1 LAB PUT A ROBOT ON A BICYCLE, A UNICYCLE AND A TILTING PLATE, AND IT'S ALL THE SAME PROBLEM.
55 seconds of clips out of a research floor. No stage, no product, no price.
A humanoid on a red tether, walking, then landing on a steel plate that rocks under it.
A machine on a bicycle, no rider, holding the line.
The same machine on a single wheel, hopping up wooden stairs, 1 contact point with the ground the entire time.
A 4-legged unit with an arm, working out of the back of a van.
An arm on a bench sorting green parts under a work light.
Then a screen with the same robot rendered in simulation, red and green dots marking the path it's about to take with its real body.
Run the list again and the pattern shows up.
A bike is unstable. A unicycle is worse. A tilting plate takes the floor away. Stairs take away the assumption that the floor is there at all.
This is the lab Boston Dynamics' founder built, and everything in the reel is 1 question: what happens when the ground lies to you.
The demos sell the walking. The labs are all working on the falling.