I agree with much else said here about the failure -> crank -> taboo problem, but it bears noting that the Fermi paradox is not really a paradox. There is this https://arxiv.org/abs/1806.02404 . The 2 minute layman's version is that multiplying several to many very uncertain numbers "drives your error bars to the fringes". So, there is not some credible, high probability estimate that we are not (EDIT: locally) alone. Without that estimate there is no paradox.