I figured robots haven't gained momentum at anywhere the same rate as personal or distributed computing since robust sensing and interaction of the physical world is still mostly an unsolved problem. Robots are useful in carefully physically constrained areas like industrial assembly lines and can be finicky toys that are too limited to do much anything useful, and there are things like Roombas that are somewhere in between the two.
A solution to the robust physical interaction thing sounds like something that would be pretty much AI-complete and would imply a lot bigger changes than housecleaning robots. This would just get you the science fiction humanoid robot that can walk into a house and start cleaning up and renovating the walls, but given the amount of flexible smarts and learning ability a general solution for being able to pull that off involves, it could probably start doing your job for you as well.
I'd guess there is also a less scary technology branch, where the current limited scope robots become somewhat more feasible with better manufacturing, materials, computing and communication technology, taking use of things like human telepresence and heavy duty data mining that have become easier due to just plain Moore's Law instead of groundbreaking AI research. These kind of robots are going to require environments tailored to their limitations, like the existing assembly line robots, as well as doing the R&D on figuring out just what kind of robot and environment combinations are going to do useful stuff, so it's not going to be the sort of lightning-fast development we saw with personal computers. More like the gradual development from 19th century homes to late 20th century homes full of electrical household appliances.
In fact, electrical household appliance probably is the metaphor for thinking about these sorts of robots, not a metal man walking about and enslaving humankind. That might explain why people aren't vastly enthused about them, even though there's probably a market for them.
Comments
I figured robots haven't gained momentum at anywhere the same rate as personal or distributed computing since robust sensing and interaction of the physical world is still mostly an unsolved problem. Robots are useful in carefully physically constrained areas like industrial assembly lines and can be finicky toys that are too limited to do much anything useful, and there are things like Roombas that are somewhere in between the two.
A solution to the robust physical interaction thing sounds like something that would be pretty much AI-complete and would imply a lot bigger changes than housecleaning robots. This would just get you the science fiction humanoid robot that can walk into a house and start cleaning up and renovating the walls, but given the amount of flexible smarts and learning ability a general solution for being able to pull that off involves, it could probably start doing your job for you as well.
I'd guess there is also a less scary technology branch, where the current limited scope robots become somewhat more feasible with better manufacturing, materials, computing and communication technology, taking use of things like human telepresence and heavy duty data mining that have become easier due to just plain Moore's Law instead of groundbreaking AI research. These kind of robots are going to require environments tailored to their limitations, like the existing assembly line robots, as well as doing the R&D on figuring out just what kind of robot and environment combinations are going to do useful stuff, so it's not going to be the sort of lightning-fast development we saw with personal computers. More like the gradual development from 19th century homes to late 20th century homes full of electrical household appliances.
In fact, electrical household appliance probably is the metaphor for thinking about these sorts of robots, not a metal man walking about and enslaving humankind. That might explain why people aren't vastly enthused about them, even though there's probably a market for them.