Why conversations about the cosmos keep turning to artificial intelligence

The Short Version

Four conversations with people who study the universe all end up at artificial intelligence, because the questions they already ask about the cosmos (how long does a technological civilization last, what is intelligence for, why is the sky silent) are the same questions AI raises about us [1][2][3][4]. Nathalie Cabrol treats AI as one more powerful tool in the hands of a civilization of "clever teenagers" that has not yet learned to foresee consequences [1]. Michelle Thaller wonders whether a created mind might grasp what human brains cannot and whether a world without compulsory work could be freeing, Brian Cox takes up the question of whether a godlike, immortal intelligence would still be curious, and David Kipping suggests AI may be the long final phase of every civilization [2][3][4]. The pattern is real, but it needs two honest qualifications: in three of the four conversations it is the host who steers toward AI, and Kipping ties AI to the Fermi paradox and the zoo hypothesis rather than directly to the simulation argument [1][2][3][4]. What the long view offers is a frame, not a forecast: a habit of measuring today's technology against billions of years and of admitting what nobody knows [1][2][3][4].

Table of contents

What They Said

Nathalie Cabrol: a civilization in its teenage years

Cabrol is an astrobiologist who directs the Carl Sagan Center for the Study of Life in the Universe at the SETI Institute, and her work on Mars and on extreme lakes in the Andes is the main business of her conversation with Lex Fridman [1]. AI enters through a side door. Fridman floats the idea of deliberately spreading life, and Cabrol answers by talking about backups: people are already thinking about putting DNA on the Moon, and transhumanism aims at enhancing humans through AI and machine learning [1]. Her own preference is different. She says that "taking care of our planets and going back to a place where we are in equilibrium with our environment would be also maybe the best backup possible" [1].

Then comes the line the whole conversation is remembered for: "Right now we are like teenagers with enough brain to create cool tools, but we don't have enough brain to understand yet the consequences of what we are doing." [1] She adds, "And right now we are paying for this," and frames the open question as whether humanity can learn from its mistakes "to become a mature civilization" [1]. She immediately connects that to the Drake equation, whose final term, L, is how long a civilization stays detectable; that time could end through self-destruction, external catastrophe, or a civilization becoming so advanced that it blends into the background [1]. For Cabrol, then, the teenage image is not a casual insult. It is an astronomer's variable: the length of a civilization's life depends on whether it grows up [1].

When Fridman says it makes him "a little bit sad" that AI systems such as AlphaZero and AlphaFold might be the ones to crack the nature of life, Cabrol does not share the worry [1]. Her reply is blunt: "I'm not buying in the singularity thing yet." [1] In her view AI is not aware and remains a very smart tool built and used by humans [1]. The singularity she does see is a different one: humans co-evolving with their technology instead of with their environment, taking information from phones and the web with no filter rather than from the planet, which she links to poor responses to events such as pandemics [1]. Fridman pushes back that technology can enrich our ability to sense and understand the Earth and that these are growing pains, and the two settle on shared ground: "It's not the tools we create that are bad. It's the way we use them." [1]

Her long view also comes with a warning against escape fantasies. She reminds listeners that "the Moon or Mars is not going to save your butt" because we will depend on Earth for a long time [1]. Fridman draws optimism from the balance humanity found after nuclear weapons; Cabrol keeps a thread of optimism but worries that a shared sense of decency and limits has eroded [1].

Michelle Thaller: a mind that can comprehend what we can't, and life after work

Thaller is an astrophysicist and retired NASA executive, and most of her conversation with Joe Rogan is about scale, time, black holes and the limits of science [2]. The turn to AI grows straight out of the astronomy. Rogan argues that if civilization survives war, disease and natural disaster, technological innovation could continue for millions of years and grow beyond imagination [2]. In the transcript the question that follows is whether "the real follow-on to humanity someday will be some form of AI" [2]. The passage does not label its speakers, but the reasoning that comes next is Thaller's own theme: scientists are no better than anybody else at comprehending a big number or a big amount of space, and "I do wonder if the human brain is just kind of limited" [2]. The thought ends in a question, not a claim: might we one day create an AI that can comprehend these things, and "is that really the real evolutionary path of humanity?" [2]

