The prevailing worldview is that robots will eventually do most of the physical work in America. There is clear evidence to support this view. The fact is that the US has greater manufacturing output now than ever, while there are fewer workers in manufacturing now than before. Advances in manufacturing technology – specifically automation – are what made this possible.
Some people believe that the best way to prepare for this future robot economy is to save their money and buy a highly capable robot – when such robots become available. The idea is that they can hire out their robot to whoever needs work done for which the robot is suited. This way people can participate in the robot economy by making money off their robot. After all, people will need some sort of income since everyone will be out of a job in the robot economy.
Naturally, the smarter and more capable a robot is, the wider the variety of work it will be able to perform. No one would want to own a robot that is suited to just a few tasks since those tasks, just like human manufacturing jobs, may be taken over by highly specialized machines. Herein lies the conundrum regarding the capabilities of these robots. Highly specialized machines take on whatever appearance and configuration they need in order to perform a task with the best speed, accuracy, and efficiency possible. Yet, it would seem difficult to build a robot to do such a job while also being sufficiently capable to do many other specialized tasks. In other words, how is a general purpose robot able to compete with highly specialized machines?
One approach to reducing this challenge is the use of modularity and re-configurability in robot design. For example, the “brain” part of the robot could be in one module while the moving parts could be in several other different modules, depending on the nature of the motion needed for certain tasks. A module with wheels could be used for rolling, while one with “legs” could be used for walking, or one with “arms” and another with “hands” could be combined to pick up and assemble parts.
The down side of this modularity approach is that robot owners would then need to decide which modules they should buy or possibly rent. A solution to this might be to buy a large collection of modules, but this would likely be too costly for all but the wealthiest people.
My thinking is that maybe owning a robot is not really the best way to participate in the robot economy. Instead, what if one owned stock in the companies that built robots? That makes the decisions a lot easier. One would have to live off the dividends but they should be significant. Even if robots start making other robots, which will surely happen, these companies will remain at the helm.