If you’re trying to understand the world you will constantly generate weird, confusing or controversial ideas. If you’re trying to influence politics you will constantly trade off conceptual clarity for memetic catchiness. Optimizing for both will typically achieve neither.

This is not me calling the paper “propaganda” btw. Instead I’m describing two opposing motivations which I think the paper authors had. But figuring out a specific threshold that qualifies as “propaganda” seems less helpful than describing these dynamics mechanistically.