Put Hydrogen and CO2 in a Tube. Thermodynamics Wants Cells. Nothing Happens.

Ask anyone who works on the origin of life how it started, Nick Lane tells Lex Fridman, and you will get a different, passionately held answer. Each view is grounded in science and still, at this stage, an opinion: the subject is too large for anyone to hold more than a slice. Lane's slice is a biologist's. Why is life made of cells and of carbon? Why is it powered by electrical charges on membranes? Look for an early-Earth environment that matches those requirements, he says, and one jumps out. Whenever you have a wet, rocky planet you get hydrothermal vents, hydrogen gas in bucket loads, and electrical charges on cell-like pores that can drive the chemistry life still does.

The potion is almost disappointingly simple. Life is made of carbon. The Lego brick taken from air or ocean is carbon dioxide, and turning it into organic molecules means strapping on hydrogen. That is what life does: it hydrogenates CO2, taking hydrogen that bubbles out of the Earth in vents and sticking it onto carbon dioxide. "It's kind of really as simple as that." Do the experiments in the lab, and the molecules you get are the molecules at the heart of biochemistry. Oxygen, he adds, matters most to the origin of life by not being there. He had just chaired a conference session on whether oxygen mattered in the history of life. Of course it did. At the start, its absence was the point.

There is no spark in the Frankenstein sense. There is a continuous chemical reaction, and a continuous electrical charge that helps drive it. We tend to think of one zap and something comes alive, Lane says, without asking what then sustains it. We are sustained by oxygen, by an ongoing reaction: put a plastic bag on your head and you have about a minute. The source of energy at the origin is the reaction between carbon dioxide and hydrogen, and most of those steps are exergonic — they release energy. Warm hydrogen and CO2 to 50°C in a Falcon tube, add a couple of catalysts, shake, and nothing happens. Thermodynamically the two gases are less stable than cells. What should happen is cells coming out. Kinetic barriers stop them. That is where you need the spark, and it has to keep running.

Could life have started more than once? Lane says yes, with two provisos. Once oxygen is in the atmosphere, hydrogen reacts with oxygen instead of CO2 — rocket fuel, an explosive reaction — so vents today will not keep originating life. Before that, all we know is a single common ancestor. There could have been many origins that vanished. What makes him lean toward once is the deep split between bacteria and archaea. They look the same under a microscope and are nearly as diverse, yet their membranes differ, and they replicate DNA with completely different enzymes and genes. The genetic code is identical. Transcription into RNA and translation into protein are basically the same. They clearly share an ancestor, and they diverged in fundamental ways very early.

Imagine millions of vents and millions of potential common ancestors. Only one lineage makes it out first, takes over the world, and wipes out the rest. You would still see a single ancestor. The strange fact is that two got out, apparently from the same place, because they share the code. Inside a vent you rely on the vent's own electrical charges to power hydrogen plus CO2 so you can grow. Leave, and you must power your own membrane: you need pumps that move ions and create charge. The pumps look different in the two groups, as if both emerged from one ancestor and then diverged fast. DNA replication looks different because it is joined to the membrane, and the membranes are different. Lane does not think the first replicators were trying to survive. They were "kind of an outgrowth of the Earth."

Panspermia, and even the milder idea that amino acids and nucleotides arrived on meteorites, he does not find helpful. Delivery from space is unequivocal. What happens next is the question. Best case, you stock a soup. Then the ingredients have to "go gloop" and become alive. "You may as well say that a miracle happened. I don't believe in soup." A vent is a continuous conversion of a flow of molecules into more of yourself, self-organization from the start. Soup, by definition, has no structure. React CO2 with hydrogen and you get carboxylic acids, including the fatty acids that form bilayer membranes like soap bubbles — spontaneous order from thermodynamics, even as the entropy of the universe increases. Panspermia only pushes the question elsewhere. What he wants is a testable sequence from a sterile planet to cells, fork by fork, experiment by experiment.

Fridman wants a framework for telling a rock from a bacterium, on Earth or on another planet. Lane would rather not answer. The question is both impossible and trivial; there is no answer, only descriptions of processes over time. There are forty or fifty definitions. NASA's working version is roughly a self-sustaining system capable of evolution. He immediately objects to "self-sustaining": life is sustained by the environment. Pedants note that by that definition a rabbit is not alive, only a pair of rabbits, because one rabbit cannot copy itself. Biologists still cannot agree whether a virus is alive. A jumping gene, simpler than a virus, can convert its surroundings into a copy of itself. That is not a definition, Lane says, but a description: life parasitizes the environment to make an informationally exact copy — plants, animals, bacteria, viruses.

