Can we automate our own work — or show that it is hard?

Computer scientists seek to understand what can be automated, but what do we know about automating our own work? Can we outsource our own research questions to computers? In this talk I will discuss this question from the perspective of the theory of distributed computing. I will present not only recent examples of human-computer-collaborations that have resulted in major breakthroughs in our understanding of distributed computing, but I will also explore the fundamental limits of such approaches.