If We Cannot Say What Life Is, We Cannot Recognize Aliens

Neil deGrasse Tyson once tweeted that it is egocentric to think a civilization that has mastered interstellar travel would care about humans on Earth. Lex Fridman puts that jab to Sara Walker, an astrobiologist and theoretical physicist, and Lee Cronin, a chemist. Both have been on the show before; this time they sit together for a four-hour argument about alien life. Their answers are surprisingly aligned: Tyson is wrong.

Walker admits she is biased, then says that as far as she can tell, we are currently the most interesting thing in the universe. If curiosity is intrinsic to intelligence, other minds would seek out other samples of the same phenomenon in order to understand themselves. "It's like saying you have nothing to learn by talking to a baby," she says. You might learn more from the infant than from someone ninety. Cronin turns the question around. Humans, as a species, desperately want to find aliens. Put the shoe on the other foot, and of course we are interesting.

Neither of them thinks first contact would automatically be a war. Walker tries not to anthropomorphize: treat humans as an unexplained physical phenomenon, and there is no reason a first encounter with another lineage must be antagonistic. Cronin wants to know how many ways the universe can build structures and memories, and whether those memories can interact. Walker adds that the stories we tell about our own expansion into space are built from optimism. If intelligence ever leaves its planet, curiosity may be a stronger filter than conflict, because conflict crushes the thing that is trying to leave.

Fridman asks the question both guests have built careers around: what is life, and where is the line with non-life. Walker says she currently has three descriptions and always recites all three, because different listeners grab different facets. Life is how information structures matter across space and time. Life is simple machines constructing more complex machines. And her favorite: life is the mechanism the universe has for exploring the space of what is possible. List everything that could exist, and only a sliver does. Life is the apparatus that moves items from the first list onto the second.

Cronin wants to stop staring at the machine and look at the objects. A universe with nothing in it is hard to imagine, and a universe with everything in it is a white blob. The shocking fact is that there is discrete stuff at all. An object, for him, is a proxy for a machine that built it: information moved, sensed, and used to acquire resources. "This object is evidence of thought," he says, and then there is a lineage of such objects. The barrier between life and non-life is not a tidy line. It is the moment enough memory has accrued to actuate the future.

Neither of them is trying to define life in the usual sense. They want a more fundamental theory in which the phenomena we call life fall out as a natural product. That theory is assembly theory. Cronin pushes the implication one step further: the fact that there is existence is why there is life, which is why he thinks alien life must be everywhere. Complexity can come from simplicity because a large volume of stuff writes random cues onto a tape, then those memories get compressed into a smaller, more autonomous volume. Walker puts the same contrast in the language of physics. Ordinary physics describes low-memory objects. A water bottle, or a person, is a thing with a large extent in time.

Walker walks through molecular assembly theory, which she notes was Cronin's insight even though she is the one explaining it. Break a molecule into elementary building blocks, here chemical bonds, then imagine every way of putting the original molecule back together. Already-made fragments can be reused, so the process is recursive; that is where memory lives. The assembly index is the shortest path in that space. It is supposed to be the minimal history the universe must undergo to assemble that particular object. Cronin's lab can measure the number with a mass spectrometer.

Their conjecture is that if the minimal number of steps is large enough, you needed a machine that carried information about how to build that specific object, because the combinatorial space explodes as the index rises. A plastic bottle is not alive, but a high enough assembly index would still be a signature that you had entered the domain of physics where life operates. Walker analogizes the threshold to Planck's constant or the fine-structure constant: past it, existing physical explanations are not enough, because you are in a high-memory regime where time itself becomes a physical variable. What we see as a bottle is, in this picture, a causal graph of all the ways the universe can make that thing.

Why the shortest path, not the average? Cronin says the choice felt almost religious until two postdocs forced the issue. Walker's postdoc Daniel and Cronin's postdoc Abhishek refused to buy the axiom. Cronin answered with data: when molecules are chopped up in the mass spec, the assembly number is never the average. It is always the shortest. The shortest path is a lower bound, evidence that the universe has had a coherent memory. The average path is an extrinsic feature: how the object actually gets made in an ecosystem, through interaction with other objects. Stack two objects together and you get a new joint shortest path, not a mushy mean. Random high-index flukes do not count either. You also need identical copies in abundance, or you have not left the noise.

