Yeah man, it is 300% more effective to pick an area, read the literature, and immediately start writing about it. I am sympathetic to in-principle visions people have about "depth" and all that, but the practical reality is that none of that matters anymore (again, for research jobs)