The moment it becomes alive, in his mind, is when you introduce genes into systems that are growing anyway. Put even random RNA into a protocell that is already doubling, and any change that lets it grow better or worse is caught by ordinary natural selection. Information is inside a cell system from the beginning. There is no story in which you build the hard drive first and then boot the machine. Michael Levin's planarian work, which Fridman was about to record, suggests evolution can also leap: knock out one ion pump and a flatworm grows a head like one from 150 million years ago, or a head no one has seen. The change is a single gene. Whether it is a hopeful monster that persists is a question of selection over generations.

The trajectory, if there is one, is planetary. Earth has iron and electrons at the center and a more oxidized outside, from the sun and from volcanic gases. The planet is a giant battery. Electrons flow from inside to outside in hydrothermal vents — the same topology a cell has. "A cell is basically just a micro version of the planet." Water is pretty fundamental, everywhere, and extremely good at what it does; carbon works in water especially well. If we found a thousand life forms, Lane would expect 995 of them to be carbon-based and living in water. A puddle, though, is not a sign that life exists or existed. You need a living planet: one that turns over its surface, brings electrons from inside to outside, and turns water into hydrogen. Once you have that, the rest is thermodynamics. Run Earth a million times and he would expect bacteria hundreds of thousands of times. His student Stuart Harrison has found patterns in the genetic code that, starting from CO2 and hydrogen, look a lot like the code we have. Life on Mars might not need to have been transferred to share a similar code.

Photosynthesis, in every form, is the same core move: stick hydrogen onto CO2, powered by the sun. The question is where the hydrogen comes from. In plants it comes from water. Sunlight splits water, hydrogen goes onto carbon dioxide, and oxygen is thrown away as waste. Lane calls it the single greatest planetary pollution event in the history of Earth. The pollutant is oxygen. It also made large, active animals possible, at least as we know them. Without an oxygenated atmosphere you can have animals of a sort; you cannot have animals that go around eating each other, or ecosystems as we know them.

The invention happened once. Splitting water takes a lot of power: you are separating charge across a membrane, and if the charge rushes back you get an explosion at the site. The wiring has to be exquisite. Plants have chloroplasts that were once cyanobacteria, and cyanobacteria are the only bacterial group that does oxygenic photosynthesis. Other bacteria take hydrogen from hydrogen sulfide and dump sulfur, or pull an electron from ferrous iron in the early oceans and bury themselves in rust. Oxygen's advantage is that the waste is a gas. It bubbles away. It did not accumulate quickly. We eat plants that got their energy from the sun, and we burn the food with their waste product. The circularity is the point.

Bacteria and archaea are prokaryotes: small cells without a nucleus. Super-resolution microscopes show astonishing molecular machinery; under an ordinary microscope they look small and simple. The earliest life you can see are stromatolites, stacked biofilms of things like cyanobacteria, large enough for the fossil record. They never produced animals or plants. Filamentous cyanobacteria are strings of cells, not organisms with tissues. The origin of the eukaryotic cell looks like an endosymbiosis: one cell got inside another. Lane personally thinks that is the single biggest invention in the whole history of life. It transformed the energetic possibilities. You end up with a supercharged cell that can support a much larger nucleus and many more genes.

Think of it as multi-bacterial power without the overhead. The bacteria living inside provide the energy currency, but each still has a genome that costs energy to express. Mitochondria, once bacteria, threw almost all of that away. Human mitochondria keep 37 genes; some protists keep 70 or 80; it is always a small number. Lane calls them a pared-down power pack whose control unit has been cut to almost nothing. Same power, far less spent on overhead, and therefore a much larger nuclear genome. Gene number went up about fourfold. Genome size and the ability to turn genes into proteins went up a hundred-thousandfold or more. Plants, animals, and even fungi start development from a single cell — oocyte and sperm fused into a zygote — so every cell has identical DNA and switches different genes on and off. The standard story is that this restricts conflict. Bacteria in biofilms are genetically different and cannot cooperate at that level. They fight.

