As has been made clear by subsequent comments, projektfu was just repeating words they had heard previously, without having any idea what they referred to, as Eliza or a scarlet macaw might do. It's not useful to project an intent onto their words to refer to anything in particular. Presumably projektfu is, unlike Eliza, a human with mental processes who is capable of forming intents, including intents to emit utterances that refer to objective or consensus reality; but such an enterprise requires efforts that were not made in this case, such as reading the six lines of text at the top of
https://en.wikipedia.org/wiki/Fire_in_a_crowded_theater and basing their comments on them†.
Moreover, I think this is true of most rhetorical devices in general, and that is part of what the podcast is criticizing. They are not attempts to collaboratively consider the evidence in order to arrive at shared conclusions informed by the knowledge of all parties; they are attempts to influence whatever conclusions may arise, by hook or by crook, to benefit the interests of the utterer, and whether either the utterances or the conclusions happen to be objectively true or false is none of the utterer's concern.
I consider this contemptible, and I try to engage in it as little as I can manage.
† Incidentally, this page explains that Schenck was not specifically overturned until Brandenburg in 1969 (although the Brandenburg decision doesn't mention Schenck), but the Supreme Court's progressively broader reading of the First Amendment was already protecting antiwar and anticapitalist activism by the 1950s, for example in Gibson (1946), Estep (1946), Dickinson (1953), and Yates (1957).