Your second point is like saying: 'I imagine a gig economy where taxi drivers will need to leverage a car to be effective.' If someone offers you a hammer to build a cabin, why would you say no?
To your point about it being good for small and medium sized businesses — let’s assume it gets rid of a significant amount of workers. Who would be the consumers of these products, other small/medium sized businesses who also mostly have AI producing work? Obviously it’s a simplified question, but I’m curious how we’ll handle the removal of “useless workforce” when AI does an objectively better job.
The previous generations could never imagine the kinds of jobs we have today.
Now I am not saying that this is a guarantee to happen again, and that new jobs will be created and such. But I think anyone claiming to know that "this time it will be different" and that "we wont create new jobs that we can't imagine today" could be just as wrong about that idea too.
Gutting a ton of jobs across many sectors in the economy won't change demand for food or rent, but will likely hurt every service sector and slow the economy down overall as people without jobs push out purchases they would have made otherwise.
As maybe a last resort, regulation could play a factor in how many AI resources a company can use, which will mimic scarcity.
You act as if humans haven’t been solving existential problems for their whole existence. It’s not like 50% of the workforce will get laid off and there will be no response to that.