It took about two billion years and, to the best of our knowledge, has not happened since. If it had not happened, Lane is certain Earth would be bacterial only. Information cannot explain the bottleneck. An *E. coli* cell may have 4,000 genes, but the strain's metagenome is about 30,000 — larger than the human genome of roughly 20,000. Bacteria have had twice as long and much larger populations, and they are still bacteria. The problem is structure. If the first cells needed charge on the outer membrane to grow, scaling up hits an engineering wall. Extreme polyploidy — tens of thousands of complete genome copies — is energetically ruinous and stops at a large bacterium. What you need is electrically charged power packs inside, control units intact, not in fatal conflict with the host. Conflict probably wins more often than not. Success changes the topology: a giant nuclear genome surrounded by small energetic genomes that run energy production. The new design does not appear overnight. It can take millions of years. Once you have cells living inside a cell, you have potential you did not have before, including enough genes in one egg to run a developmental program.

Why do all eukaryotes do sex? Lane's PhD student Marco Colnaghi had just published two papers on it in PNAS. "Everyone's interested in sex," Lane says. "The job of biologists is to make sex dull." It was invented with eukaryotes about two billion years ago. All eukaryotes share the same ground plan. Haploid gametes fuse, restoring two copies of the genome. Chromosomes line up, everything doubles to four copies, stretches are swapped — recombination — then two rounds of cell division produce haploid gametes again. Meiosis and syngamy. Single-celled eukaryotes do it; plants and animals do it almost the same way. Bacteria do not. They shuffle DNA with the same molecular machinery, taking bits from the environment and swapping them in and out.

Colnaghi's argument is an engineering one. You are an *E. coli* cell with 4,000 genes. You want 20,000, and you need to stop mutations from shooting the genome to pieces, using lateral gene transfer. You know something is wrong because regulation screams; you do not know which gene. With a small genome, picking up the right fragment is plausible. At 20,000 genes you would have to pick up DNA all day, in large chunks, and you would still get the wrong DNA. In the end you have to line chromosomes up. "You're forced into sex, to coin a phrase." Mitochondria gave the energy budget for a large genome. A large genome immediately creates a quality problem. Sex is the only way out. Sexual selection — peacocks, female mate choice, handicaps as advertisements — comes later. Even algal blooms run a related filter: cells release a toxin; those that can cope form a resistant spore, and those that cannot trigger programmed cell death. Genetically almost identical cells, different life histories, a binary choice.

DNA is a huge invention as a hard drive, but the code lives in RNA, and RNA almost certainly came first. The RNA-world idea is decades old: RNA can copy itself and catalyze reactions, cutting the chicken-and-egg loop. The ribosome is still a giant ribozyme. Getting to RNA, Lane suspects, is not that hard. Distinguishing messenger, transfer, and ribosomal RNA in a random population is an open question, maybe a phase transition. RNA genomes top out at viruses such as HIV. DNA is chemically a small change and a vastly more stable store. How easy the transition was, he does not know. Work on xeno-nucleic acids shows you can swap bases, backbones, even try arsenate for phosphate, within limits. He increasingly thinks we were channeled toward ribose, phosphates, and the letters we use, not that they won a free-for-all among competing chemistries. Fridman offers JavaScript: a messy language that took over because it ran in the browser first. Lane likes the analogy. Biology, too, is underpinned by mess. The human genome is full of broken sequence and old viruses, and some of it still becomes functional genes.

Predation starts in bacteria. Bdellovibrio, a favorite of Lynn Margulis, drills through a bacterial wall and membrane and eats the cell from the periplasmic space while making copies of itself. Predation is killing; parasitism is living on someone. Go back 560 to 570 million years, before the Cambrian explosion, and you find the Ediacaran fauna, sometimes called vendobionts. Stalked things with leaf-like fronds and fractal branching, living more than 200 meters deep — no waves, no storm damage, so not photosynthetic. They were filter feeders, stuck to the spot, like sponges and corals. Lane describes a gentle, rather beautiful, rather limited world. Then oxygen changed, and the fossil record fills with animals we recognize: eyes, claws, shells, plainly killing or running away. A vicious world we know, and evolutionary arms races. A cheetah does not run that fast unless the zebra can escape. Predation forces adaptation that would never have been possible without it, and in the end a more beautiful world.

