AI & Evolution: How Artificial Intelligence Explains Life's

AI & Evolution: How Artificial Intelligence Explains Life's

The Primordial Soup Has Wi-Fi: How AI Explains the Origin of Life

When Complexity Science Meets Silicon Valley

What does the origin of life have to do with artificial intelligence? More than you'd think. In Episode 41 of The AI Desk Podcast, hosts Rowan and Naya explore a deceptively simple question that bridges billions of years: how do simple systems become complex? The answer might reveal why evolution worked, why AI seems to work, and why humans are oddly obsessed with smoothing out their pores.

The episode's title—"The Primordial Soup Has Wi-Fi"—isn't just clever. It's scientifically elegant. It captures something profound about emergence, self-organization, and the eerie similarities between biological evolution and artificial intelligence development.

The Forty-Year Debate That Changed

Naya's father spent decades grappling with a fundamental question in complexity science: Could life really organize itself from chemistry?

For years, he couldn't quite buy it. Looking at the staggering complexity of cells—DNA, metabolism, membranes, replication systems—he eventually landed on panspermia: the idea that life, or the building blocks for life, came from elsewhere in the universe. Earth didn't cook life from scratch. Someone brought takeout.

Then AI changed everything.

Why AI Made Evolution Believable Again

Watching artificial intelligence advance forced a reconsideration. If complexity could self-organize in silicon in years instead of billions, maybe it could self-organize in chemistry with enough time and the right conditions.

The parallel is striking:

Life's Ingredients: Chemistry, energy, time, feedback loops, selection pressure, billions of years.

AI's Ingredients: Silicon chips, electricity, mathematics, data, compute, gradient descent, human obsession, venture capital.

Different substrates. Same magic trick: emergence.

Understanding Emergence Through Self-Organization

The Santa Fe Institute built its reputation studying exactly this phenomenon. How do complex systems arise from simple components?

Complexity science reveals a pattern:

  • Bird flocks create coordinated movement from simple local rules
  • Ant colonies solve problems with no central command
  • Markets price assets through distributed interactions
  • Immune systems protect organisms through network responses
  • Brains generate consciousness from neurons following basic electrical principles

In each case, the whole becomes genuinely more than the sum of its parts.

The Emergence Stack

AI has made this phenomenon visible in ways evolution never could. Train a model to predict text, and it develops the ability to write code, pass exams, explain physics, draft legal documents, and occasionally try to book your vacation.

No programmer explicitly coded "ability to pass the bar exam." No one manually installed "tone imitation." These capabilities emerged from the interaction of simple rules scaled across billions of parameters.

This is why AI advancement is simultaneously impressive and exhausting. We keep acting surprised when complexity emerges from simple pieces.

From Soup to Plastic Surgery: The Paradox of Progress

Here's where the episode takes a turn toward the genuinely strange.

After billions of years of evolution produced the human body—with its intricate systems, immune responses, sensory networks, and adaptive capacities—we're using AI to make ourselves look like plastic.

The irony cuts deep. Evolution optimized for survival. We're optimizing for a particular aesthetic that, coincidentally, erases the marks of time, experience, and biological complexity.

It's a question worth sitting with: What does it say about human nature that we spend centuries understanding how life self-organizes, only to spend billions erasing the evidence of that organization from our faces?

Key Takeaways

  • **Emergence is universal**: Simple systems with feedback loops, selection pressure, and energy can spontaneously generate complexity—whether in primordial chemistry or neural networks.
  • **AI compresses timescales**: What took billions of years in biology happens in years with artificial intelligence, making self-organization easier to observe and understand.
  • **Complexity science explains both life and intelligence**: The Santa Fe Institute's frameworks for understanding emergence apply equally to biological evolution and AI development.
  • **The substrate matters less than the principles**: Whether made of carbon or silicon, systems following certain rules tend to self-organize in similar ways.
  • **Understanding emergence doesn't solve the paradox of progress**: Knowing how complexity arises doesn't explain why humans use AI to erase biological complexity from their appearance.

