Over the last 50 years, the Supreme Court's First Amendment cases have treated content-based speech restrictions very differently from content-neutral ones. Even a relatively modest content-based restriction, generally speaking, must either fit within a First Amendment exception (such as for true threats or for defamation) or must pass the very demanding "strict scrutiny" test. But a relatively modest content-neutral restriction—such as a restriction on sound amplification, or a limit on the number of people who can picket in a particular place—is generally constitutional if it passes the considerably less demanding "intermediate scrutiny" test. (Harsher restrictions, which fail to leave open "ample alternative channels" for speech, are generally harder to justify, even if they are content-neutral.)
This of course means that it's crucial to define what's content-based and what's content-neutral. That's often clear: Restrictions that turn on the viewpoint of speech (e.g., even modest restrictions on racist speech or anti-government speech or some such) are certainly content-based. So are subject matter restrictions, for instance ones that restrict picketing but exempt labor union picketing, or ones that treat political signs different from other signs. So are restrictions on saying particular words (e.g., vulgarities), or restrictions that turn on whether the speaker conveyed certain facts (e.g., the names of rape victims, or the name of the author of a leaflet). But sometimes the matter is less clear.
In two cases in the mid-2010s, the Court set forth two rules that seemed to define content discrimination quite broadly: Under those rules (which also had ample precedent in past cases), a restriction is content-based if it
