The switch from downvoting to flagging in the UI was a mistake. It contradicts the notion of a platform where stakeholders decide on the fate of content as a consensus between people who may agree or disagree and turns it into one where a UI designer has imposed a non-neutral value system on what it means to downvote and when that should be used.

The reputation system has its own issues, but at least all the inputs to reputation are those that come from stakeholders (over time), not a UI designer. Even if, as is the case, the rules for aggregating this input are are somewhat arbitrary and broken.

Finally, hiding content based on voting and/or rep should be configurable by the end user. I may want to filter out all content that doesn't get enough upvotes (for example at least 1 MV worth). You may want to filter out content from anyone with less than a 70 rep. Someone else might want to see the whole fire hose. And finally someone may wish to see only that which has been approved by a specified set of voters (perhaps this could be used for "parental controls" with a trusted child-friendly approval service). There is no harm in providing this choice (other than the costs of implementing it of course).