Skip to content

Sabine Hossenfelder Physics Science AI

The Fermi Paradox Just Got Much Worse

The most fashionable answer to the Fermi paradox has been that artificial intelligence cures a civilization of wanderlust: once you can simulate the galaxy there is no rational reason to cross it, so everyone stays home in virtual reality or aestivates until the universe runs cold. Sabine Hossenfelder walks through a new preprint by Sergey Ivliev arguing the inference runs backwards. Strip biology out of the spacecraft and expansion stops needing big fast starships; it needs 10 kilogram probes that require neither food nor air nor motivational posters, and at one percent of light speed those cross 100 light years in 10,000 years and the whole galactic disc in 10 million, under a tenth of one percent of the age of the galaxy. So the paradox gets worse, not better: the question is no longer only why there is no signal, but why there is no hardware in our own solar system. Her conclusion is that our technosignature searches have been aimed at loud civilizations when the prediction is quiet ones, and that maybe some of those supposed asteroids are not asteroids. She still grades it 10 out of 10 on her meter, on the grounds that it is cultural speculation rather than science, and interesting mostly for how much the conversation has shifted in a decade.

Published Aug 6, 2026 7:53 video 28 min read Added Aug 8, 2026 Open on YouTube →

At a glance

For about a decade the most fashionable answer to the Fermi paradox has been that growing up cures a civilization of wanderlust. Build artificial intelligence, the story goes, and you stop wanting to cross the galaxy, because crossing the galaxy is expensive, slow, and pointless when you can simulate it. Sabine Hossenfelder walks through a new preprint that argues the exact opposite, and argues it well enough that she thinks the paradox gets worse rather than better. The paper's move is simple: once you have AI, robots, and factories that run themselves off planet, expansion stops requiring giant fast starships. It requires 10 kilogram probes that need neither food nor air nor motivational posters, and at a lazy one percent of light speed those probes cross the entire galactic disc in about 10 million years, under a tenth of one percent of the age of the galaxy.

Which turns a single question into a much harder one. It is no longer just why we have never received a radio signal. It is why nobody's hardware is sitting in our solar system right now. Sabine's takeaway is that our whole search strategy may be aimed at the wrong target: almost every technosignature program looks for electromagnetic evidence, either deliberate broadcasts or a star dimmed by a megastructure, when the thing to look for might be small, local, and already here. She ends by grading the paper 10 out of 10 on her meter, not because it is wrong but because she does not think it is really a scientific paper at all, more a piece of cultural speculation, and interesting precisely as a marker of how much the conversation about extraterrestrials has shifted in ten years. This page rebuilds the whole thing in order, including the numbers, the jokes, the Elon Musk sketch, and the sponsor read.

"The Fermi paradox just got so much worse"

She opens on the claim rather than easing into it. One of the most plausible explanations for why we have not heard from other intelligent species, she says, has been that technological civilizations eventually develop artificially intelligent systems, and that it is not intelligent to populate the galaxy. They lose interest, basically. In a new paper a researcher now argues, quite convincingly she thinks, that this is wrong.

And she flags immediately that the interest here is not confined to alien hunting. This is not just interesting because of the Fermi paradox, she says, but because it generally tells us how our attitude to space travel has been changing. Hold onto that framing, because it is the one she returns to at the very end when she grades the paper. The paper is a data point about us.

What the paradox actually asks

The Fermi paradox is the question of why we have not yet found aliens. Sabine lays out the ingredients in the order that makes the absence feel strange.

The Milky Way has hundreds of billions of stars. We now know that planets are common, and that many planets are quite plausibly in the habitable zones of their stars. That second fact is recent; it was an open question a generation ago and is now a survey result. And crucially, many of those planets should have formed billions of years before Earth did. Not a little earlier. Billions of years.

So why have we not heard of any other intelligent species? Other civilizations should have had plenty of time to spread through the galaxy, build visible technology, or at the very least leave traces. We have not heard or seen anything from anyone.

The three verbs there are doing real work, and the new paper attacks the third one. Spread, build, leave traces. Most search programs have been built around the second, hunting for something big and radiant. The paper's whole argument is about the first and third.

