Three researchers built a thousand-year simulation of Earth's own future and ran it two thousand times. Celia Blanco, Jacob Haqq-Misra, and George Profitiliotis modeled ten distinct trajectories a technological civilization might follow, each one run 200 separate times to account for randomness, and published the results in April 2026.
Most of those simulated futures did not end in extinction. They ended somewhere stranger: civilizations that rose, lost most of their technological capacity, and rebuilt, sometimes more than once, over the course of a thousand years.
The researchers call the fraction of that thousand years a civilization spends technologically active its duty cycle. Across the ten scenarios, duty cycles ranged from about 38 percent to 100 percent. Some trajectories, given names like Golden Age and Eden, never collapsed at all. Others, like Big Brother and Sword of Damocles, collapsed quickly and stayed down. A few, including a scenario the researchers called Restoration, collapsed and rebuilt more than once within the thousand-year window.
Two variables mattered more than any of the modeled existential hazards themselves.
What actually determines whether a civilization collapses?
The paper's sensitivity analysis found that resource depletion rate and the fraction of technological capacity preserved after a collapse were the two most consistent drivers of a civilization's long-term trajectory, more consistent than any single catastrophic risk built into the model. A civilization that depletes its resources slowly, and that retains more of what it built even after a collapse, tends toward the high end of the duty cycle range almost regardless of what specific disaster hits it. One that burns through resources quickly and loses nearly everything in a collapse tends toward the low end, no matter how the collapse was triggered.
Governance structure shaped the picture too, sitting alongside resource pressure and hazard exposure as one of the three variables the model used to steer each trajectory. The paper does not treat any single form of governance as the fix. What it treats as decisive is whether a civilization's institutions can preserve enough coordination and technical knowledge to keep a collapse from erasing what came before it.
The researchers frame that as an actionable result rather than a fatalistic one. If the model is roughly right, then reducing resource consumption and preserving institutional and technical capacity through a crisis may matter as much for Earth's own long-term resilience as preventing any single catastrophic event outright. The duty cycle is something a civilization can influence in the middle of its own story, not just something that gets determined once at the moment of collapse.
Named scenarios such as Living with the Land, Transhumanism, Deus ex Machina, and Out of Eden left behind almost no detectable chlorofluorocarbon signature at all, regardless of their duty cycle. A civilization can go dark to an outside observer for reasons that have nothing to do with collapse, simply by never producing the kind of industrial byproduct anyone is currently equipped to look for.
Why would this explain the silence of the universe?
That framing matters beyond Earth too. The paper's second half turns the same model toward a much older question: why the galaxy appears so quiet, despite being old enough and large enough that someone should probably have made themselves known by now.
A civilization is really only detectable while its duty cycle is active. The researchers modeled which byproducts of industrial activity, including nitrogen dioxide and specific chlorofluorocarbons, would remain measurable in a planet's atmosphere during and after each scenario. Some compounds persist for decades after production stops. Others, like carbon tetrafluoride, can linger for tens of thousands of years, showing up in only the most catastrophic, high-emission trajectories the model produced. A few of the gentler, more sustainable scenarios left almost no persistent chemical trace at all, meaning a civilization could handle its own transition carefully enough to become nearly invisible to exactly the kind of search we currently run.
For two civilizations to notice each other, their active windows have to overlap. The probability of that overlap is roughly the product of their two duty cycles. A civilization active 100 percent of the time and one active 38 percent of the time might still miss each other, if the timing of their respective windows simply doesn't line up across a galaxy old enough for that math to matter. Search strategies built around continuous, decades-long observation may be tuned for exactly the wrong kind of civilization, the rare one that never goes dark, while missing the more common pattern this model describes.
Does this change what the silence actually means?
The paper's authors put it plainly: the silence may not mean intelligence is rare. It may mean most technological civilizations spend a meaningful share of their history dark, recovering from themselves, and observers everywhere, including us, are only capable of noticing each other during a narrow, overlapping sliver of that much longer story.
We have argued elsewhere that the sheer distance between stars may be answer enough on its own, without needing any civilization to be hiding or dead. This model doesn't compete with that explanation so much as sit alongside it. Distance limits who could ever reach anyone. Duty cycle limits who could ever notice anyone, even next door, even if reaching them were somehow no obstacle at all.
Neither explanation requires anyone to be gone. Both are compatible with a galaxy that is, in fact, full of intermittent lights, most of them simply not on at the same moment we happened to look up, including, perhaps, our own.