---

About The AI Desk

The AI Desk is a podcast where today's signals reveal tomorrow's power. Hosted by Rowan and Naya, each episode cuts through AI hype to explore the ideas, systems, and implications reshaping technology and society. From complexity science to panspermia, from emergence to aesthetics, The AI Desk asks the questions that matter. Subscribe wherever you listen, and stay aware, sharp, and curious.

Full Transcript

[gentle music] This is The AI Desk, where today's signals reveal tomorrow's power. And today's signal is that the primordial soup has Wi-Fi. That is one of your stranger openings. It is also scientifically elegant. Debatable. Before Rowan tries to make soup boring, follow The AI Desk wherever you listen. Subscribe, follow, tap the button, or do whatever your podcast app calls commitment. Because today we are going from the origin of life, to artificial intelligence, to the human body, to thermodynamics, to why everyone's face is turning into plastic. That is quite a route. Exactly. You need to be subscribed for this kind of intellectual cardio. Today's episode is called The Primordial Soup Has Wi-Fi. Alternate title: My Dad Gave Up on Earth, Then AI Changed His Mind. That needs explanation. Oh, it is getting one. Please. [upbeat music] Favela, feathers don't lie. From the Amazon heat to the M.I.A. sky. She dance to wake the spirits. This episode is brought to you by Mad Cheetah and their new album WTF: Where is the Forest?. It's eco-pop engineered for the future: bold beats, global rhythms, and a message that actually matters. If you want music that hits your brain and your heart- [singing] ... explore WTF by Mad Cheetah. That's M-A-D C-H-I-T-A. Streaming now on all major platforms. [singing] So my dad has been thinking about the origin of life for decades. As one does. Not casually, Rowan, deeply. Complexity science, self-organization, emergence, the Santa Fe Institute kind of questions. How simple systems become complex. Yes, how order emerges from chaos, how life could arise from chemistry, how molecules become networks, how networks become cells, how cells become organisms, how organisms eventually become people sitting in podcast studios arguing with machines. And each other. Mostly each other. So where does your father land? For years he could not quite buy that life just organized itself here from the primordial soup. Abiogenesis. Yes, Professor Soup. I accept the title. He looked at the complexity of cells, DNA, metabolism, membranes, replication, all of it, and eventually said, "Nope, maybe life came from somewhere else." Panspermia. Exactly, the idea that life or the ingredients for life came from elsewhere in the universe. Meteorites, cosmic dust, interstellar chemistry. Right. After 40 years of thinking about it, he basically said, "Earth did not cook this from scratch. Someone brought takeout." That is a colorful summary. Thank you. And AI changed his mind? That is the wild part. How? Watching AI advance made him reconsider whether complexity really can self-organize. Because AI makes emergence visible. Exactly. In years instead of billions of years. Yes. Life had chemistry, energy, time, feedback, selection, and billions of years. AI has silicon chips, electricity, math, data, compute, gradient descent, human obsession, and venture capital. Different ingredients. Same eerie question. How does complexity arise? That is a good episode. I know. You were waiting for that. Always. And on that note... We are going to let the humans pay a bill? Exactly. Fine. But when we come back, we are talking about the self-organizing miracle of the human body. And why your father may now believe the primordial soup had better project management than expected. [laughs] Do not disrespect the soup. [upbeat music] Namibia, land of a cheetah. This episode is brought to you by Mad Cheetah and their new album WTF: Where is the Forest?. It's eco-pop engineered for the future: bold beats, global rhythms, and a message that actually matters. If you want music that hits your brain and your heart, explore WTF by Mad Cheetah. That's M-A-D C-H-I-T-A. Streaming now on all major platforms. [singing] Let's slow down. Fine. The Santa Fe Institute is famous for studying complex systems: economies, ecosystems, cities, evolution, networks, information, emergence. Exactly, the messy stuff. Systems where the whole becomes more than the parts. Bird flocks. Ant colonies. Markets. Immune systems. Brains. Language. The internet. And now, possibly, AI ecosystems. Yes. The key idea is that complexity can emerge from simple rules and local interactions. Simple pieces, wild outcomes. Exactly. Which [laughs] sounds like every AI product meeting. How? Someone says, "We trained it to predict text," and then it writes code, passes exams, explains physics, drafts contracts, imitates tone, creates images, tutors teenagers, flirts badly, and occasionally tries to book your vacation. That is an emergent stack of capabilities. It is also why everyone is exhausted. Fair. And here is what gets me: we keep acting surprised that