Full automation is not necessary. That's one of the easiest oversights I've ever seen in the business world.
I was part of a team once that desperately needed to do quality assurance testing as quickly as possible on a massive influx of software product. They were about to outsource quality assurance testing - and for a 6 figure amount - because we couldn't fully automate it in house. When I pointed out that we could automate 75% and split the remaining work among our staff of 5 people, they treated that idea like it was ground-breaking when in reality it's just common sense. We ended up not spending a dime on testing, accomplished the goals with the 75% automation, and completed it with only 2 of our people doing manual work.
If the Amazon warehouses become fully automated and a big group of people lose their jobs and there's no safety net, the warehouses will be burned to the ground within 24 hours.
It definitely won't of course, and I'm not saying it's right, but people who think that a mass people-to-automaton transition is going to go smoothly without a plan in place for the 'people' may lack perspective.