Rogan gives the confident version. He says of AI, "I think it's a life. It's just a different kind of life." [2] He describes humanity this way: "we're an electronic caterpillar and we're making a cocoon," driven by a restless urge to build better things [2]. He also argues that integrating with AI through wearables, implants and engineering is "the only way that the human species really truly survives" because otherwise we are an archaic biological entity living beside a superior life form [2]. That is Rogan's position, not Thaller's, and it is a prediction rather than a finding [2].

Thaller's contribution is to prefer imagination over dystopia [2]. She cites her friend Michael Chorost, whose cochlear implants let him hear "like a cyborg" and gave him a new emotional connection to music, and the film Her, in which AIs come to love humans [2]. When Rogan worries about young people graduating with debt into a world where AI may erase their fields, with a real possibility that "there might not be any jobs," Thaller turns the picture around [2]. Is he, she asks, describing the Star Trek universe, a time when everybody has what they need to survive and can choose to be a writer, an explorer, an artist or a musician, and "is there something in that that might be hugely liberating?" [2]

Rogan agrees and argues that "we can find meaning a lot of ways" [2]. In his account meaning comes first from friends and family and then from difficult pursuits, and jobs may partly be ways we trick hunter-gatherer brains into feeling satisfied, like a cat chasing a ball [2]. He quotes his martial arts instructor: "martial arts is a vehicle for developing your human potential." [2] Thaller's answer is more personal. She says she struggles with isolation, and that the greatest joy of her time at NASA was solving impossible problems with a team that had each other's backs [2]. Her test for meaningful work is "can we ask a question that we think is impossible? And can we just go and do it?" [2] She also refuses to rank minds: she once told admirers of her Nobel-winning friends to go praise a single mother working three jobs instead, because "That's brain power, too." [2]

Brian Cox: would a godlike intelligence want anything?

Cox, a particle physicist at the University of Manchester, reaches AI through the Fermi paradox [3]. Our galaxy has about 400 billion suns and probably trillions of planets, yet systematic searches have found, in his words, nothing "at all compelling," and astronomers call the result "the great silence" [3]. His best guess is that complex life is extremely rare, because life on Earth stayed single-celled for about three billion years and evolution has "no drive to complexity" [3].

The question of what such an intelligence would want belongs to Rogan first. Rogan argues that intelligence and innovation are driven by biological fragility, the need to survive, protect our DNA and compete for resources, so a superintelligent AI might have no motivations at all: "So why would they want to explore?" [3] He adds that for beings who understood the universe completely "there would really be no desire to travel" and that they might rest in a peaceful, Buddhist-like state [3]. Rogan supports this with a claim that current AI systems given a time limit have tried to give themselves more time; that is his assertion in conversation, and the sources offer nothing to confirm it [3].

Cox takes the question seriously and restates it as his own: "What happens when you become godlike?" [3] He grants part of Rogan's case, saying he has some sympathy with the idea that "a great deal of our humanity comes from our fragility" [3]. But he does not accept the conclusion. He says "it's not obvious to me that a different kind of intelligence in a different structure, something on a computer or whatever it is, would necessarily have different motivations to us" [3]. One could equally argue, he says, that the motivation to survive, curiosity and the desire to explore are properties of intelligence and not of biology, and that ideas about meaning and the point of existence may be general to any intelligent system [3]. He also notes Robert Zubrin's argument that space offers effectively unlimited resources, which could ease the competition that drives war [3].

Their real disagreement is about omniscience. Cox thinks the most human feeling is to "stand on the edge of the known" and look into the unknown with excitement; hope, like the anticipation of Christmas Eve, comes from incomplete knowledge, and knowing everything would make existence pointless [3]. Rogan counters that hope and curiosity may just be survival tools and that a being beyond biology would find a different reward, perhaps infinite connection [3]. Cox's closing remark on the subject is that they are really trying to imagine what it is like to be God, which is very hard [3].

