Max value entropy search recasts Bayesian optimization as reducing entropy over the best possible output value rather than reducing entropy over the location of the minimizer. This results in having to compute mutual information over the output distribution rather than over the much higher dimensional input space. This paper extends max value entropy search to the multi-objective setting, where now the authors guide the search through reducing entropy over the distribution over the value at the optimal Pareto frontier. Overall, the reviewers found the paper very well written, the approach sensible and the experiments convincing. The major source of criticism was that the reviewers found the contribution to be incremental over the original Max-value entropy search without making that clear in the paper. However, they did enjoy reading the paper and would recommend it to others interested in multi-objective Bayesian optimization. The approach is indeed novel and given the strong empirical results and a reviewer willing to champion the paper, I recommend acceptance. Please take both R2 and R3's comments into account in preparing the camera ready version, particularly with respect to discussion of existing work.