My proposal is not designed to diminish the rewards for content that voters are truly interested in up voting. My proposal is merely to untie votes from a one-size-fits-all preference, so that different voters can be grouped (clustered) automatically (algorithmically) by their relative like-mindedness, as measured by the similarity of their voting patterns over time.

In short, I am proposing an improvement which will be computed automatically behind the the scenes and which will give us more relevant rankings so that we all hopefully see the sort of content we are most interested in.

Soon there will be so much content that we otherwise will not be able to find the content we are interested in.