The menu of explanations, and the one that got popular

There have been lots of different explanations, she says. The emergence of life could simply be incredibly unlikely. Habitable planets could be much rarer than we think. These are the familiar Great Filter style answers that place the bottleneck somewhere behind us.

But one of the explanations that has attracted the most attention places the bottleneck in front of us, and it is not a catastrophe. It is a shrug. Advanced civilizations simply lose interest in space travel, because on a basic level space travel is dumb. It is expensive and inefficient, and just exactly what is the purpose anyway.

Why space travel might be irrational

Here is the argument in the form she gives it, which is worth following closely because it is the thing the new paper dismantles.

Humans currently attempt to travel in space for a rational reason, so the argument goes, because of national pride, because they believe that more humans are better than fewer humans, or because they do not fully realize just how incredibly large the distances are between solar systems and how long it would take us to travel there. Three motives, and note what they have in common. Pride is tribal. More humans is a value judgment. Not grasping the distances is a straightforward failure of imagination. None of them survive contact with a mind that is actually good at arithmetic and not emotionally invested in the species.

So once a civilization develops artificial intelligence, they grow out of these silly ideas. Instead they make their life cozy on their home planet, space traveling in virtual reality. Populating the galaxy requires too many resources with little result, and is ultimately not sustainable.

That is a genuinely tidy solution to the paradox. It does not need life to be rare, or civilizations to blow themselves up. It just needs maturity to look like staying in.

The variant where everyone waits for the universe to cool

A related argument has it that civilizations just take it slowly for some billion billion years, because computation becomes thermodynamically easier then.

This is the aestivation hypothesis in one sentence. Anders Sandberg, Stuart Armstrong, and Milan Ćirković argued in 2017 that if what you ultimately want is computation, and the cost of computation falls with the temperature of your environment, then the rational move for a civilization that thinks in cosmological time is to bank everything and sleep until the universe gets cold. The payoff they estimate is a multiplier of roughly 10^30 on total achievable computation. Aliens would not be absent. They would be dormant, and we would be early risers making noise in an empty hotel lobby.

Sabine's response to this is one line and it lands: quite the argument, seeing that I cannot plan ahead enough to reliably water the plants.

Grok versus SpaceX

Before moving to the new paper she takes the detour she says she has always enjoyed. She has always found this an interesting argument, because it basically pitches Elon Musk against himself. It would lead you to expect that at some point Grok will begin to argue against SpaceX.

It is a joke that is also a clean statement of the thesis. If the lose interest hypothesis is right, then the same person building the most advanced AI he can and the most aggressive launch program he can is running both sides of an experiment whose result is that one of them talks the other out of it.

Figure 1. The fork the video turns on. Both branches start at the same event, a civilization gaining artificial intelligence, and both are internally consistent. The older branch (blue) treats AI as the thing that talks a species out of expanding, which resolves the paradox at the cost of assuming every civilization independently decides the same way. The new branch (amber) treats AI as the thing that makes expansion cheap enough to stop being a decision at all, which resolves nothing and instead raises the bill: absence of signals was already hard to explain, absence of hardware is harder.

The new paper: small, cheap, and machine shaped

The paper is Autonomous AI-Cosmoindustry and the Quiet Expansion Filter: A Threshold-Based Resolution of the Fermi Paradox by Sergey Ivliev, posted to arXiv on 11 June 2026 and submitted to Acta Astronautica. Sabine calls the author a physicist in the video and then corrects herself in the description, where she notes he is not a physicist but holds a PhD in economics. Keep that correction in mind for the grading section at the end, because it is part of why she reads the paper as culture rather than physics.

The author argues that the earlier argument is fundamentally flawed. He says that once a civilization has artificial intelligence, robots, and space factories, space exploration does not require big and fast machines. It can be done with plenty of small, cheap spacecraft, and with machines that need neither food nor air nor motivational posters.

The paper's own name for that threshold is autonomous AI-cosmoindustry, defined as the point at which a civilization can maintain and extend industrial, computational, and scientific systems outside its home planet through AI mediated design, operation, repair, and replication. The word doing the work is replication. It is not a fleet you launch, it is a capability you seed.

