[pasabagi]

I think Dijkstra made a point that when somebody sees something new, they try to project something they know onto it - so with computers, the first thing people tried to do was make machines that thought like people- a project that's still bearing very little fruit. In the meantime, the things you could do really well with computers, like discrete maths, have been fabulously useful.

I suspect robots are a bit like that. They make very poor substitutes for people. But, they can do a lot of stuff ('dishwashers', for example), that are basically orthogonal to what people do, but reduce workload immensely.