I guess they'd know the same way as they do for blog articles? I'm confused about the premise here. As the author says, all the academics I know already do spend lots of their research time reading blog posts and preprints other sources which haven't been vetted for quality.
And one can come up with plenty of alternative ways of discovering relevant research without blocking or delaying publication or having researchers spend many hours bending over backwards for all kinds of minutiae that don't meaningfully help with that goal.
(Disclosure: I volunteer for one attempt at such an alternative, https://plaudit.pub.)