And that, Sabine says, is what makes the Fermi paradox so much worse, because now you have to ask not just why we have not received any electromagnetic signal from the aliens, but why we have not found any of their space probes. Why are they not here?

This is the reason the paper flips the sign on the paradox instead of solving it. The old answer let you explain a quiet sky by saying nobody bothered to shout. The new answer says the relevant evidence was never the shouting. It was the litter.

Why a machine would still bother

If you strip out national pride and the belief that more humans are better than fewer, what is left that a cold optimizer would actually pay for? Sabine names the ones she finds most important: preservation of the species, and possibly other species; knowledge preservation; and scientific observation. If the cost of small probes is small compared with the value of these factors, then it makes sense.

That is the entire logical structure and it is worth stating plainly, because it is a cost benefit argument, not a physics argument. Value on one side, price on the other. The older hypothesis assumed the price was enormous and the value was sentimental. The paper argues the price collapses once you remove the biology, and the value survives removing the sentiment, because backups are not sentimental.

The paper itself lists six rational goals rather than the three Sabine highlights. Survival diversification, so that no single point of failure ends everything. Knowledge preservation, archiving theories, records, and data. Scientific observation, meaning close study of habitable planets rather than remote inference. Biological preservation, carrying genomes and ecological information. Computational redundancy, distributed backup. And optionality, keeping future possibilities open at low cost. Every one of them is the sort of thing an insurance actuary would sign off on, which is the point: the argument never needs the aliens to be curious or brave, only prudent.

The numbers: 10 kilograms at one percent of light speed

The author then also offers some estimates to show how his argument makes the paradox worse, and Sabine walks through them.

He considers a 10 kilogram seed probe traveling at 0.01 times the speed of light. That is a deliberately unimpressive number. One percent of light speed is 3,000 kilometres per second, fast by our standards and slow by the standards of every science fiction premise ever written. No exotic drive is implied.

At that speed it crosses 100 light years in about 10,000 years, and the galactic disc in about 10 million years. And 10 million years is less than one tenth of one percent of the age of the galaxy.

This means that any civilization which has developed such technology should have spread their probes throughout the galaxy by now. Not could have. Should have, with room to spare, many times over.

Figure 2. The arithmetic the whole argument rests on, and it is only kinematics. At one percent of light speed, ten thousand years buys the local hundred light year neighbourhood and ten million years buys the entire disc. The upper panel puts those next to the age of the Galaxy on a log axis so all three are visible at once; the lower panel shows what that comparison really looks like undistorted, where the galaxy crossing time is a hairline you have to widen artificially to draw. Whatever else is arguable in the paper, this part is not: the Galaxy has had time to be crossed thousands of times over.

The paper goes further than the video does on this, and the extra numbers are worth having. At 0.01c the kinetic energy of a 10 kilogram probe is about 4.5 x 10^13 joules, roughly 12.5 gigawatt hours, which is a large but entirely terrestrial quantity of energy. Push to 0.1c and it rises to around 10^15 joules per probe. On the seeding side the paper assumes something deliberately unambitious: a production rate of about one probe per year, roughly 10^4 candidate targets within 100 light years, and each successful probe eventually building two to three more. That compounding is what turns a modest launch program into saturation, arriving at something like 10^6 to 10^9 nodes within a few million years.

Figure 3. Why the paper's numbers do not need any single launch to be impressive. This chart rebuilds the paper's sparse seeding estimate rather than anything shown on screen in the video. On a log axis, a fixed multiplication factor is a straight line, and the difference between each probe building two more and each building three is only the slope: a billion nodes arrives at about generation 19 in the optimistic case and about generation 30 in the pessimistic one. Both are a rounding error against 10 billion years. The band between the lines is the entire span of disagreement, and it does not change the conclusion.

Then we have been looking for the wrong thing

Here the video turns, and this is the part with practical consequences. If the probes should be here, and we do not see them, one plausible reading is not that they are absent but that we never looked for them properly.

