More on the Robopocalypse

One more thing: there is some chance that in the long run human capital will adapt to the Robopocalypse and worker skills will complement the capabilities of our future new robot overlords. LinkedIn founder and executive Reid Hoffman sums it up well (trying a little too hard to be cool) in this interview with The Daily Show:

Trevor Noah: If robots are doing the jobs...what do humans do? What is our purpose?
RH: ...That's a possible universe, but there are many other possible universes...So it's not that the robots do all the teaching, but it enables [sic] people to be much more productive. 

Finally, we probably don't need to worry about robots killing us and ending human existence as we know it, because...

TN: Will Artificial Intelligence kill us all?
RH: Very unlikely...Asteroids may kill us all; nuclear weapons may kill us all...
...
TN: Why would the robot not want to make me not exist?
RN: ...Then you get to the whole question of the robot "wanting."... When do robots become fully wanting, sentient people, and what language of "want" describes them?

In other words, robot preferences will probably not be the same as human preferences, and will probably be less spiteful, hateful, and vengeful. Maybe a bit more like... homo economicus? At least our robot overlords might be programmed to have preferences that are complete, montonic, and transitive (even when we aggregate them!).