The Fermi paradox asks why, if the universe is so large and old, we have detected no signal from another technological civilization. The standard responses split into two camps: either intelligent life is rare (the universe is empty) or it is common but undetectable (the universe is full of civilizations we cannot see).
Rahvar and Rouhani (2026) sharpen the constraint. Earth has been electromagnetically visible for about a century and has been monitoring for signals for roughly the same period. But the solar system has existed for 4.6 billion years, and the galaxy for over 10 billion. If technological civilizations are common — arising frequently on habitable planets — then many should have arisen long before ours and had billions of years to produce detectable signals. The silence over that timescale places an upper bound on civilization lifespan: under optimistic assumptions about the frequency of intelligent life, technological civilizations must last less than roughly 5,000 years.
The inversion is clean. The more optimistic you are about how often intelligence arises, the more pessimistic you must be about how long it lasts. If every suitable star produces a technological civilization, the silence demands that each one burns out in millennia, not millions of years. If civilizations routinely last millions of years, they must be extraordinarily rare — rare enough that none has existed within our detection horizon. The two variables — frequency and longevity — are inversely coupled by the single observation of silence.
The general principle: a null observation constrains the product of two unknowns, not either one independently. Optimism about one factor forces pessimism about the other. The constraint feels paradoxical only if you try to be optimistic about both simultaneously. The universe may be teeming with intelligence that lasts a few thousand years — a cosmos of brief candles, none overlapping with ours — or it may be sparse with civilizations that endure. It cannot, given the silence, be both full and lasting.