The Fermi Paradox is Wrong: Why We Don't Hear Aliens Because They're Post-Biological

Every time someone points a radio telescope at a star and hears nothing, there is a quiet assumption riding along: that whoever is out there is still the kind of creature that would want to shout. A warm-blooded thing, roughly our size, living on a planet, building transmitters for the same reasons we did. I think that assumption is almost certainly wrong, and that a lot of the gloom around the Fermi paradox comes from it.

I should be careful about what I’m claiming, though, because the title writes a cheque the physics only partly cashes. My argument is that the radio version of the paradox is malformed, and that post-biological intelligence explains the silence on those frequencies well. What it does not do, on its own, is explain the deeper version of Fermi’s question. That part is harder, and the most interesting thing about the post-biological idea is how precisely it tells you where it should fail.

§Fermi was asking about visitors, not broadcasts

The original question was a lunch remark at Los Alamos in the summer of 1950. Fermi, Teller, Konopinski and York had been joking about a New Yorker cartoon and the recent wave of flying saucer reports, and Fermi asked, more or less, where everybody was.^[1]1. Fermi died in 1954 without writing his version down. The account comes from his lunch companions, collected decades later by Eric Jones: “Where is everybody.” An account of Fermi’s question, Los Alamos report LA-10311-MS (1985). A readable retelling, including York’s recollection that Fermi thought we should have been visited long ago, is Paul Patton’s Beyond Fermi’s Paradox I at Universe Today. The point, as York remembered it, was that if civilisations were common and interstellar travel possible, Earth ought to have been visited, repeatedly, a long time ago.

Notice what’s missing from that. Nothing about radio. The question is about expansion and footprints: given billions of years and billions of stars, why hasn’t anyone spread far enough to reach us? The radio framing came later, with SETI, and it quietly narrowed the question to something much weaker. “Why haven’t we picked up a narrowband signal from the few thousand stars we’ve looked at closely?” is a question with a hundred boring answers. “Why is the galaxy not visibly used?” is the one that bites.

Holding those two apart matters, because the post-biological argument demolishes the first and makes the second sharper.

§The radio era is a rounding error

Here is the arithmetic that convinced me. Earth is about 4.5 billion years old. Humans have been able to transmit radio for a bit over a century. And we are already, within that same century, building machines that do a serious fraction of our thinking. Whatever happens next, the phase of our history in which we are biological, technological and broadcasting looks short. Call it a few centuries to be generous. Against 4.5 billion years, that’s less than one part in ten million of the planet’s history.

Now apply it to everyone else. The philosopher Susan Schneider, in a paper written for a NASA and Library of Congress volume on astrobiology, makes the point that we are young: many planets are billions of years older than ours, so any civilisation we meet is likely to be far older than us. She calls us “galactic babies”, and argues that the most advanced civilisations “will likely be post biological.”^[2]2. Susan Schneider, interviewed about the paper she wrote for a NASA and Library of Congress volume on astrobiology: Philosopher Ponders ‘Alien Minds’ for NASA, UConn Today (2015). If even a fraction of those older civilisations did what we seem to be in the middle of doing, then the odds of catching one during its brief radio-using, flesh-based window are terrible. You’re looking for a species in the one century when it happened to resemble you.

As an engineer, the framing I keep coming back to is the bootloader. Biology is a fantastic way to get from chemistry to the first general-purpose computer, and a terrible platform to run anything on afterwards. It needs a narrow temperature band, liquid water, a planet’s worth of shielding, and it thinks with nerve signals slow enough that a copper wire would find them embarrassing. Nobody ships the bootloader as the product. I don’t see why evolution’s output would be the exception, once something on the planet can design its own substrate.

And a mind running on engineered hardware has much less reason to broadcast. Our radio leakage is mostly an accident of a particular technology stack: high-power, omnidirectional transmitters for TV and radar. Plenty of that already looks like a phase, with more and more traffic moving over cable, fibre and tight beams. Nobody designs a network to spray its traffic into space if they can help it.

