In humans, the beer is usually fetched first and then the baby-making follows sometime later. In robots, the process seems to be reversed. Engineers at the University of Cambridge built a “mother” robot that is able to make “baby robots” on her own, test them and pass preferred traits down to them. Later, researchers at MIT announced the development of a team of robots that work together to fetch beer. Should we get them all together or make sure they never meet?
The baby-making robot isn’t as scary as it sounds (yet). The Cambridge scientists used a robotic arm as the mother and motorized plastic cubes as the babies. The mother robot assembled the babies into movable robots with human involvement and watched how far they could travel.
It (she?) then reassembled the babies in new configurations that improved the baby’s mobility. After ten generations, the youngest baby robot moved twice as far as the oldest one. It sounds pretty simplistic but PLOS One reported that study leader Dr Fumiya Iida was excited:
Natural selection is basically reproduction, assessment, reproduction, assessment and so on. That’s essentially what this robot is doing – we can actually watch the improvement and diversification of the species.
Can the babies eventually learn to fetch beer? That’s where the team from MIT comes in. They built three robots: two mobile “waiters to take orders and deliver the beer and one with robotic arms to grab the beer from the fridge and place it in the waiters’ coolers for delivery. The key was a special algorithm that helped the robots work together without knowing what the other robots were doing. This high-level reasoning algorithm probably has many uses, but the money is in beer delivery.
What would happen if the robot mother arm met the robot with beer-picking arms? Would they share a brew and create babies that delivered beer? Would the babies eventually learn to mix drinks? Call a taxi for their drunk parents?
If this doesn’t convince more people to go into robotics, nothing will.