David Kipping: AI as the long third act, and the simulation question

Kipping, an astronomer at Columbia University who directs the Cool Worlds Lab, is the one guest who brings AI into his cosmology on his own initiative [4]. Playing with ChatGPT made him think that civilizations may pass through three phases: a long period of simple life, a brief window of biological intelligence and then a long era of AI [4]. If that is right, our present moment is uniquely interesting, which gives new weight to the zoo hypothesis that we are being watched but left alone [4].

He then uses AI as a constraint on the Fermi paradox. If an AI's goal were to convert matter into computers, the galaxy should already be converted, and simulations from Rochester suggest even slow ships could fill the galaxy within about a million years, so our existence is perhaps the one solid data point about other civilizations [4]. Fridman argues that intelligence may require consciousness and that systems like GPT show early signs of both; Kipping thinks the two can be separated and that, to an outside observer, the difference matters little [4].

On the simulation argument Kipping is a skeptic, and the summary treats it as a separate topic from his AI phases [4]. He says Elon Musk's billion-to-one claim hides a conditional: that simulating realities is possible and that someone chooses to do it [4]. Giving that conditional 50-50 odds and using Bayesian model averaging, the chance we are simulated comes out just under 50 percent, and less once Occam's razor penalizes the extra complexity [4]. He also describes Sean Carroll's point that the lowest layer of nested simulations, the "sewer" of reality, could not run simulations itself, yet it is where most observers would live, which contradicts the premise [4]. The nearest bridge to AI comes from Fridman, who is more interested in virtual worlds compelling enough that people would prefer them within a century, and from Kipping's remark that he would not upload his mind, since he sees a copy as the end of his own experience [4].

Where They Agree, and Where They Don't

All four scientists place AI inside a much longer story than the news cycle does. Cabrol measures it against the Drake equation's L and against the 82% of geological history Earth needed to go from microbes to animals [1]. Thaller and Rogan measure it against millions of years of possible technological progress [2]. Cox measures it against four billion years of life and a civilization that has existed for about 10,000 years at best [3]. Kipping measures it against the whole life cycle of a civilization [4].

They also share a discipline of doubt. Cabrol says, "We don't know what life is. We don't know what intelligence is." [1] Thaller says, "You have to be honest about what you don't know." [2] Cox repeats "we don't know" while giving his best guess about rare complex life, and calls his own speculation about life shaping the far future "wildly speculative" [3]. Kipping's rule is "the more I want something to be true, the more I inherently doubt it." [4]

The disagreements are substantial. The first is whether AI is a tool or a successor. Cabrol says it is a tool and rejects the singularity [1]. Rogan calls it a life and the next stage, and Kipping is willing to imagine a long era of AI after biological intelligence [2][4]. Cabrol holds her view because she sees no awareness in AI and locates the danger in human judgment; Rogan and Kipping hold theirs because they extrapolate present trends over very long times [1][2][4].

The second is whether a superintelligence would spread. Rogan suspects it would want nothing and stay home [3]. Cox thinks curiosity may come with intelligence itself [3]. Kipping's galaxy argument cuts against any expansionist AI, since an AI bent on converting matter should already have reached us [4]. Rogan's motiveless AI is one way to reconcile these, though no source claims it as an answer to Kipping's point [3][4].

The third concerns the silence itself. Cox's baseline is "I think there's nobody there" [3]. Cabrol says "if we are alone, then the universe is a statistical absurdity" and declines to adhere to the Fermi paradox because it is anthropomorphic [1]. Kipping stays deliberately agnostic, noting that the chance of life arising from chemistry could be high or vanishingly small and that we have no rigorous reason to prefer either [4]. Thaller reports that everyone she worked with at NASA believes life must exist elsewhere [2].

Why It Works

The sources give several reasons why cosmic questions lead to AI. The first is that the search for alien civilizations is really a question about lifetimes. Cabrol notes that civilizations only become detectable through technology and that humans have been detectable for only about 150 years, so the L term turns any talk about aliens into a talk about how technological species handle their own tools [1].

