Dacyn comments on Computational efficiency reasons not to model VNM-rational preference relations with utility functions