This. The result is utterly trivial since they are skipping coherence: It remains to be seen if we can extend the linear (or polynomial) efficiency to coherent

This. The result is utterly trivial since they are skipping coherence: It remains to be seen if we can extend the linear (or polynomial) efficiency to coherent control gates. The efficiency of our implementation of coherent control-U gates requires further study. Our results so far show that in the worst case the number of edges grows as 2n where n is the number of control qubits. Yup. Nothing to see here.