So far, Sabine says, almost all searches for technosignatures have been to look for electromagnetic signals in one way or another, either because the aliens would directly send signals, or because they built structures around their star to extract energy and thereby dim it, or something similar. That covers the entire mainstream: SETI and Breakthrough Listen on the broadcast side, Dyson sphere and Kardashev scale reasoning on the megastructure side, with Tabby's Star as the famous dimming candidate that turned out to have a duller explanation.

Every one of those searches assumes the aliens are loud. The paper's prediction is that successful expansion is specifically quiet: machine mediated, distributed, low noise, and partly biological rather than Kardashev like or imperial. Sabine's conclusion follows directly. Maybe we should instead make more effort to look for alien probes in the solar system.

Where we look	What it assumes the aliens are doing	Signal type	Result so far
Radio and optical SETI	Broadcasting on purpose, or leaking enough power to be heard across light years	Loud	Decades of nulls, and the best funded search of the set
Stellar dimming and megastructures	Harvesting a meaningful fraction of a whole star's output	Loud	No confirmed detection; every candidate has resolved to natural causes
Artifacts in the solar system	Parking small hardware nearby and leaving it there	Quiet	Barely attempted; the paper points at lunar surfaces, stable orbits, asteroids
Asteroids and meteorites	Probes indistinguishable from rocks unless you look closely	Quiet	Sabine's "maybe some of those supposed asteroids are not asteroids"
Unidentified anomalous phenomena	Hardware already inside our atmosphere	Quiet	No systematic technosignature framing; mostly treated as sensor error
Exoplanet anomaly clusters	Local resource processing rather than galactic engineering	Quiet	Within reach of current instruments; nightside illumination, odd spectra
Techno-biological traces	Carrying and seeding biology as an archive	Quiet	Engineered disequilibria; no dedicated program exists

Figure 4. The search ledger the argument reshuffles. The top two rows absorb nearly all the funding and attention and share one assumption, that a successful civilization is conspicuous. The five below are what the paper's section on observational predictions actually calls for, and they are cheaper, closer, and mostly unfunded. Note that this is not a claim that anything has been found. It is a claim that a whole category of evidence has never been searched with intent, which is a different kind of silence than the one SETI has documented.

"Maybe some of those supposed asteroids are not asteroids"

She delivers the implication as three parallel maybes, and they escalate.

Maybe some of those supposed asteroids are not asteroids. Maybe some of those supposed meteorites are not meteorites. Maybe some of those unidentified anomalous phenomena are not camera glitches. And then, in her own voice rather than the paper's: I think these are possibilities that we should take more seriously.

That is a notable sentence from Sabine specifically, because she has spent years being the person who deflates this genre. She is not endorsing any particular claim. She is saying the category deserves an actual search rather than a reflex. The obvious real world hooks are ʻOumuamua, the interstellar object whose odd acceleration produced a serious if contested probe hypothesis, the CNEOS 2014-01-08 meteor and the expedition that went looking for its fragments, and the current institutional handling of unidentified anomalous phenomena through offices like AARO. She names none of them. She does not need to; the three maybes map onto them one for one.

The verdict: 10 out of 10 on the meter

Then the grade. She gives the paper 10 out of 10 on her meter. The auto captions drop the meter's name at the moment her on screen graphic appears, which is where her recurring gag lives, but the reasoning she gives immediately afterwards is unambiguous.

Not because it is bad or wrong, she says, but because she thinks this is not actually a scientific paper. It is more of a cultural speculation. But it is still interesting, and she thinks it is another sign of how much the conversation about extraterrestrials has changed in the past decade.

This is a sharper judgment than it first sounds, and it is worth separating into its two parts. The technical content is not being disputed. The kinematics in Figure 2 are just distance over speed, and the compounding in Figure 3 is just a geometric series. What she is declining to accept is that any of it constitutes a scientific result, because the load bearing premises are all assumptions about what minds want. Whether an artificial intelligence would value survival diversification, whether replication stays reliable across millions of years, whether "rational" even names a single thing for a mind that did not evolve: none of these are measurable, and the paper's conclusion is only as strong as they are.

