Hanson's argument: nearly all of the survival probability along the chain from non-life to interstellar civilisation must collapse somewhere. Set how hard each step is; the tool shows what fraction of the filter must lie behind us versus ahead — and why finding past microbial life elsewhere in the solar system would be bad news.
Robin Hanson's 1998 essay framed the Fermi observation as follows: if you start with ~10¹⁰–10¹¹ habitable planets in the galaxy and you observe zero interstellar civilisations, the product of per-step survival probabilities along the chain from dead chemistry to galactic civilisation must be at most ~10⁻¹⁰. Somewhere on the chain, the probabilities collapse. That collapse — concentrated in one or two steps or spread across many — is the Great Filter.
If the filter is behind us, it means at least one of the early steps (abiogenesis, oxygenation, eukaryogenesis, multicellularity) was implausibly hard, and our universe is mostly dead because life almost never starts. Future humanity is probably fine; we got lucky. If the filter is ahead, it means the early steps are easy and the late steps — industrial-civ sustainability, AI alignment, weapons restraint, climate survival, interstellar engineering — are the bottleneck. Future humanity is probably doomed; the silence is what awaits us too.
Finding a second instance of life — even microbial — in the solar system would constrain the early-filter hypothesis. Two independent abiogenesis events in one tiny stellar neighbourhood implies abiogenesis is not the bottleneck, which means the bottleneck has to be somewhere else, which (since we're here at the civilisation step) means it has to be ahead. Hence Bostrom's title: "Why I hope the search for extraterrestrial life finds nothing." This tool models the effect by letting you slide the abiogenesis step toward 1 and watching the implied "filter ahead" probability spike.
For each of 8 steps you supply P(survive). The product gives the total survival probability from dead chemistry to interstellar civilisation. Given a habitable-planet count H (default 10¹⁰), the implied number of interstellar civilisations is N = H · ∏ Pi. "Filter passed" = product over the steps before "you are here"; "filter ahead" = product over the remaining steps. These two numbers tell you which side of you the missing log-decades sit on.
One way to escape the "filter ahead" prediction is the Macro Transcension Hypothesis: civilisations don't disappear, they compress into compact substrates (e.g. the ergosphere of an IMBH) that don't emit observable signatures. In Great Filter terms, this redefines the "interstellar" step — civilisations achieve something more interesting than colonising volumes, and so the absence of visible expansion is not absence of civilisations. See BZ/Kardashev calculator and Bekenstein-Landauer-Lloyd explorer for the engineering side, and the Aestivation calculator for an adjacent hypothesis (civilisations wait for the CMB to cool rather than compressing into ergospheres).
Robin Hanson 1998 ("The Great Filter — Are We Almost Past It?") observed that the Fermi paradox implies at least one extreme-difficulty step ("filter") between abiogenesis and observable galactic-scale civilisation. The filter could be behind us (rare-Earth-style: life or intelligence is staggeringly unlikely) or ahead of us (existential risk: technological civilisations destroy themselves). The strategic implication is asymmetric — discovering simple microbial life on Mars would be terrifying, because it would suggest the filter must be at a later stage and therefore likely ahead of us.
Standard candidate filters (with rough estimates of step difficulty from Carter, Hanson, others): abiogenesis itself (~10⁻¹⁰ per Earth-like planet, but with huge uncertainty); prokaryote → eukaryote transition (took ~2 Gyr on Earth, may be a hard step); multicellularity (independent evolution multiple times on Earth, so probably easy); animal-grade nervous systems; tool-using intelligence; agriculture; industrial revolution; nuclear weapons; AI alignment; biosphere collapse. Bostrom 2008 ("Where Are They?", Technology Review) made the "Mars-discovery would be bad news" argument popular.
The 2010s–2020s exoplanet revolution (Kepler, TESS, JWST atmospheric spectroscopy) has substantially shifted Bayesian priors. f_p ≈ 1 and n_e ≈ 0.1–0.5 mean that "Earth-like" environments are abundant; the filter must therefore be biological, behavioural, or technological rather than planetary. JWST searches for biosignatures (O₃, CH₄ disequilibria) in habitable-zone exoplanet atmospheres through the 2020s and 2030s will provide the first empirical constraints on f_l (life origination rate). LISA and future GW observatories may detect technosignatures from megastructures or unusual orbital configurations.