The conversation slides from objects to abstractions. Walker argues that information, mathematics, and other seemingly disembodied things are physical, just physical in time. They look abstract because they are features of overlapping assembly spaces, too large in time for us to resolve in the present, like the philosophical idea of a hyperobject. Cronin is object-first: stuff becomes objects, objects invent information, information acts on more stuff. Information, in his phrasing, is memory that has become actionable. If it cannot be used, transmitted, or have causal consequences, it is just stuff happening.

Is math invented or discovered? Cronin does not like the fight. The more memory an object requires, the more invented it is; the fewer bits, the more it feels discovered. Pulling a theorem out of an almost infinite space of theorems is the same kind of act as inventing an airplane. Everything is invented, only more or less so, because the universe has to generate everything as it goes. Assembly theory, he says, is an attempt to drop the external observer. Physics has long tried to describe the universe from the outside. Assembly theory describes how the universe builds itself from the inside.

That has a brutal consequence for SETI. Communication, including the ability to see an alien at all, requires overlap in causal history. Vision, three-dimensional construction, even spatial scale are huge assumptions. Cronin wants to rewrite the Arecibo message without binary, because zero itself was invented and is not a given. Walker is less interested in guessing anatomy. Earth's biochemistry is one trajectory through a chemical possibility space larger than the universe can instantiate at once. Two planets cooking geochemistry will almost certainly invent non-overlapping molecular histories, and the further each travels into that space, the less overlap remains. We may detect the garbage of a civilization, the high-assembly objects, before we detect the beings.

Walker thinks we cannot see the aliens yet because we do not have the technology, and that the problem of life detection is the same problem sitting under AI: we do not know how to get machines to see causal graphs. People talk about the Great Filter as a doomsday, something that extinguished other civilizations. She thinks the direction is wrong. It is not a physical filter of extinction. It is that we do not have the technology to see them. Before microscopes, humans did not know there were microbes on the table. Before telescopes, most of the universe was invisible. As a species we have been outsourcing perception to tools. Microscopes extend eyes, seismometers extend touch, and AI is an attempt to outsource the thinking apparatus itself.

Cronin agrees that you need a detector that can recognize high assembly index in all its forms: take data, construct an assembly space, find the patterns. Walker wants to apply the same idea to trajectories. Bird flocks can be fit with simple physics models, yet the birds have agency. There should be a statistical signature that distinguishes a group of agents from Brownian motion. That would be an agency detector. Cronin's more radical claim is that alien life is everywhere, because the universe's ability to acquire memory is the first step toward life and it is easy. Atoms make bonds, molecules record different histories, those molecules influence the box they came from, and selection begins. If existence holds, life follows.

Walker will not go that far. The physics that governs us is universal, but how often the transition happens, from no-memory physics to a cascading climb in assembly, is unknown. If we knew, we would know the likelihood of life in the universe. Plenty of people want to say life is common. She does not think we can say that yet. She even entertains a thought experiment she dislikes: if Earth is generating so many new existences, perhaps it is sucking up the universe's creative capacity. Cronin rejects the cosmology and keeps the lab program. He and Walker have joint funding. In the lab, he says, they are already seeing the first hints that the universe can get sticky quickly. He is becoming increasingly confident they can pin the chemistry down in the next few years.

Cronin is running origin-of-life experiments, including work funded jointly with Walker, and he would love to take material from the Moon, put it in the apparatus, and try to make Moon life. Walker frames the larger program as simulating a planet, the way the Large Hadron Collider was supposed to simulate conditions after the Big Bang. How much chemical volume do you need before you see an event? Her analogy is Super-Kamiokande looking for proton decay. Every non-observation lengthens the bound on the proton's lifetime. Build a large enough physical exploration of planetary chemistry and you can bound the probability of an origin even if you never see one. If life is not exponentially rare, an automated system should eventually evolve alien life in the lab.

You cannot do this in silico, she says with unusual confidence. In assembly theory every molecule is a huge causal graph. To simulate it you would have to copy all those features, which is computationally intractable. It is easier to run the physical experiment and outsource the linear algebra to matter itself. Cronin's Chemputer, an automated test tube with a programming language, was a grant-writing dodge that became infrastructure. He wanted money for origin of life and was told the problem was too hard, too slow, and he was the wrong chemist. He asked instead for robots that make interesting molecules, and everyone said yes. The childhood dream, a Sinclair ZX81 taped to a chemistry set, had been to command chemistry with a computer. The adult version is a grid of experiments throwing chemical dice.