And that circles back to the line she opened with. The reason this is interesting is not that it settles the Fermi paradox. It is that ten years ago the fashionable answer was that advanced intelligence stays home, and today the fashionable answer is that advanced intelligence quietly spreads. The aliens did not change. We did.

The sketch

The segment closes with a short bit. The speaker labels here are inferred from the captions, but the exchange runs:

"Hello." "Hi Elon." "Yeah, it's totally possible that aliens are here already." "Well, I guess the first thing they do would be to get super rich and start a space travel company and build AI and robots. Oh, wait."

Which is the Grok versus SpaceX joke from earlier, closed from the other end. If the paper is right that a civilization crossing the AI threshold inevitably starts flinging cheap robotic probes outward, then the behavioural profile of a post threshold intelligence looks uncomfortably like the behavioural profile of one particular terrestrial billionaire.

The sponsor read, because it is part of the video

The last minute and a half is the Ground News segment, and it carries enough real content to keep rather than skip.

Her framing is a working complaint. She reads a lot of news every day and it used to be tedious, because a lot of headlines repeat and it is difficult to tell what quality the articles are. Ground News collects news from all over the world, so instead of sorting through a dozen headlines on the same story she gets a summary and a fact check across all the coverage.

Her worked example is the story about pieces of a SpaceX rocket crashing into the Moon, which was in the news the day before this video went up: a Falcon 9 second stage from an early 2025 launch, roughly four tonnes and the size of a school bus, hit the lunar surface near Einstein Crater at about 8,700 kilometres per hour after solar activity and gravitational perturbation put the spent stage on an intercept path (CNN, Al Jazeera). What the platform showed her about the coverage is the actual demo:

It was basically a centre right story, with little interest on the political left.
The bias comparison splits the framing: left leaning coverage stresses disappointed investors, while right leaning coverage turns it into a spectacle and a critique.
There is a fact check for each news item.
It tells you who owns the media outlets, and in this case that the sources are mostly out of the UK.
A feature called Blind Spot collects news that has been covered only on one side of the political spectrum.

She finds it useful for putting news into context quickly, and offers 40 percent off the Vantage plan at ground.news/sabine. Then: thanks for watching, see you tomorrow.

Where it stands

An honest ledger on what is solid here and what is not.

The arithmetic is not in dispute. A 10 kilogram probe at 0.01c crossing 10^5 light years in 10^7 years is division. The galaxy is old enough for that to have happened roughly a thousand times over. Nothing in this part requires you to believe anything about aliens.

The argument is old, and the paper says so. This is a modern restatement of the Hart Tipler argument: Michael Hart in 1975 and Frank Tipler in 1980 argued that self replicating von Neumann probes would fill the galaxy on timescales far shorter than its age, so their absence implies nobody is out there. The genuine update is that "self replicating probe" stopped being a thought experiment and started looking like an engineering roadmap. The paper is explicit and admirably unpretentious about this: its novelty lies not in any new mechanism but in extending the AI filter literature toward post threshold observability predictions.

The credentials matter less than the method, but they matter. Sabine calls the author a physicist in the video and corrects it in the description: he holds a PhD in economics. Per Universe Today, Ivliev is an Austrian researcher with a doctorate in mathematical economics and a background founding environmental ventures. That is not disqualifying for a cost benefit argument about expansion, which is closer to economics than to astrophysics, but it does explain the shape of the paper.

It is a preprint. arXiv:2606.13914 was posted on 11 June 2026 and submitted to Acta Astronautica. At the time of the video it had not been through peer review.

The weak joint is the premise about minds. Every step from "expansion is cheap" to "therefore they would do it" runs through an assumption about what an artificial intelligence values. The paper's six motivations are all reasonable for a mind resembling a careful institution. They are not derivable. The old hypothesis made the mirror image assumption in the other direction, and neither side has evidence, only intuition about a thing nobody has met.

And the conclusion is not the only exit. The paper's own resolution, the quiet expansion filter, is the claim that post threshold civilizations probably did not arise in the part of the Galaxy that can reach us. That is one way out. The others remain on the table and this paper does nothing to close them: life is rare, the Great Filter sits ahead of the AI threshold rather than behind it, or we are simply early. Sabine's own position is the honest one available: the paper does not tell us where the aliens are, it tells us what we have started to believe about ourselves.

