, 2026, Jamie Condliffe confronts the industry's enthusiasm with the reality on the ground. Morgan Stanley imagines nearly a billion humanoids in 2050 for a market of more than $5,000 billion; but the techniques that have transformed language understanding do not easily transpose to the physical world, where other approaches will undoubtedly be needed.
Deceptive demonstrations and unreliable robots
The author reminds us that several spectacular demonstrations, notably those of Tesla's Optimus, were based in part on human piloting. And reliability remains the crux of the problem: for Marc Raibert, founder of Boston Dynamics, a robot that succeeds 70% of the time is like a robot that doesn't work. Concrete deployments, such as those of Agility Robotics in warehouses, remain limited.
However, the article notes encouraging signs, such as Physical Intelligence's π0.7 model, which is capable of combining know-how for tasks never seen before in training. But skeptical voices dominate: Yann LeCun believes that roboticists do not yet know how to make robots really useful, and Jonathan Hurst, from Agility, estimates that it will take about ten years before they can provide real services at home.
Read the full article on the original source
Go to the source →