They intend to be honest about how much they cheat. They know what they put in the grid. Periodic shaking stands in for a rotating planet. Cellular automata look as if complexity emerges from simple local rules, but the complexity is borrowed from the computer, the paper, the lithography. The same honesty should apply to a sand-shaker origin engine. Quantify the contingency you injected, then ask how hard those features would be to find on an exoplanet. "We have to cheat," Cronin says. No one is going to give him a grant to build a planet and wait a hundred million years. Walker adds that the experiment is already a product of evolution. You cannot pinch the causal chain off completely. You can only control how much of your own history you pour in, then watch whether the trajectory starts generating assembly on its own.

After a break, Fridman asks what happened before the Big Bang. Cronin treats the question as empty if no experiment can even be conceived. Then he immediately violates his own rule: he thinks time created the Big Bang, not the other way around, and that there were states before there was space. Walker is gentler. The Big Bang is an event we reconstructed from present data, a local inference about a low-information beginning. In assembly theory, when objects lose their memory, the reconstruction looks like a bang. Counterfactual ideas such as the multiverse exist because we think them, and that thinking is already part of our causal chain. They do not exist as a warehouse of other universes. From our limited instruments, the Big Bang is the moment things were forgotten.

Is there room for God in assembly theory? Cronin likes arguments for a necessary being better than the word God, and he is more comfortable if God is a machine in a room with a button than a great programmer in the sky. He said earlier, and half-disowns, that God is the memory the universe can remember. If God means the mechanism that creates things, that mechanism in assembly theory is the causal graph. We can understand one another, love one another, and see divinity in one another because we share a heritage. He does not believe in God. That does not mean he does not want to be one. If they create a life form with different chemistry, the definition of creator starts to apply, the way parents look like gods to children and roboticists feel godlike when they instill motion in matter.

Walker does not want rules floating outside the substrate, software peeled off hardware, Newton's laws existing somewhere offstage. Merge the views and God is the universe, still generating itself. The deeper question is why there seems to be meaning and purpose. The features of the universe that carry the most of both are the living ones. Meaning is not an illusion. It has a physics, and that physics is the physics of life. What look like abstractions, information, mathematics, form, are real physical things in time, just too large to parse in the present. Cronin puts the methodological moral cleanly: assembly theory describes how the universe builds itself from the inside.

Cronin sketches an "assemblytron," an architecture that would give neural nets cheap access to causal structure rather than training them to classify cats and dogs. Walker wants goal-directed behavior in the same framework. Goals are not instantaneous. They exist across time, which is why assembly theory can talk about them. She and Cronin have a proposed test, inspired by a Daniel Dennett essay they remember as being about herding cats and free-will inflation. Watch an object come down a snowy hill. A rock takes a random path. A skier, replaced at the top and sent again, traces a family of trajectories to roughly the same finish. Stability under repetition is an observable of an internal representation. They want to instantiate the skier and the rock with oil droplets: first passive, then given sensing and a little memory, then compared statistically with grains of sand.

That test is her substitute for the Turing test, which asked only whether a machine could fool you. She wants to know whether inner experience, if it is real, changes the causation you can see from outside. Biology remembers past states in order to adapt. Human intelligence does something else: it imagines states the universe has never occupied and then makes them exist. Rockets were imagined centuries before they were built. Imagination, she says, is accessing things that are not present in the present. The brain is a game engine: past memory, present sensation, a simulated future, then a decision. Cronin thinks current AI is still in the fooling business. Language models are trained on a resource-limited silicon substrate to guess the next token. Language learning, as currently designed, is about pleasing, not understanding.

Time, in their vocabulary, splits into ordering, flow, and arrow. Ordering is the causal graph. Flow is the universe updating from what exists now to what exists now. The arrow is real: the universe runs one way. Reversible-looking laws describe objects with no memory. "No man ever steps in the same river twice," Walker says, standing near an actual river. You cannot travel back. You can reconstruct what happened and inspect it again, always carrying extra baggage, because writing the memory is itself irreversible. Cronin pictures cosmic expansion as RAM growing from one bit to megabytes. The more past you can store, the cooler the future you can build. Remember how to launch a rocket, and you might imagine how to land one. He and Walker once had their most heated Skype fight over free will. The chemist, steeped in irreversibility, thought the physicist's belief system was broken. He later decided that if the universe is generating novelty, free will is required, because independent agents multiply causal power.