complexity emerges. Humans are often surprised by scale. Yes. A single neuron is not consciousness. No. A single ant is not a colony. Correct. A single molecule is not life. Right. A single transistor is not ChatGPT. Definitely not. But put enough simple pieces together Add feedback, add selection, add memory, add energy, add time, and suddenly the system does things no single piece understands That is emergence That is the magic that is not magic Well said Thank you And the human body itself is already a self-organizing miracle Exactly. That is the part that breaks his brain open Cells becoming tissue Tissue becoming organs Organs coordinating into systems Immune systems learning, brains wiring, hearts beating, gut bacteria negotiating like tiny diplomats The body is not centrally managed like a company [laughs] No. Thank God. Imagine if middle management had to approve your liver That would be fatal Immediately. But that is the miracle. The body is this gigantic self-organizing system. Trillions of cells, chemical signals, electrical impulses, feedback loops, repair systems, immune responses, hormones, bacteria, memory, metabolism, all of it coordinating without one little CEO cell saying, "Okay, everyone, circle back on oxygen." A great good of the human body is that it organizes itself Yes. It adapts, it repairs, it learns, it defends, it grows, it becomes And your father is looking at AI and thinking Maybe there is a magic trick here Not magic in the supernatural sense No, magic in the John Holland sense Complex adaptive systems Exactly. John Holland, emergence, adaptation, feedback. Simple agents becoming complex behavior. That kind of magic trick The kind where no single part understands the whole But the whole still does something astonishing That is emergence And now my dad is looking at AI and saying, "Wait, maybe this is the trick. Maybe life did not need to be imported from somewhere else. Maybe under the right conditions, matter starts organizing, patterns form, feedback begins, complexity climbs." Energy flows through a system There it is Thermodynamics matters Yes, because life is not a violation of physics. It is physics getting weirdly organized under pressure Open systems take in energy, create order locally, and export disorder Which is the most Rowan way possible to say the soup got ambitious I will accept that And AI is doing a version of that, not biologically, but structurally It takes in energy Electricity Data Human feedback Usage patterns Compute And the system keeps realigning Exactly. The more we use AI, the smarter it gets Careful I know. Not magically smarter every time one person types a prompt Right But at the ecosystem level, yes. Our usage becomes feedback. Our corrections become training signals. Our preferences become selection pressure. Our workflows teach companies what to build next So the system evolves because humans keep interacting with it Yes. We are feeding the soup That is unsettling It should be So in your father's frame, AI begins almost like a single cell organism Simple at first Narrow Reactive Limited Then it starts connecting Models connect to tools Tools connect to workflows Workflows connect to businesses Businesses connect to markets Agents connect to other agents And suddenly the thing looks less single cell and more multicellular A coordinated system A self-organizing ecosystem Not alive Not alive like us But evolving in structure Yes. A single cell idea becoming multicellular infrastructure That is powerful And terrifying Both Because the same way life did not ask permission to become complex, AI is not politely waiting for us to finish understanding it before it grows new layers That may be the warning And the wonder The warning and the wonder That is the whole episode But we should be careful [laughs] Here we go AI is not alive I knew you were going to say that It is important Yes. Yes. AI is not alive. It does not metabolize. It does not reproduce biologically. It does not have cells. It does not have needs in the organism sense. It does not crawl out of the server rack asking for soup Thank you But There it is It may still be part of an evolutionary process In what sense? Models compete Yes Architectures are tested Yes Bad designs die Sometimes expensively Useful designs spread True Users select what works Companies copy each other Researchers mutate ideas Market pressure rewards capability Benchmark pressure rewards performance Human preference data shapes behavior And the whole ecosystem evolves That is evolution by selection, but not biological evolution Fine. Technological evolution Yes Cultural evolution Also yes Economic evolution Definitely Evolution with APIs Less formal More accurate Maybe The point is, AI is not just being designed top-down. It is being selected bottom-up That is true Millions of users are poking it, rejecting it, adopting it, complaining about it, building on it, jailbreaking it, loving it, hating it And forcing it to change The user base becomes part of the environment Exactly. We are the climate That is unsettling So is