You cannot have that world without oxygen, because it is about how much energy you extract from food. Without an oxygenated planet, perhaps 10 percent. With one, about 40 percent. That is the difference between one or two trophic levels and five or six: things eating things that eat other things. Fridman quotes Hunter S. Thompson on beauty requiring trampled souls. Lane does not flinch. Trillions of organisms were murdered in the struggle, and that violence is a catalyst for evolutionary creativity. "Evolution only happens because of death." Survival of the fittest, he notes, means death of the weakest. Morally, we cannot base society on the way evolution works. In some respects we do: science is competing hypotheses that die if they do not get better; city-states, universities, and companies compete and innovate. The problem is keeping the innovation without the death. That takes social rules, and someone has to write them. Do you let banks go to the wall, as evolution would, or impose regulation that the bankers will then exploit?

Wind the clock back five or six million years and let it run, and humans as we know them are not a high-probability outcome. An alien finding Earth full of zebras and giraffes would see a wonderful planet, not a spacefaring civilization. The extra steps from large mammals to humans, Lane thinks, are about the brain and about groups. Mark Thomas at UCL has argued that population density and interactions between groups — people moving like lateral gene transfer — let expertise accumulate. Isolated groups, Easter Island among them, often lose cultural complexity. Fire, tools, and language do not track the archaeological trajectory as well as density does. Lane's other observation is that humans are both deeply social and deeply selfish: the tug between those sides is the tug between socialist and capitalist instincts. Chimps run aggressive alpha-male societies; bonobos a female society that admits males on its terms. We have a lot in common with both. Neanderthal mitochondrial DNA diverged perhaps 600,000 years ago and looks unmixed; nuclear genomes show plenty of mating. Lane reads one-way gene traffic from humans into Neanderthals. Why they disappeared, he does not know. He suspects they were less violent, less clever, less populous, less willing to fight, and were driven to extinction at the margins of Europe.

Vision and movement, Lane says, are extremely important, and they show up with animals in the Cambrian. The old taunt — what use is half an eye? — does not survive a simple model. From a light-sensitive spot on flat tissue to an eyeball with a lens, assuming about 1 percent change per generation, you can get there in half a million years. That is not how evolution actually walked; it shows the steps can be reconstructed. Once you have an eye, there is nowhere to hide, and predator-prey logic takes over. Bats with echolocation, he suspects, form an image of the world much as we do; the difference is mental reconstruction. Some single-celled organisms have a lens, retina, and cornea — a camera eye in one cell, with no brain. They can build the structure. Once eyes exist, the central nervous system is under huge pressure to figure out what the signals mean, and how to run or stay put.

Fridman raises Donald Hoffman's claim that evolution does not reward seeing the world as it is, so perception is an illusion far from physics. Lane prefers "bubble" to "illusion." We take a fraction of possible sensory input and interpret what keeps us alive. The tree is physically there; walk into it and the collision is not a delusion. A child bangs into a table, feels pain, touches the edge, sees it, and binds the three inputs: do not do that again. Keep someone in the dark for years and the wiring for vision never develops. Limited reality, Lane says, but true reality. Fail to learn it and you get eaten or hit by a bus. A Matrix-style construction is hard to call impossible, but a simulation that convinces would have to join up touch, pain, vision, and the social world that tells you what things are. Babies are not born knowing how to use their eyes. They boot the system by bumping into trees. Fridman thinks humans hunger for a simple theory of whatever reality they are handed, which might make us cheap to trick. Lane agrees we try to integrate inputs into a hypothesis, and that we do not do it alone.

The conversation lands on consciousness, and Lane has a view even if he does not have a strong one. Ten years earlier he wrote a chapter for *Life Ascending*, subtitled "The Ten Great Inventions of Evolution," because a book with that subtitle could not skip consciousness. He does not work on it. He could not not be curious. He came away thinking nobody knows, and that two schools dominate. One says consciousness is a property of matter, an unknown law of physics — panpsychism, the sun and the rock conscious by degree. He finds that unpersuasive. We can tell living from not. The other says it is an emergent property of a complex nervous system. He dislikes how often "emergence" plasters over ignorance. As a biochemist his question is concrete. A depolarizing neuron gives rise to pain, love, anger. What is a feeling in biophysical terms? Which bit of the wiring? He has never seen an answer that makes sense to him. If we want to understand consciousness, he thinks that is the only question worth answering.

