Do you really believe the success (at this time in technology) of general-purpose AI has a strong correlation to the potential for human labor to be replaced by machine labor?
It would seem with the plethora of startups being created every day at providing services and products in different markets by utilizing technology means that there will be plenty of different companies willing to focus on a particular market / product / service to develop their AI. Does the success of general-purpose AI really limit the possibility of developing specialized machine labor AI? I can see how it will be slower to reach the point of replacing all of these jobs than if general-purpose AI worked well, but I don't think it really limits the possibility.
So long as humans can do general purpose thinking and machines can't, any task that requires general purpose thinking will require humans.
Many of these tasks may be things that go undone now, but will become affordable when machine-made stuff gets cheaper and removes expenses from people's budgets.
Assuming the economics doesn't mess everything up. Which it easily could. But the fundamental problem is self-solving until we run out of applications of general purpose thinking.