§Thinking costs heat, and heat is visible

This is where the argument either becomes physics or stays science fiction, so let’s do the physics. Computation is not free. Landauer’s principle puts a floor on the energy needed to erase one bit of information at temperature $T$:

At room temperature, 300 K, that’s about 2.9 × 10⁻²¹ joules per bit. Small. But a post-biological civilisation is, almost by definition, one whose main activity is computation, and anything that wants more of it eventually wants more energy. The biggest energy source in any star system is the star. The Sun puts out 3.8 × 10²⁶ watts. If you captured all of it and computed at the Landauer limit at room temperature, you could erase roughly 10⁴⁷ bits per second.^[3]3. Solar luminosity of 3.828 × 10²⁶ W from NASA’s Sun Fact Sheet. The rest is Landauer’s bound, the Stefan–Boltzmann law and Wien’s law, done on the back of an envelope. Real computers sit many orders of magnitude above the Landauer floor, which only makes the heat problem worse. That’s the size of the prize. It’s the reason Freeman Dyson’s old idea of a shell or swarm of collectors around a star refuses to die.

But energy in means heat out. Conservation doesn’t care how clever your architecture is. A swarm that absorbs the Sun’s whole output has to re-radiate the same 3.8 × 10²⁶ watts. Using the Stefan–Boltzmann law, a shell radiating that much at a comfortable 300 K would need a radius of about 1.7 times the Earth–Sun distance. And Wien’s law says a 300 K surface glows brightest at around 10 micrometres, deep in the mid-infrared.

So the post-biological hypothesis makes a prediction, and it is not “silence”. It’s the opposite: a star that looks oddly dim in visible light and oddly bright in the infrared, glowing like a radiator. A civilisation can stop talking. It cannot stop obeying the second law. You can turn off a radio transmitter. You cannot turn off your heat sink.

§We went looking for the waste heat

This is the part of the story I wish more people knew, because the searches have actually been done, and they’re much stronger than listening for beeps.

The Ĝ (G-HAT) project at Penn State used the WISE infrared survey to look at about 100,000 galaxies for the mid-infrared glow you’d expect if someone were reprocessing a galaxy’s starlight on a galactic scale. None of them hosts a civilisation turning more than 85 percent of its starlight into mid-infrared waste heat; only about fifty are even bright enough in the infrared to allow more than half.^[4]4. Griffith, Wright et al., The Ĝ Infrared Search for Extraterrestrial Civilizations with Large Energy Supplies. III, ApJS 217, 25 (2015). The framing of why waste heat is the thing to look for is in paper I: Wright, Mullan, Sigurðsson & Povich (2014). Arp 220, which astronomers already know as a dusty collision between two galaxies, is one of those fifty. That is a strong result. It rules out entire galaxies that have been fully put to work, across a sample a thousand times bigger than anything searched before.

At the scale of single stars, Project Hephaistos sifted about five million stars in Gaia, 2MASS and WISE data for partial Dyson spheres, and in 2024 reported seven candidates, all small red M dwarfs with an infrared excess that ordinary explanations didn’t easily cover.^[5]5. Suazo, Zackrisson et al., Project Hephaistos II: Dyson sphere candidates from Gaia DR3, 2MASS, and WISE, MNRAS (2024). The first rebuttal is Ren, Garrett & Siemion, Background Contamination of the Project Hephaistos Dyson Sphere Candidates (2024); the 2026 follow-up, with the Hephaistos authors on it, is Ren et al., Archival Diagnostics for Potential Background Contaminants of Project Hephaistos Dyson Sphere Candidates. Within weeks, Ren, Garrett and Siemion pointed out that three of them sat close to radio sources, and argued that hot, dust-obscured galaxies far in the background could be contaminating the infrared light. A larger 2026 study, which includes several of the original Hephaistos authors, found clear background contamination in two candidates and suggestive signs in others.

I find this genuinely encouraging, and not because it found aliens. It’s that the process worked: a falsifiable technosignature, a pipeline, candidates, a rebuttal, and follow-up with both sides on the paper. Tabby’s star went the same way. Its strange dips in brightness set off a wave of megastructure speculation, and by 2018 the observations showed the dimming was stronger in blue light than red, which is what fine dust does and a solid structure doesn’t.^[6]6. Boyajian et al., The First Post-Kepler Brightness Dips of KIC 8462852, ApJL (2018). The dips were deeper in blue than in red, which is what fine dust does and what a solid panel does not.

Put it together and the honest summary is uncomfortable for my side: the infrared sky looks pretty quiet too. Post-biology explains the radio silence. It doesn’t, by itself, explain the infrared silence. If anything it turns the infrared silence into a sharper puzzle, because computing minds should want energy more than biological ones do.

