I think the interesting question is are there scenarios where the practically achievable precision is good enough that the other benefits of analogue/mechanica

I think the interesting question is are there scenarios where the practically achievable precision is good enough that the other benefits of analogue/mechanical results in a net win. This is a point pondering hard about. Especially so, when we ourselves in the midst of programming with differentiable functions/programs approximated on digital computers. One can see that one of the tools that will help us grapple with this question is rate distortion theory in general and quantization in particular.