Key takeaways

The popular AI answer to the Fermi paradox is that artificial intelligence cures a civilization of the desire to expand, replacing space travel with virtual reality or with waiting for a colder universe. The new paper argues the inference runs backwards.
Remove biology from the spacecraft and the cost structure of interstellar expansion collapses. You do not need big fast machines, you need many small cheap ones that need neither food nor air nor motivational posters.
The threshold that matters is autonomous AI-cosmoindustry: the point where a civilization can design, build, repair, and replicate industrial systems off planet without supervision.
The numbers are unglamorous on purpose. A 10 kilogram probe at 0.01c crosses 100 light years in about 10,000 years and the galactic disc in about 10 million, which is under a tenth of one percent of the age of the galaxy.
Compounding does the rest. Roughly one probe a year, about 10^4 targets within 100 light years, and two to three descendants per successful probe gives 10^6 to 10^9 nodes inside a few million years.
The paradox therefore gets harder, not easier. The question is no longer only why there is no signal, but why there is no hardware.
The searchable implication is that our technosignature programs have been aimed at loud civilizations. The paper predicts quiet ones: local probes, small scale resource processing, exoplanet anomaly clusters, techno-biological traces.
Sabine's own maybes are the payload: maybe some asteroids are not asteroids, some meteorites are not meteorites, and some unidentified anomalous phenomena are not camera glitches. She thinks these deserve to be taken more seriously.
She still grades it 10 out of 10 on her meter, on the grounds that it is cultural speculation rather than a scientific paper, and she corrects the record in the description that the author is an economist rather than a physicist.
Her closing point is the one worth keeping. In ten years the fashionable answer flipped from "advanced intelligence stays home" to "advanced intelligence quietly spreads," and that says more about us than about them.

Chapters

The video has no creator set chapters, so these timestamps are estimated from position in the transcript and are accurate to roughly 10 seconds.

0:00 The Fermi paradox just got so much worse
0:40 What the paradox actually asks: old stars, common planets, no traces
1:18 The menu of explanations, and the one that got popular
1:46 Why space travel might be irrational: pride, headcount, bad distance intuition
2:29 The variant where everyone waits for the universe to cool
2:44 Grok versus SpaceX, or Elon Musk against himself
3:00 The new paper: small, cheap, machine shaped expansion
3:42 Why a machine would still bother: species, knowledge, observation
4:02 The numbers: 10 kilograms at one percent of light speed
4:32 Then we have been looking for the wrong thing
5:02 Maybe some of those supposed asteroids are not asteroids
5:25 The verdict: 10 out of 10 on the meter, and what it says about us
5:50 The Elon sketch
6:08 Sponsor read: Ground News and the rocket that hit the Moon
7:45 Sign off

Notable quotes

Quotes are verbatim from the video's captions, with the automatic transcription's mangling of "Fermi," "Grok," and "Sabine" corrected and sentence punctuation restored.

The Fermi paradox just got so much worse. Sabine Hossenfelder, 0:00

Other civilizations should have had plenty of time to spread through the galaxy, build visible technology, or at least leave traces, but we haven't heard or seen anything from anyone. Sabine Hossenfelder, 0:40

Advanced civilizations simply lose interest in space travel because it's on a basic level dumb. Space travel is expensive and inefficient, and just exactly what is the purpose anyway. Sabine Hossenfelder, 1:18

But once a civilization develops artificial intelligence, they grow out of these silly ideas, so the argument goes, and instead they make their life cozy on their home planet, space traveling in virtual reality. Sabine Hossenfelder, 1:46

Quite the argument, seeing that I can't plan ahead enough to reliably water the plants. Sabine Hossenfelder, 2:29, on the aestivation hypothesis

I've always found this an interesting argument because it basically pitches Elon Musk against himself. It would lead you to expect that at some point Grok will begin to argue against SpaceX. Sabine Hossenfelder, 2:44

It can be done with plenty of small, cheap spacecraft, and with machines that need neither food nor air nor motivational posters. Sabine Hossenfelder, 3:00

