>Why
>is it that it seems to most people more promising to build up smart
>cooperative agents from scratch, as in an AI approach, than to
>domesticate existing very smart but not cooperative-enough agents? Is
>domestication really that hard compared to learning how to organize
>intelligence and acquiring all that common sense knowledge?
Partly it's that we've tried and failed to replace humans with animals in the
routine but abstract purposes we're discussing here (cleaning, caretaking,
machine operation, etc.) We have a new possibility so we're giving it a go.
It's worth a try.
The other advantage to building the agents is that they might be far more
easily duplicated. Monkeys are cheaper than people, but a well-trained
monkey still requires a lot of investment and upkeep. Further, animals are
more variable than machines and that doesn't sit well with industrial design
theory, which works hard to minimize nonessential variability.