Well, a turkey predicting death would get shunned by other turkeys every time the farmer just comes and feeds them. And then Thankgiving comes and it turns out that turkey was right.

Failure of something to happen in the past is not really a good way to predict whether it happen in the future. The similarities between periods don't tell us anything at all without considering the differences.

Back then there was still factory jobs (the second wave of industrialization) and tons of services jobs (which made most of the population into office and service workers). Both of those are things that are now being automated away. So where exactly do you see those jobs being created?

We mechanised the farm removing the need for hundreds/thousands of people to work on farms. Those people went to work in factories and mines.

The mines have moved to larger machines with fewer workers, preferring to use large open-cut techniques rather than underground mining wherever possible.

Then we automated large numbers of factories, and there were some more transport jobs moving stuff around en-masse.

Now we're looking to automate the transport jobs with self-driving vehicles.

When we automate transport jobs, where do all those people go to work?

Businesses are not automating things because they think it's cool, it's because it's a lower cost alternative to having humans do the work.

But we have to reeducate people when their job gets rationalised away before they pensioners

Always being "a couple of times during stages of the industrial revolution" -- back when we had crude automation and no AI.

Hardly a long range of samples to make any safe deductions from.