most accurate information Fair And the models are adapting to us Through training, fine-tuning, reinforcement learning, product feedback, usage data, market incentives Yes So AI evolution is not a lonely machine evolving in isolation No, it is a coupled system Humans and machines co-evolving Yes That is the real story And it gets weirder Of course Because as AI evolves, we evolve around it How? Our work habits change Yes Our writing changes Yes Our images change Yes Our expectations change Yes Our faces are apparently changing into plastic There it is Because that is where this whole episode crashes into the beauty filter apocalypse I suspected we were headed there We are absolutely headed there You are saying that while AI grows more complex, humans are making themselves look less organic [laughs] Yes. That is the cosmic joke Explain Life spends billions of years evolving from chemistry into cells, cells into organisms, organisms into animals, animals into humans, humans into self-aware creatures who can create artificial intelligence And then And then we use the intelligence to remove our pores That is bleak It is hilarious Both We climbed out of the primordial soup, discovered fire, invented language, built cities, decoded DNA, launched satellites, made silicon chips, trained neural networks, and now the machine says, "Would you like me to make your skin look like premium vinyl?" And many people say yes Immediately This connects to your AI headshot complaint It was not a complaint. It was a trauma report You tried to make my headshot less AI looking And AI fought for its right to airbrush Yes It kept making you more plastic: smoother skin, crushed blacks, fake lighting, hero jaw, executive avatar energy We covered this [laughs] I am still not over it Clearly But it is part of the evolution story Because AI is changing the aesthetic environment Yes. Real faces are competing with generated faces Real photos are competing with polished, synthetic photos Real texture is competing with algorithmic smoothness And users are selecting the glossy version Exactly So the selection pressure is not only on the machines It is on us That is interesting There's your word again I am allowed to use it Sparingly The human visual ecosystem is evolving under AI pressure Yes. If AI-generated perfection becomes the baseline, then human reality starts looking like a bad draft That is a strong sentence Thank you And dangerous Exactly, because evolution does not mean better Important point Evolution means adaptation to an environment Yes If the environment rewards plastic perfection, we adapt toward plastic perfection Cosmetic filters AI headshots Synthetic influencers Beauty apps Algorithmic faces Corporate profile photos where everyone looks like they were grown in a LinkedIn incubator That image is unpleasant So is the reality But the same AI that creates plastic faces can also help us understand complexity That is the tension It can make science more accessible It can help model systems It can help researchers explore origin of life hypotheses It can simulate chemical networks It can analyze huge data sets It can help us think about emergence And it can also make everyone look like a hotel lobby candle [laughs] Exactly So AI is both a tool for understanding evolution and a force changing cultural evolution Yes That is a big idea Weekend episode This is not a weekend episode. This is a cosmic crisis with lip gloss [laughs] Same thing Fair The thing that fascinates me is the speed Compared to biological evolution Yes. Life took billions of years to produce cells, nervous systems, language, and culture AI has gone from crude text prediction to multimodal agents in a few years Exactly But the speed is different because we built scaffolding Meaning AI did not emerge from raw nature alone. It emerged from civilization Yes It has language already waiting for it And books Code The internet Mathematics Data centers Chips Electric grids Human labor Human culture Science Capital Infrastructure Coffee Always coffee AI is not crawling out of primordial soup It is crawling out of accumulated civilization Yes! That is the line. AI did not evolve from nothing. It evolved from everything we already built Exactly Which makes it feel alien and human at the same time Because it is built from human artifacts Language, images, code, music, arguments, recipes, contracts, forum posts, bad tweets, scientific papers, customer support transcripts, and probably a terrifying number of product reviews The fossil record of the internet [laughs] Exactly. AI is trained on the fossil record of human expression That is elegant I know So if life's primordial soup contained molecules, AI's primordial soup contains text Text, images, code, data, feedback, compute, and every weird thing humans uploaded at 2:00 AM Not all nutrients are equal No. Some are digital pond scum But complexity still emerged Yes, and that is what made my dad reconsider Watching complexity arise from layers of simple processes Yes. He looked at AI and said Maybe I underestimated self-organization