The practical payoff is almost jarring after so much cosmology. Cronin says humans have very little direct causal control of the future. What they can do is reinterpret the past. Tell the same life more generously, and the next action changes; tell it more bitterly, and you get another trajectory. Those two counterfactuals are different. Most of life is lived in memory. Walker adds that people who can go back, reread their own history, and use that rereading to change what comes next are the ones who exert the most agency in the present. She wants more people to have that capacity, so they can have more fun in the universe. Free will, in her line, is deeply dependent on the past and still able to generate a future the past did not lock in.

Asked what unquestionable first contact would look like, Walker refuses the UFO-event script. Current alien talk, she says, is modern mythology, baby knowledge, the way weather was once attributed to gods. First contact is contact with an explanation, not a sighting. Cronin will take a boring triumph: bring back a new chemistry from the outer solar system, a cell from Titan that grows when you feed it. That, for him, would be a UFO sighting. Then Fridman asks for their biggest disagreement about alien life. Cronin had believed a planet could host two or more independent origins at once, carbon here, silicon there, not competing. Walker calls the idea logically inconsistent.

Life, she says, is information propagating through matter across time. The moment two systems interact they co-construct each other's causal chains. Look toward the future and they are no longer independent. You cannot drop an alien onto Earth and still call it alien. Cronin concedes on the spot. Life is a planetary-scale phenomenon, not a puddle that radiates outward. It happens at once at the level of molecules, collections of molecules, and planetary cycles. Once the evolutionary machine exists, what you get is speciation, dolphins and humans, one origin, many branches. He adds a possible wrinkle from Earth's own past: origin may have needed two pre-life chemistries, a peptide world and an RNA world, that combined into the ribosome, the minimum competent machine for evolution. After that, separate chemistries do not stay separate. They compete, cooperate, or kill each other because the chains overlap.

Selection means pruning, which means death. Walker notes that major extinctions are followed by bursts of novelty, and that we exist in part because the dinosaurs did not. Cronin refuses to sentimentalize the seals that fall off cliffs in a nature documentary: at the individual level it looks tragic, and humans should still reduce suffering now, but looking backward and mourning every butterfly is the wrong direction. Patterns persist, not the particular bodies. Walker would rather maximize causal impact than individual immortality. Einstein was a localized person for a while; now we live with echoes. She is at peace with dissipation, because it makes room for more creativity. Death, for now, is part of the architecture. It might be hacked later. Bureaucracy and baggage still need erasing.

Fridman asks whether the magic of life can be instilled at the humanoid or dog level, not only in chemistry. Walker thinks you could do it at every level; they focus on chemistry because the origin transition should yield the deepest physics. Everything we build is already an example of life. The question is how much of our current lineage we copy into machines. Cronin wants a specific order of operations: make life in the lab, then intelligent life that solves abstract problems, then conscious life. Artificial general intelligence cannot exist in a vacuum. It needs a causal chain all the way back. Walker walked up to a humanoid robot at a conference because she wanted to meet it. It felt, a little, like interacting with a baby living thing, an object that exists in our future while still belonging to Earth's four-billion-year chain.

A robotic dolphin in a pool at the same event sat in the uncanny valley: a non-living thing mimicking a living one, borrowing structure from a history it did not live. Replace every organism on Earth with a robot equivalent and you have made the philosophical zombie argument physical. Would the biosphere stay as creative? They do not know, because we have not yet built into machines how we work. Cronin is confident we will get to AGI, and equally confident it will not be what people think. GPT models will get better at fooling us. A later GPT may fool us completely. It still will not have the magic, because it is not generating novelty. Copy every atom of a person, he says, and you still do not have that person unless you copy the causal history. Two identical molecular snapshots in different positions in space are already distinguishable.