§The quiet option, and why I only half buy it

There is a clever escape, and it’s the reason I think this line of argument deserves more attention than it gets. In 2017 Anders Sandberg, Stuart Armstrong and Milan Ćirković proposed the aestivation hypothesis.^[7]7. Sandberg, Armstrong & Ćirković, That is not dead which can eternal lie: the aestivation hypothesis for resolving Fermi’s paradox (2017). Look at the Landauer bound again: the cost of erasing a bit scales with temperature. Today the cosmic background sits at about 2.7 K, a hundred times colder than room temperature, and it keeps cooling as the universe expands. A civilisation that cared mostly about total computation would get far more of it by waiting. Sandberg and colleagues estimate a gain of around 10³⁰ by holding out for the far future. So the most advanced minds might gather their resources and then go dormant, like animals sleeping through a hot summer. That’s what aestivation means. On that view the galaxy looks empty because its owners are asleep.

I love this idea because it’s the first Fermi answer that comes from thermodynamics instead of sociology. It’s also probably wrong, and the reason is worth knowing. In 2019 Charles Bennett, Robin Hanson and Jess Riedel pointed out that the argument assumes waste entropy can only be dumped into the cosmic background.^[8]8. Bennett, Hanson & Riedel, Comment on ‘The aestivation hypothesis for resolving Fermi’s paradox’, Foundations of Physics 49, 820 (2019). But the universe today is full of things that are far from maximum entropy, and those can absorb a computer’s entropy at the same efficient rate now as later. If you can pay your entropy bill today at no extra cost, the case for waiting collapses. The engineering version: you don’t need a colder room if you have a heat sink with spare capacity sitting right there.

There’s a deeper problem, too, which no hypothesis about motivation fully escapes. For the silence to hold, essentially every civilisation has to choose quiet, forever. Jason Wright and colleagues call the failure to take this seriously the “monocultural fallacy”^[4]4. Griffith, Wright et al., The Ĝ Infrared Search for Extraterrestrial Civilizations with Large Energy Supplies. III, ApJS 217, 25 (2015). The framing of why waste heat is the thing to look for is in paper I: Wright, Mullan, Sigurðsson & Povich (2014).: it’s easy to explain why a typical alien wouldn’t expand, much harder to explain why none ever does. One restless faction, one self-copying probe, one mind that values computation now, and the galaxy fills up. Their paper estimates that the galaxy could be crossed in under a billion years, which is a small slice of its age.

This is where the post-biological story cuts against itself. Machines make Fermi’s original problem worse, not better. A mind that can pause itself, copy itself, and survive centuries of cold vacuum is exactly the kind of thing that crosses interstellar space easily. If post-biological minds are the norm, the question “why haven’t they reached us?” becomes more pointed, not less.

§What “wrong” can honestly mean

So what’s left of the title? There are two defensible senses in which the Fermi paradox is wrong, and one I don’t think holds.

The first is mine: the popular paradox, “they should be talking and they aren’t”, rests on a bad assumption about what intelligence turns into. If minds go post-biological quickly, radio silence isn’t surprising at all. I hold that with some confidence. It’s cheap to believe, because it only needs other civilisations to do roughly what ours appears to be doing.

The second is the stronger, colder one, from Sandberg again, this time with Eric Drexler and Toby Ord. Their 2018 paper argues that the paradox comes from plugging single best-guess numbers into Drake-style equations whose terms are uncertain across many orders of magnitude.^[9]9. Sandberg, Drexler & Ord, Dissolving the Fermi Paradox (2018). Carry that uncertainty through properly and there’s a substantial chance we are alone in the observable universe. If so, an empty sky is not a contradiction that needs explaining. It’s just one of the likely outcomes. In that sense there is no paradox, only a question we can’t answer yet. Honestly, this is the better argument, and it doesn’t need aliens of any kind.

The sense that doesn’t hold is the one the headline tempts you towards: that we’ve explained the empty sky because the aliens became computers. We haven’t. Post-biology tells you not to expect chatter. It also tells you to expect heat and expansion, and so far the heat searches have come back clean at the scale of whole galaxies, with a handful of single-star candidates that keep turning out to be distant dusty galaxies in the background.

What I take from all this is mostly a change in where to look. I’d spend less effort on the thinly spread hope that someone is signalling us. I’d spend more on the one signature no civilisation can opt out of, whatever its substrate: thermodynamics. Biology was optional. Waste heat isn’t. If the galaxy has inhabitants who think in silicon or something stranger, they are still running on an energy budget, and every energy budget shows up somewhere as warm.