>The similarities are abstract: both use iteration to concentrate probability mass, both employ external oracles, both amplify marked states through operations on the full state space.

Pardon my reading comprehension, but I don't understand the analogy with Grover's beyond just reweighing probabilities