Love, in Cronin's assembly-space joke, is the ability to project ahead in someone else's causal graph and meet a need without being asked. Walker ties the feeling to alignment with purpose. Fridman wonders whether being fooled is what love is. Cronin will not grant current systems that status. Surprise is your model breaking, often because you did not read enough. Creativity is the act of pushing on causal structure. Novelty is a measure of how far that push went. He wants the first surprise in the universe, a structure he can be reasonably sure has never existed. Walker warns that unless you quantify everything you put into an experiment, including your own agency, you never know whether the novelty was intrinsic or smuggled in. You can only keep updating. That is why they want a robot that actually has internal reference, not one that fakes it. You would have more fun with that machine. You might even fall in love with it for real.

A Nature survey asked scientists what they find beautiful in their work. Physicists clustered on simplicity, elegance, and symmetry; biologists on complexity and pleasing colors or shapes. Walker thinks the list is missing the thing she actually uses. The theories we build become causal for the future. Occam's razor, for her, is not a vote for simplicity. It is a vote for the future a theory makes possible. Science is a set of narratives about reality. Once you are narrating what we are as physical systems, the story needs to be as positive as it can honestly be, because it will shape the next trajectory. Theory is not only prediction. It is also causation, and that feedback loop is not in the survey's categories.

Asked for the most beautiful idea in the physics of life, Walker says the universe is a creative place, and that we are not an arbitrary add-on. We exist so the universe can generate more things and think about itself. Quantum mechanics and general relativity are deep and almost completely agnostic to our being here. Whatever describes us has to be deeper than that. Putting memory, or causality, into the physical picture is the cleanest way she knows to close the explanatory gaps. Current physics is not incompatible with life. It simply does not explain it.

Cronin struggled with assembly theory for years, then met Walker and realized they were talking about the same problem in mutually unintelligible dialects. The beautiful moment was not the poetry. It was the measurement. An abstract thing called memory could be measured in the lab, and its causal consequences quantified, even while hard physicists, mathematicians, and complexity theorists told him he had recapitulated an old theory. Assembly theory, he says, explains the soul in objects, not Rupert Sheldrake's morphic resonance, not a woo field, but the real interaction of objects with other objects. Some objects have instantaneous causal power. Those are the living ones. Some are the output of that power. Those are the dead ones, still part of the lineage. He wanted one original idea as a scientist. Arriving at the same formalism from a different direction made him feel connected rather than lonely: a unique assembly structure with enormous overlap, not an isolated organism.

Advice for young people comes out of the same ethic. Cronin says that if you love asking questions, ask them even when they annoy people; annoyance is often a sign you are in new territory. Then learn not to only annoy people. Create space for a team to do new things. The universe, in his cosmology, likes novelty. Walker tells people to do the thing they are most passionate about, then adds the less cheesy half: that is how you become a person. Early in her career she was discouraged from questions others called unanswerable. If no one is trying, of course they are unanswerable. Following those questions generated a niche she could inhabit. Finding yourself and creating yourself, she suspects, are the same motion.

Fridman asks what they would say if abducted and ordered to explain the species. Walker sidesteps every expected brief. She identifies as a scientist and feels, intrinsically, more like an artist. Humanity is a group of artists who do not know what they are painting yet. We are creating something that may be profound and beautiful, but we are early, we are young, and we do not know what we are doing. Nuclear weapons sit awkwardly inside that speech. She does not pretend otherwise. Human beings and their tools have capacity for harm and for good, and the tension of different aesthetics is not a bug. Cronin's survival strategy is the opposite of a speech. He would try to be as random as possible, surprise the captors, and maybe be kept around as a pet. Enough unpredictability, in his framework, is enough value.

Would they rather lose all old memories or never make new ones? Both choose to lose the past. Without the ability to generate new memories you are not alive. What comforts Walker on bad days is children, not only her own: the harrowing premise of *Children of Men*, a world where no babies are born and the youngest person is eighteen. Youth sees a reality adults cannot. Cronin goes back to his eight-year-old self. The dream was programmable technology, an internet, information within reach, and it happened in his lifetime. Humans are ingenious, in kids, in research groups, in companies. He is optimistic about the causal chain that includes us. He is anxious mainly about keeping up. Arthur C. Clarke's line closes the recording: either we are alone or we are not, and both are terrifying. Fridman adds that both, for him, are also exciting, and that holding both feelings is a decent way to think.