Actual strategist here (I automate things humans do for a living). I agree! Building a machine to do what humans are good at is dumb. Building a machine to do

Actual strategist here (I automate things humans do for a living). I agree! Building a machine to do what humans are good at is dumb. Building a machine to do what humans are not good at is smart. Humans are great at noticing things as they walk around. Machines are OK, but miss the context way too frequently. If a person can do it alone, it'll be really hard to make a robot better at it, at least in the real world. We should build more machines that are good at what machines are good at, and let humans fill in the blanks around those machines. For example, a robot that unloads heavy stuff from trucks would save hours of time so that human can move stuff to the shelves (and answer questions from shoppers while they're out there). Lo and behold, the profitable robotics companies here and elsewhere do that: move heavy stuff around (and only that, and only in warehouses where things are controlled). Now, arguably their case is that they want something "acceptably worse than a human for much less cost", but they erred by missing all the different things a human has to do while on shift, and building up a whole machine for this, the one thing that humans do without thinking.