Re: 80-20 rule?
Take 100 expert surgeons, each doing 20 operations. Video those 2000 operations. From that program a machine to do them, with greater dexterity and vision than a human. With machine learning, soon you have a super-expert.Use that to act as knowledge repository for every operation. Soon you don't need a surgeon, just a machine and a human standing by to take the blame if things go wrong/outside the machine's remit.
The 80% which go are the good jobs, the 20% which stay are the boring ones.