To fall back on an example that is not my own, imagine a conversation in 1900 between two horses:
Horse 1: I'm worried about these new automobile things, they might take our jobs
Horse 2: Don't worry, every technological development that displaced horse work in the past has led to more newer and better jobs for horses, there will always be work for us. Just look at trains, we don't have to pull big loads anymore, now we just draw carts and support riders, who knows what great jobs are just around the corner.
The world's horse population peaked in 1915, and is now down by 70% in the US.
Much of the new present day automation does not involve new machines. It often instead involves just interfacing a computer to an existing machine and writing software to let that computer replace the human operator of the machine. Those computers and interfaces are commodity items, largely built in automated factories.
This seems to me to be a fundamental difference.
Retail changed from a career that offered full time jobs to one making use of a just in time workforce, and has suffered tremendous consolidation over that period.Like entre chains and concept stores have declined; something like Media Play is dead and no longer able to be revived. Even Gamestop suffers, as the amount of physical releases declines.
I think retail shows it can be pretty dire indeed.
However I think the problem is we're at a point where things are churning so rapidly that many are being left behind. Your 50-year-old truck driver is not in a position to learn a new skill every 5-10 years to stay afloat, or uproot his life and move to a new city where they need workers.
Furthermore, there's conceivably a point when we won't need very many unskilled workers at all -- when we finally run out of things for average people to do -- and humans don't tend to treat other humans very well when they aren't needed. However, that's further off than the churn economy we have right now.