"Rockstar reviewers" seem like a cure that's almost as bad as the disease. Some scientific fields already have a problem with groupthink, with a few well-defined and vigorously-opposed schools of thought. I would vastly prefer a broader reviewer pool to the usual suspects from the same few labs.

Everybody likes money, but I'm also not sure that's the way to go either. It would be great if reviewing directly impacted people's academic/research careers; I suspect the ability to review well is highly correlated with the ability to successfully run a research group. However, there are lots of thorny issues involving power and interpersonal relationships.