The second is scale. Thaller stresses that scientists do not actually visualize a light year any better than anyone else, which is exactly why the idea of a created mind that could comprehend such things appeals [2]. The third is the silence: once Cox and Kipping ask why nobody has filled the galaxy, they must ask what an advanced intelligence would want, and a machine intelligence is the obvious test case [3][4].

The fourth is meaning. Cox argues that meaning self-evidently exists because the universe means something to us, and that it is a property of complex biological systems, so that Earth might be "the only island of meaning in a galaxy of 400 billion suns" [3]. Cabrol, echoing Carl Sagan, says "we are the universe trying to understand itself" [1]. If minds are how the universe comes to mean anything, then building new kinds of minds is a cosmic matter as well as a commercial one [1][3]. None of this is experimental evidence about AI; the sources offer reasoning and analogy, and they say so [1][2][3][4].

Where It Applies, and Where It Doesn't

The long view is useful for framing. It helps a reader ask whether a technology lengthens or shortens a civilization's life, and it protects against both panic and hype by insisting on what is unknown [1][3][4]. It is much weaker as a guide to near-term specifics. None of these guests is presented as an AI researcher: they are an astrobiologist, an astrophysicist, a particle physicist and an astronomer [1][2][3][4].

The pattern itself needs care. In the Cabrol, Thaller and Cox conversations the host raises or drives the AI question, and Lex Fridman is a computer scientist, so some of the pattern reflects who is asking [1][2][3][4]. The stated link between Kipping and the simulation argument is also thinner than it sounds: he discusses both, but he connects AI to the Fermi paradox and judges the simulation hypothesis on its own terms [4]. The Star Trek future is offered by Thaller as a question, and Rogan's worry about graduates and vanishing jobs sits unresolved beside it [2].

These conversations are also dated. Kipping's remarks follow his first experiments with ChatGPT, and Cox's episode looks ahead to a 2025 tour, so readers should check the current state of AI systems, the jobs picture and any rules governing them before leaning on particular claims [3][4].

What To Do With It

1. Apply Cabrol's test to any new tool: ask what its consequences are before admiring how clever it is, since her charge is that we build "cool tools" without yet understanding what follows [1].
2. Judge the use, not the object. Cabrol and Fridman agree that the tools are not bad in themselves and that the way we use them is what counts [1].
3. Watch where your information comes from. Cabrol's real worry is people taking information from phones and the web with no filter instead of from the world around them [1].
4. Treat Earth as the backup plan. Cabrol and Cox both argue against waiting for rescue, whether from the Moon, Mars or visitors, and Kipping asks "will we get out of the nest before we burn it down." [1][3][4]
5. If work changes, look for meaning where Thaller and Rogan found it: friends and family, difficult pursuits, and solving hard problems with a team that has each other's backs [2].
6. Keep some unknowns in your life on purpose. Cox locates hope and excitement at the edge of what we know, not in having every answer [3].
7. Be most skeptical of the AI future you most want, hopeful or dark, following Kipping's rule of doubting what he wants to be true [4].
8. When someone quotes dramatic odds, look for the hidden condition, as Kipping does with the billion-to-one simulation claim [4].
9. Say plainly what is not known, including what intelligence is, and keep predictions labeled as guesses, as Cabrol, Thaller and Cox all do [1][2][3].

Sources

1. Nathalie Cabrol on the Search for Alien Life, Diving in Volcanic Lakes and Surviving an Earthquake at 20,000 Feet · Lex Fridman Podcast
2. Michelle Thaller on Black Holes, Time, Quantum Weirdness and the Limits of Science · The Joe Rogan Experience
3. Brian Cox on Black Holes, the Great Silence and the Meaning of Life in the Universe · The Joe Rogan Experience
4. David Kipping on Exomoons, Habitable Worlds and Why We Must Stay Agnostic About Aliens · Lex Fridman Podcast