An AI that out-thinks us on information processing is, in his view, a matter of time, maybe already here. Designing a mind at least the equal of a human mind is not the obstacle. What people value most — feelings, emotions, a personal sense of the world — may matter as much as computation, and that is where he does not think AI will necessarily become conscious. He thinks it is a property of life. Fridman reads from the epilogue of *Transformer*: we can understand in principle how parallel processing could be wondrously intelligent, and we cannot answer in principle whether that intelligence would experience joy or melancholy. "What is a quantum of solace?" Lane's own conjecture goes back to single cells. A cell does not know it exists, yet it acts as a unit with astonishing precision. He links that to electrical fields on membranes: real-time feedback on how am I doing relative to my environment. Get it wrong, and a bacterium has an electrical discharge that says panic, and it runs. Bind that phase to that response over generations, and feelings come in through the back door. Complexify the system; he has no problem with a phase transition. Whether language alone can do it, he does not know.

Fridman wants to leave zombies and love to the philosophers and keep the feeling, which "sure as hell feels very real." If a toaster made you miss it, maybe that is a conscious toaster. Lane asks whether that is psychologically different from a dog. Most people grant dogs consciousness and deny it to toasters; you can look into a dog's eyes. Fridman argues that chess and protein folding are not the same order of intelligence as emotional conversation, driving, or walking. Humans discount what they are good at. Lane's fell-running memory makes the same point later: you can fly down a mountain with the conscious mind shut out, and the moment you try to place each foot you crash. Chess asks us to calculate many moves ahead, which we were never selected to do. Two or three or four is enough. Death, unlike society, hangs a penalty over every biological decision. That richness — environment, lifetime, generations, sequence change — is intrinsically hard to simulate. A sufficiently intelligent being, he thinks, watching HAL unplugged in *2001*, will have some form of consciousness. Whether it would be like biological consciousness, he does not know.

Yuval Noah Harari's *21 Lessons for the 21st Century* troubles Lane. Harari treats biochemistry as an algorithm that AI will hack better than humans: music beyond Mozart, prose beyond Shakespeare, because you only need to know which bits of the algorithm to play. As a biochemist, Lane finds the argument "close to irrefutable and not very enjoyable." He does not like the sound of it; that is a human reaction. Harari's last chapter then switches register — meditation, inner consciousness — as if that were a way out of the algorithm. Lane's distinction: the feedback loops are undoubtedly algorithmic; what we decide to do can run on another currency, feelings, over which we have little personal control. Consciousness, in his view, is what overrides a rational decision that cannot be made because the information conflicts: do this and see what happens. Fridman says the slogan is irrefutable the moment you utter it, and it simplifies away something fundamental, the way calling the universe an information-processing system misses how entropy makes local pockets of machines that then evolve. Reverse-engineering the whole biochemical algorithm of a human, he thinks, is centuries away. Capturing some of the magic of a dog's greeting could come much sooner.

DeepMind and AlphaFold will help biology, Lane is sure. He still thinks the field often asks the wrong question. A protein emerging from a ribosome comes out of a charged tunnel, one amino acid at a time, forced into a shape by that environment; chaperones step in when the default fold is not wanted. From first principles a sequence has a vast conformational space. Biology makes the problem simpler. Pattern-finding on all structures may recover the answer; looking at the exit tunnel tells you why it is this answer. Genome-wide association studies make the same mistake in his own area. Twin studies say a disease such as epilepsy is perhaps 50 percent genetic; the SNPs people find often explain less than a tenth. One response is to hunt a million weaker variants. Another is that researchers throw away mitochondrial DNA because it is small, even though a few letter changes there control energy and biosynthesis — and cancer cells rewire mitochondria toward building blocks rather than ATP. Fruit flies show nuclear-mitochondrial interactions that shift risk. Human intuition, bloody-mindedness about a neglected system, is what notices the missing piece. A hypothesis, Lane says, is a testable story, a leap into the unknown. He is not sure pattern-finding invents those stories. Magnus Carlsen told Fridman that AlphaZero did things that could easily be mistaken for creativity. Lane's fell-running moral still stands: solving chess may not mean we have solved downhill running.

Electrical fields in development are, for Lane, twenty-first-century biology that most biologists still treat as mumbo jumbo. So little is known about how a genome becomes a wired organism. EEG tells us about brain function without telling us which cells generate it; mitochondrial membranes inside neurons have more surface and a greater potential than the plasma membrane, with parallel cristae that could reinforce fields over distance. Drosophila mitochondria in the brain form astonishing rectangular arrays. Nobody knows if that plays a role in consciousness. Organoids, he has been told, can show feelings or even memories of a former brain. Until we control the fields that control development, he doubts we will turn an organoid into a real system. Reverse-engineering a human mind, he suspects, may be beyond the human mind. He is a biologist anyway. He is trying to understand the rudimentary outlines.

