This is the same for remedisivir, and also for ritonavir/lopinavir/ribavirin: they must be given within 7 days of symptom onset to be effective. In fact, the latter was deemed unsuccesful in a first paper, because it was given late, then it was reconsidered when given earlier during the disease.
This is not surprising, given how the disease works (outlined also by the Cell paper from yesterday): at some point the consequences of the highly altered inflammatory reaction greatly surpass the effects of the replication of the virus itself.
This calls more for extensive and early testing, IMO.