You don't need a bigger machine. You just need more machines. Think parallel. A bunch of little snow-eating-ice-excreting robots is more nimble and effective than one big machine. There have been plenty of projects with robots that work autonomously, form ad hoc networks, coordinate with each other without human intervention, and go back to base when their ice beds are full or their batteries are getting low.
Not bigger machines, more machines. The hive is the organism. Build enough, and some are always recharging while others are working. They can run 24/7/365. And when one place needs fewer and another needs more, you throw them on a truck or a train car and ship them to where they're needed.
Picture a swarm of robots acting in a seemingly random way, but with an internal logic and a human override that can send them all back to their docks. A beehive and an ant mound are remarkably efficient models of cooperation among simple, small, weak creatures. Replicate that with machines, and you've got something.
It's a fundamental re-thinking of the whole model of automation. Instead of building a few big machines that function like humans, build a lot of small machines that work together like an ant mound or a prairie dog colony or a flock of geese or a herd of elk.
I don't have to I work there minus the internal logic.
Absolutely correct. The Japanese are at the forefront in the coming field of popular robotics, where robots serve the ordinary person. In ten years time, we’ll see astonishing accomplishments. I desperately need a robot to clean my house - my wife is threatening to divorce me if I don’t get the first one off the assembly line.