Karpathy, Fridman reports, votes that Lane's next book should be the Fermi paradox. How many alien civilizations? The universe is very big, so some. Not as many as most people would like. There is no inevitable march from bacteria to humans. The eukaryotic cell is the biggest gap; photosynthesis is another. A billion years separate eukaryotes from the first animals. From the Great Oxidation Event to enough oxygen for animals to respire took close to two billion years, apparently for planetary and geological reasons we do not really know. Without eukaryotes, Lane says, nothing else would have happened: it is tantamount to saying you cannot build a human from bacterial cells. Most biologists would not agree. An archaeal cell takes up an endosymbiont, and then basically everything to do with complexity happens. Plants, animals, and fungi share the same cell type despite utterly different ways of life, which suggests adaptation to pesky endosymbionts inside, not to ocean versus land. Maybe the process is nearly inevitable once you have endosymbionts. Maybe conflict produces war and extinction almost every time. Both readings are a bit grim. He cannot say which is true.

He expects billions, tens of billions, of planets with bacteria. Five to ten planets per star, and he would hope at least one has bacteria. Complex life will not follow automatically. Wind back five million years and a planet of giraffes is impressive to colonize, not to converse with. Finding life on Mars would delight him; what he would most want to know is how similar it is. Transfer between Mars and Earth is quite feasible, so a similar-but-not-identical code would be read as common origin. The other reading is deterministic chemistry from CO2 and water on a wet, rocky planet, producing a code not far from ours. Stereochemistry would settle it. Life uses D-sugars and L-amino acids; bacterial and archaeal lipids already use opposite stereo isomers, so both work. George Church has been trying to build life on the opposite isomers. Opposite stereochemistry on Mars would be unequivocal independent origin. Europa and Enceladus look like acidic oceans with hydrothermal systems. Titan's liquid methane would demand something very different.

Photosynthesis is his second-hardest leap, probably because the wiring is difficult, not because the niche was taken. Unequivocal oxygenic photosynthesis by 2.2 billion years ago, perhaps earlier; then more than a billion and a half years of stasis, the Boring Billion, oxygen producers balanced by consumers, 21 percent for hundreds of millions of years. Evolution does not crawl up a ramp. It sits in equilibrium until a tipping point, and the next stable state can be a long way off. After the Great Oxidation Event that state looked like today's Black Sea: oxygenated at the surface, sulfurous below, stable for more than a billion years, not obviously progressive. Snowball Earth, ice at the equator, forced change. One parochial-sounding mechanism Lane discusses: sulfate washed into the oceans. Enzymes prefer carbon-12 to carbon-13, so buried organic matter leaves oxygen behind and a carbon-13 signal in limestone. A huge negative isotope excursion just before the Cambrian would, if it were oxygen oxidizing carbon-12, strip the atmosphere of oxygen — yet other indicators say oxygen was there. Mass balance points to sulfate as the oxidant, reduced by bacteria, buried as pyrite. Continental collision lifted evaporitic calcium sulfate into mountains and into the sea. Without that, he says, the Cambrian explosion almost certainly would not have happened, and neither would animals. Evolution can invent an eye in half a million years. It can also do nothing for a billion. The Cambrian is not biology's great mystery. Consciousness is.

Why have we not seen visitors? The simplest answer, and the one he finds most probable, is that they are not in the right part of the universe at the right time. Humans happened once. The chance we survive the next few hundred years, possibly the next fifty or a hundred, seems to him quite small. Maybe the only thing that survives from humanity is AI, and once it can copy itself and cut humans out, maybe that takes over the universe. There is a catch. He does not see how a sterile planet gives rise directly to AI. You need organic life first, then an intelligent designer. Fred Hoyle's *The Black Cloud* challenged the teenage Lane to stop being parochial about carbon; speed-of-light limits and the need for populations under selection make a solar-system-scale dust mind, in Lane's later view, a poor bet. Richard Dawkins was right: nothing else generates that complexity. Earth is not Gaia the organism. James Lovelock's poetry became Earth-system science, regulation that keeps the world hospitable. There is only one Earth, so it cannot evolve by natural selection. Aliens arriving at night would notice cities first. Lane still would not call cities alive.

