This is an insightful article into the process. It is possible to test ahead of making any coding changes, that would involve convincing any high-stake friends to stop upvoting and only downvote bad posts, to see how that affects the Steemit post payouts and economy.

I like it since it doesn't take anything away from high-stake users, or blame them for the economy, and it proposes a fix that can be tested to a reasonable extent without any coding changes or payout changes.