Well, if you believe tangential dicta that was grounded in no preexisting law offered in a since-overturned case allowing the repression of core political speech, sure...
But best not use that example.
https://en.wikipedia.org/wiki/Shouting_fire_in_a_crowded_the...
If you're feeling pedantic, perhaps you can replace the phrase with "yelling things to incite an imminent lawless action" (to crib from Wikipedia's summarization of Brandenburg) any time anyone ever says it. Does that work for you? Because the underlying point remains: that there are limits to free speech.
It is a misquote (leaving out "falsely", a modifier which is key to the meaning of the original quote) of a statement the original of which is inaccurate as a description of the prior state of the law when the case it was in was decided, or the state of the law once the case was decided (being dicta, it itself had no binding effect), or the current state of the law, from a decision now widely recognized as anathema to the central protection of the First Amendment.
It is, almost literally, the worst example that you could use.
But it's the one everybody knows. Society has chosen it as the common term for the concept by now. Its original source or accuracy is entirely irrelevant -- in the same way it would be missing the point to complain that "black holes" aren't technically black.
After all, we're not having a discussion between lawyers about the intracacies of US free speech law. It's just a phrase for referring to the general concept of rights not being absolute.
>spreading blatant viral lies on social media
You even admit this is arguable. You're assuming people are morons who need big brother to help them out - you don't really believe in democratic principles if this is your stance. The enforcement of punishing such lies is inconsistent as well. One needs only to look at the blatant left wing bias of Twitter and Facebook, where conservatives are banned for expressing opinions whereas lies told by media outlets (with small corrections added days later) are left up and not punished at all.
And the idea that the "general consensus" is against all forms of censorship is quite false. In America it is, but in Europe and other countries censorship of hate speech (e.g. racism, Nazism, etc.) is quite accepted as part of the general consensus.
And even in the US, "yelling fire in a crowded movie theater" isn't protected either. And arguably, spreading blatant viral lies on social media close to an election is akin to yelling fire in a crowded movie theater, since the national consequences could be so dire.
There are many intelligent arguments to be made that censorship of speech that is either a) primarily hate-directed rather than information-directed, or b) outrageously false but capable of swinging an election, could be outlawed, and neither of these would be incompatible with modern-day political liberalism, which is more commonly called "social democracy" to distinguish it from the classical liberalism that Mill did so much to defend.
And the idea that this would somehow depend on an "oracle of truth" is nonsense. Courts judge things like libel and defamation cases all the time. Sure, there are gray cases that could go either way, but drawing lines in gray areas is what courts have done ever since they existed in the first place. Holding Facebook moderators ultimately responsible to judges, for example, isn't inherently difficult to do if we wanted to.