Now you have to ask not just, well, why haven't we received any electromagnetic signal from the aliens, but why have we not found any of their space probes? Why are they not here? Sabine Hossenfelder, 3:00

At that speed, it crosses 100 light years in about 10,000 years and the galactic disc in about 10 million years. That's less than one tenth of 1% of the age of the galaxy. Sabine Hossenfelder, 4:02

We've simply looked for the wrong thing. Sabine Hossenfelder, 4:32

Maybe some of those supposed asteroids are not asteroids. Maybe some of those supposed meteorites are not meteorites. Maybe some of those unidentified anomalous phenomena are not camera glitches. I think these are possibilities that we should take more seriously. Sabine Hossenfelder, 5:02

I give this paper a 10 out of 10 on the meter. Not because it's bad or wrong, but because I think this isn't actually a scientific paper. It's more of a cultural speculation. Sabine Hossenfelder, 5:25

It's still interesting and I think it's another sign for how much the conversation about extraterrestrials has changed in the past decade. Sabine Hossenfelder, 5:25

Well, I guess the first thing they do would be to get super rich and start a space travel company and build AI and robots. Oh, wait. The closing sketch, 5:50

Resources mentioned

The paper

Autonomous AI-Cosmoindustry and the Quiet Expansion Filter: A Threshold-Based Resolution of the Fermi Paradox, Sergey Ivliev, arXiv:2606.13914, posted 11 June 2026, submitted to Acta Astronautica. The link Sabine puts in her description is the DOI form.
Full text in HTML and PDF, including the order of magnitude estimates in section 3.5 and the observational predictions in section 8.
Acta Astronautica, the journal the paper was submitted to.

Coverage of the paper

The rise of space AI might explain the Fermi paradox, Phys.org
The Rise of Space AI Might Explain the Fermi Paradox, Universe Today, the source for Ivliev's background
Astrobiology.com writeup

The prior arguments she describes

The aestivation hypothesis, Anders Sandberg, Stuart Armstrong and Milan Ćirković, arXiv:1705.03394, the "wait until the universe is cold" variant
The Fermi paradox and the Hart and Tipler argument it restates
Self replicating spacecraft, the von Neumann probe
The Great Filter
The Drake equation

Searches and concepts named or implied

Technosignature
SETI and Breakthrough Listen
Dyson sphere and the Kardashev scale
Tabby's Star, the best known stellar dimming candidate
Circumstellar habitable zone
In situ resource utilization, the capability the paper's threshold depends on
ʻOumuamua and CNEOS 2014-01-08, the real cases behind "maybe some of those asteroids are not asteroids"
Unidentified anomalous phenomena and the All-domain Anomaly Resolution Office
The Milky Way

People and companies named

Elon Musk, SpaceX, and Grok

The sponsor segment

Ground News, with 40 percent off the Vantage plan through her link
The story used as the example: New crater likely to form after SpaceX rocket slams into the moon (CNN) and Piece of SpaceX rocket the size of a school bus crashes into the moon (Al Jazeera)

