I have no problem with running trials trying to answer the question. IMO, one of the (many) big failures of our government response to COVID has been the lack of central coordination of well designed clinical trials. Even though we have had way more cases than anywhere else, the lack of data is pretty shocking. A place like the UK has done a much better job of putting a huge percentage of their patients on study, which is why we have gotten the answers we have on Dex (it works!) and on HCQ (it doesn't, at least in hospitalized patients). The problem with just prescribing random drugs before the data is available is not just the potential harm it could do to patients, and the wasted money, but it makes it harder to do the research needed to know whether you are actually helping or not.
That said, there are a finite numbers of patients and dollars that are available for clinical research, even if things were better organized. So far, a hugely disproportionate amount of those dollars and patients have gone into HCQ studies. Given that the available evidence suggests it is highly unlikely to be helpful, I think it makes more sense to divert those patients and dollars to testing other options with a higher pretest probability of success.