Lane opens *Transformer* with cities seen from space: grey, crystalline, parasitic scratches that obliterate the living Earth's blue-green, ugly until they light up at night. Controlled energy flow, information, metabolism of a kind — are they alive? Of course not. They are cities. Smart enough aliens, he thinks, would know that. We forget cells and think of biology as information; he wanted the comparison of energy flow through a city and through a cell. Cities are loosely governed. After the Great Fire of London, Christopher Wren planned Parisian boulevards without pubs or coffee houses. The plan died, probably for lack of money. Real London grew in jumbled streets, and the business of the City was done in those unplanned rooms. A cell is unplanned in the same way, not run by nuclear genes as textbooks imply, but an evolved entity with flux and animation. Lane will not push the metaphor through. Cities do not give rise to offspring cities. They propagate, if at all, as memes, through human societies, which exist because individual humans propagate. Without the humans, none of the rest exists. "I never came across a better idea than evolution." He rather hopes he is wrong about a lot of this, because the worldview is not uplifting. Science is about what is real. He thinks there is beauty in that too.

Why can human minds go to such dark places — depression, suicide, meaninglessness? Feature of life or bug? He does not know. If it is a feature, other organisms should have it. Dogs can mooch for days after an owner dies. He suspects it is a feature of biology, probably of mortality. You can also do the math, conclude it is all pointless, notice that society will be gone soon enough, and arrive at a bleak place by logic alone. Fridman finds it surprising we find any meaning at all. Lane wonders whether that is where consciousness comes in: transient joy, and with it transient misery. Biology almost never gets regulation right. There will always be a bell curve. Some people sit at the joy end, some at the misery end. That is how brains are wired, and he doubts there is an escape. Sex is the same. You cannot regulate it, so anything goes.

His writing method is a handful of questions. What is the story? What do I want to know? What do I want to convey? Why does it matter to anybody? The biggest questions are the childlike ones everyone wants to ask and does not, in case they look stupid. The longer you stay in science, the more you realize nobody knows, and asking is not stupid. He tries to ask what he would have asked at fifteen or sixteen, hungry, knowing little, wanting the edge of knowledge with a clean eye on the question. Who is the reader? In the end, only himself at that age, because you cannot know whether they are in the bath, on the metro, or half-listening to an audiobook. All you can do is please yourself. First say what we know; he was shocked, in the early books, how fast he hit what we do not. Supervising physics and maths PhDs, he found modeling is the same craft: what goes in the equation, the simplest unit, then the realization that if this is like that, that cannot be true. You navigate a landscape without knowing where it goes. He thinks that ride is why the books work. The editor of *Transformer* waited two months and said the first two chapters were a formidable hurdle. Lane put the manuscript down for a summer, then rewrote those chapters almost from scratch. The editor came back: now it had a barreling energy. Because the man had told the truth the first time, Lane believed him the second.

The only advice he gives students is to follow what they are interested in. There is a career path in science and there isn't: PhD, postdoc on another continent, two-year contracts, no guarantee of faculty, then tenure, prizes, and in physics the Nobel forever out of reach. "Happiness does not lie there." He was made redundant twice in a brief stretch of industry; that path has no guarantee either. People help you if they see light in the eye. You may not have a job in two years. If you care about what you are doing now, it works itself out. Fridman adds that you must also hunt mentors, not only topics; hallway conversations set a life's direction. Lane agrees, and adds obsession: if you love it, it is not work. Do not hunt happiness in success. He quotes Robert Louis Stevenson: "Nothing in life is so disenchanting as attainment." He has known people eaten by caustic resentment over a prize they think they deserved. Fridman notes the other failure mode: win the prizes, grow an ego too large to enjoy surprise. "Oh, man, to that, I couldn't agree more." Asked to summarize Earth, the first word is living. The distinctive word is eukaryotic, which empties rooms as fast as "Krebs cycle," so he reaches for "complex" and watches it fail too. On a universal scale, "mostly harmless" is probably accurate. Humanity is in no danger of disturbing the universe at the moment. Karpathy's gag — photons in, rockets out — gets a correction: that is a hell of a lot of photons before a launch. Lane's fear is not nuclear war so much as walking past a tipping point we cannot locate, assuming we can fix it in twenty years. He does not doubt we can solve it. He doubts we have the political will to do it as a planet, sharpish. He would rather be wrong than merely uplifting.