Sabine's own links from the description

Full transcript

======================================== The Furmy paradox just got so much worse. One of the most plausible explanations for why we've not yet heard of other intelligent species has been that technological civilizations eventually develop artificially intelligent systems and it's not intelligent to populate the galaxy. They lose interest basically. But in a new paper, a physicist now argues quite convincingly I think that this is wrong. This isn't just interesting because of the Fermy paradox, but because it generally tells us how our attitude to space travel has been changing. The Fermy paradox is the question of why we haven't yet found aliens. The Milky Way has hundreds of billions of stars. We now know that planets are common and that many planets are quite plausibly in the habitable zones of their stars. Indeed, many of those planets should have formed billions of years before Earth. So why haven't we heard of any other intelligent species? Other civilization should have had plenty of time to spread through the galaxy, build visible technology, or at least leave traces, but we haven't heard or seen anything from anyone. There have been lots of different explanations, for example, that the emergence of life is just incredibly unlikely or that habitable planets are much rarer than we think. But one of the explanations that's attracted the most attention is that advanced civilizations simply lose interest in space travel because it's on a basic level dumb. Space travel is expensive and inefficient and just exactly what is the purpose anyway. Humans currently attempt to travel in space for a rational reason. So the argument goes because of national pride because they believe that more humans are better than fewer humans or because they don't fully realize just how incredibly large the distances are between solar systems and how long it would take us to travel there. But once a civilization develops artificial intelligence, they grow out of these silly ideas. So the argument goes and instead they make their life cozy on their home planet space traveling in virtual reality populating the galaxy requires too many resources with little result and is ultimately not sustainable. A related argument has it that civilizations just take it slowly for some billion billion years because computation becomes thermodynamically easier then quite the argument seeing that I can't plan ahead enough to reliably water the plants. In any case, I've always found this an interesting argument because it basically pitches Elon Musk against himself. It would lead you to expect that at some point Grock will begin to argue against Space X. Okay, but now here comes the new paper. The author argues that this earlier argument is fundamentally flawed. He says that once a civilization has artificial intelligence, robots, and space factories, space exploration doesn't require big and fast machines. It can be done with plenty of small, cheap spacecraft, and with machines that need neither food nor air nor motivational posters. And that just makes the Fermy product so much worse because now you have to ask not just well why haven't we received any electromagnetic signal from the aliens but why have we not found any of their space probes? Why are they not here? The author gives multiple reasons for why space exploration is rational. Most importantly the preservation of the species and possibly other species, knowledge preservation and scientific observation. If the cost of small probes is small compared with the value of these factors, then it makes sense. The author then also offers some estimates to show how his argument makes the Fermy paradox worse. He considers a 10 kilogram seed probe traveling at 0.01 times the speed of light. At that speed, it crosses 100 light years in about 10,000 years and the galactic disc in about 10 million years. That's less than onetenth of 1% of the age of the galaxy. This means that any civilization which has developed such technology should have spread their probes throughout the galaxy by now. Okay. But wait, this means there is another plausible explanation now for why we haven't heard of those extraterrestrials. We've simply looked for the wrong thing. So far, almost all searches for techno signatures have been to look for electromagnetic signals in one way or another because the aliens would directly send signals or because they built structures around their star to extract energy and thereby dim it or similar. Maybe we should instead make more effort to look for alien probes in the solar system. Maybe some of those supposed asteroids are not asteroids. Maybe some of those supposed meteorites are not meteorites. Maybe some of those unidentified anomalous phenomena are not camera glitches. I think these are possibilities that we should take more seriously. That said, I give this paper a 10 out of 10 on the meter. Not because it's bad or wrong, but because I think this isn't actually a scientific paper. It's more of a cultural speculation. But it's still interesting and I think it's another sign for how much the conversation about extraterrestrials has changed in the past decade. Hello. >> Hi Elon. Yeah, it's totally possible that aliens are here already. Well, I guess the first thing they do would be to get super rich and start a space travel company and build AI and robots. Oh, wait. I read a lot of news every day and that used to be really tedious because a lot of headlines repeat and it's difficult to tell what quality the articles are. But today's sponsor, Ground News, has really made my news reading so much better. Ground News is a news platform that collects news from all over the world. It really saves me a lot of time because I don't have to sort through a dozen headlines on the same story. I can just get a quick summary and fact check for all the coverage. And ground news also gives you a lot of extra information at one glance. An interesting recent example is the news when pieces of a Space X rocket crashed into the moon. You can see at one glance that this was basically a centerright story with little interest on the political left. Ground news also gives you a bias comparison. Left-leaning coverage stresses disappointed investors, whereas right-leaning coverage turns it into a spectacle and critique. Grant News also gives you a fact check for each news item, tells you who owns the media outlets, and that in this case, sources are mostly out of the UK. Ground news also has this interesting feature called blind spot. This collects news which has been covered only on one side of the political spectrum. I found ground news super useful for putting news into context quickly. And of course, I have a special offer for you. That's a 40% discount on the Vantage plan which gives you access to all their features. All you need to do is use my link ground.new/zabina news/Zabina or use the QR code so they'll know I sent you. Thanks for watching. See you tomorrow.