Stagnation, or progress being slowed down by Luddites or people in positions of power being overly cautious and pussyfooting around, is what terrifies me.
This factor is exactly why all of Cloned's predictions for the impact of 5G are going to come through way later than 2030.
It doesn't matter if we can automate all driving, for example. Because millions of people would be very upset at having their car taken away and being banned from driving.
And if your automated cars need to worry about humans driving randomly and unpredictably around the place, then we have to delay them by half a decade to test them before anyone is legally allowed to lift their hands off the wheel.
Even if that's all sorted, the average age of cars (uk) is about 7-8 years, so there would be at least a decade getting everyone switched over. Assuming you could persuade everyone to want to.
Same with autonomous education and healthcare:
You need to 100% automate very very important processes, mess up healthcare and people die, mess up education and you destroy the future of your nation/society, along with the lives and prospects of potentially millions of children.
(there's a reason schools are still using the whole "sit 30 children in a room and lecture them for hours" methods in spite of all the better ways we have had to teach people since the invention of TV)
You also need to do it with little/reluctant assistance from the majority of existing teachers or doctors/nurses (since they won't want to put themselves out of a job)
You need to consider more factors than "Would this be technically feasible", which is why Ray Kurzweil has missed a lot of predictions.
Just because we could doesn't mean we will.