This sounds like a pretty good principle. In the long-term, actual engagement with worthwhile content is significantly more valuable than a blind upvote for unseen content. There is a big information asymmetry problem from bot-voting for e.g. author lists (especially when it involves a federated publishing model as described by @nonlinearone). To me, it sounds like the short term hit in profits is actually the rational long term play. I would actually be in favour of vote power degrading at a much faster rate, plus a system of delegated curation to deal with whale-sized positions, to encourage engagement over mass curation to avoid this problem.