Predicting discoveries—enabling research or killing novelty?