Can this not be at least partially solved by relying more heavily on a user's feedback(like/ dislike buttons, numerical rating systems, and any other feedback mechanisms) to better understand preferences of the viewer? Along with simply being more patient and diligent when trying to discern actual user desire?

It seems like the problem of asymmetry in recommendation platforms is more a choice from a design perspective than a challenge yet to be figured out.

I'm far from an expert tho so just a thought from a laymen