If you're so sure that new jobs will appear (and -- critical omission -- that they will be any good), surely you would be willing to ask the capital interests for whom these arguments are self-serving to put money where their mouth is and backstop a guarantee?
I suspect that the reason might be that the Industrial Revolution happened over 200 years ago. That provides a lot of time for 97% of jobs to progressively disappear without disrupting society too much (except for all the revolutions and world wars). That would be quite different than if AI caused any significant percentage of jobs to disappear in a much shorter period of time.
This analogy happens a lot, and it might be true, but it's not clear to me that they're comparable.
The Industrial Revolution mostly ate mechanical labor and created more 'thinking' and knowledge worker jobs closer to the top of the stack. AGI goes after the information / decision-making layer itself. And it's unclear how much remains once those are automated.
I consider the Industrial Revolution to still be ongoing, since jobs have constantly been automated away by technology for 250 years. Some like to split that time into separate eras. In that paradigm we're now in the Fifth Industrial Revolution (Industry 5.0).
Whatever you call it, jobs keep getting "stolen" by technology, and yet employment rates stay high and average living standard keeps rising.
I'm genuinely fascinated by how this keeps happening, decade after decade, and yet most people are convinced the opposite is happening. I'm old enough to remember this exact discussion from 50 years ago.
We all see and interact with jobs that did not exist 20 years ago, and many of us work those jobs. And yet... this knowledge is somehow compartmentalized away from future expectations.
If you want a theoretical framework for why this keeps happening, my thought is that unemployed humans are an unused resource. And capitalism is really good at finding ways to use those.
I have two ways to think of it, and both give similar numbers.
A: 250 years ago, 98% worked in farming. Today it's 2% (who produce more food!). Assume that the other 2% are at least twice as productive, and you get that 3% of the population now produces as much as 100% back then.
B: It's hard to directly estimate how much GDP per person has increased in 250 years. But the typical number economists get when trying is that it's 30x as big. Which means 3.3% of today's workforce produces as much (per person) as the whole workforce did back then.
Both A and B can be critiqued, but the precise numbers don't really matter for the argument.
Until you understand how something that counter intuitive happened, you should not speculate on how AI replacing current jobs will play out!