Again, the argument is not to do nothing. The argument is use masks where they help, not enforce usage where they don't, and further study the in betweens.
The point of my comment wasn't to argue for policy however. It was to point out the inaccuracy of the statement that "we know masks work, therefore making masks mandatory is objectively good (for public health)". There are several examples where this reasoning have turned out to be not only empirically false, but actively detrimental. Bike helmets and infant mortality come to mind.
The comparison with bike helmets is apt. They are great, it is easy to understand why they work, but still enforced usage damages public health. One theory is that lesser people bike, and those that do take more risks. The facts are that areas with the least cycling risks have the least helmet usage.
Enforced masks everywhere all the time has the obvious risks of more people going outside even with slight symptoms, and taking more risks by not avoided crowded places.
These arguments are not mine, look at the meta studies WHO made. I simply parrot what little articles I can find, and compare with other similar public policy, including bike helmets.
The relevant question to ask before making masks mandatory by law is not whether they help reduce virus spread some of the time, or even most of the time, but rather if there instances where the enforcement actually increase spread. Judging from published literature, that is far from unlikely.