I don't see how they don't address the point being brought up. They explicitly point out past scenarios where these sort of predictions were wrong. I think the peddlers of these predictions are disingenuous and intellectually lazy. If they want others to take them seriously they need to explain why their predictions don't fall under the same criticisms.

I like the idea from the CP Grey video on automation: Better technology did not make more better jobs for horses.

Horses are still around, but the population peaked around the turn of the century (19th-20th).

How good does automation have to get before large portions of the population find themselves in the same position as early 20th century horses.

People can do mental labor. Horses can't. Automation can't.

Automation can absolutely do manual labor. How do you think cars used to be made in the past? By workers manually attaching the parts together. Today however we have giant machines that perform those tasks.

It's only a matter of time before other forms of manual labor are automated. It won't happen all at once, but gradually.

I said mental labour, not manual.

Yes, most physical labor may be automated... eventually, a lot of that is still fairly far off - we don't even have completely automated farms yet.

I'm more talking things like development, writing, design, marketing and sales. Things like that we'll probably always want some human touch to, even if it can be machine assisted in some ways.