Point is that the study is done at the wrong stage, too late.
There's also the fact that study makes it less likely that pre-hospitalization HCQ will have any effect either. Applying a Bayesian analysis here, it's now likelier that the true ground truth is simply that HCQ doesn't do anything to COVID-19 at all, not necessarily that it works only in a narrow early window. HCQ has severe side effects too, that require close monitoring and can be fatal. If it's not clearly doing good then it definitely shouldn't be being given to COVID-19 patients.
Personally, my hopes are more on remdesivir and favipiravir, both of which are actual antivirals (HCQ is not) and thus were always more promising. HCQ was never the most promising drug candidate, and the only reason we're even talking about it as much as we are is that, for whatever random reason, that's the one Trump heard about first and then fixated on.
I'd agree that we really don't have any evidence that giving it to hospitalized patients helps at all. Most antivirals have to be taken at the early stages of an infection.
The best looking treatment for people seriously ill with Covid-19 right now seems to be IL-6 inhibitors. That won't stop you from getting sick but might let you avoid intubation.
We'll see with the U of Minnesota trial, which attempts to use it as post-exposure prophylaxis. Data (hopefully) will tell us.
On favipiravir, the data on the Chinese study wasn't exceptional, but wasn't that bad either. Perhaps the Japanese phase 3 study will be more enlightening.