I've been asked to review papers, which I hate, because I feel like I'm writing a lot of BS when I could just be saying which papers are better than others.
But I feel bad declining reviews for that reason, because "it's the best system we've got".
So, I want to build a better system: Research Tomatoes
The idea so far is a monthly journal of new research where everyone who submits needs to rank-order at least 3 other submissions, say. And the journal just shows all the papers in rank order at the end of the month.
I think it would also be best to make it non-archival, where Research Tomatoes doesn't own any publication rights, so that people can submit there without ruining their chances of submitting to some top-tier conference (until Research Tomatoes becomes a top-tier conference).
And then when people ask me to review papers, I can say "I will review this paper if you submit something to Research Tomatoes, or at least rank order 3 papers on that site".