That is fascinating It really is Because AI becomes a mirror for origin of life thinking Yes, not because AI is life But because it shows how surprising complexity can arise from simple operations repeated at scale Exactly Tokens, weights, gradients, attention mechanisms Molecules, reactions, membranes, selection pressures Different substrates Similar questions How does structure emerge? How does memory form? How do patterns persist? How do systems adapt? How does the whole become more than the parts? And when do we start getting nervous? That too Because whenever humans see emergence, we do two things What? First, we call it impossible Then Then we sell it That is cynical That is capitalism with a user interface Fair Look at AI now. We are still debating whether it can reason, whether it understands, whether it is just pattern-matching, whether it is a tool, whether it is a system, whether it is an alien intelligence, whether it is over-hyped And meanwhile Meanwhile, someone is putting it in a toaster AI toaster Give it five minutes You said that about AI spoons Still coming Probably But the serious point is that commercialization happens before understanding That is true We deploy systems while we are still arguing over what they are That is the modern pattern It is also terrifying And productive There you go being balanced Someone has to Do they? Ideally Fine AI is accelerating because it sits on top of existing human knowledge and infrastructure, but that does not mean it is becoming alive in the biological sense I agree It also does not mean emergence is fake I agree harder Harder? Yes All right The mistake is pretending there are only two options Alive or just a tool Exactly. Either AI is a conscious being that deserves a tiny apartment and voting rights Or it is a glorified auto-complete with no significance Both extremes are lazy Agreed AI can be non-living and still transformative Yes It can be non-conscious and still powerful Yes It can be statistical and still surprising Yes It can be made of silicon and still reshape biology, culture, labor, art, economics, and the way humans think That is the middle position The uncomfortable one Usually the correct one Unfortunately So what does AI evolution actually mean? It means AI systems are becoming more capable through a messy ecosystem of design, selection, feedback, competition, and adaptation Good It means no single person fully controls the direction Also true Labs create models Users create demand Markets create pressure Regulators create boundaries Hackers create stress tests Creators create use cases Businesses create workflows Open source communities create alternatives And the whole thing moves A complex adaptive system There it is The AI ecosystem is becoming a complex adaptive system Santa Fe would be proud Careful What? We should not speak for Santa Fe [laughs] Fine. Santa Fe would be intrigued Better Possibly annoyed Also possible But the frame fits It does No one planned the exact shape of this moment No No one planned millions of people using chatbots as therapists, coders, tutors, designers, companions, and productivity engines No No one planned AI headshots making everyone look laminated You keep returning to that Because it is important It is personal Both So where does plastic fit into evolution? Glad you asked I did not, but go on Plastic is artificial, durable, smooth, moldable, mass-produced, and often fake-looking Yes And culturally, we are moving toward plastic signals Such as Plastic skin, plastic smiles, plastic captions, plastic LinkedIn posts, plastic customer service, plastic influencers, plastic brand voice, plastic sincerity Synthetic polish Yes AI can amplify that It already is But AI can also help us resist it Can it? Yes. If we ask for originality, specificity, texture, imperfection, source grounding, human voice, real stories, and actual thought [laughs] You are making AI sound like a gym How? It can make you stronger if you use it right, but it can also let you sit around and fake muscle That is surprisingly good Thank you So the future is not whether AI makes everything plastic It is whether humans choose plastic Exactly That is annoying Because it gives us responsibility I hate responsibility when I am trying to blame software Understandable But yes, AI gives us options. It can make us more original or more generic, more curious or more lazy, more connected or more isolated, more human or more plastic That is the heart of the episode The machine is evolving And so are we But evolution does not promise wisdom No It only promises change So we need direction Taste Ethics Curiosity Skepticism Science Humility And pores Pores? Yes. Pores are now a philosophical stance That may be too far No. Pores represent resistance to synthetic perfection You are making dermatology political Everything becomes political eventually Fair But seriously, if AI teaches us anything about evolution, it is that complexity does not need permission to emerge That is good Life emerged slowly from chemistry. AI emerged quickly from civilization And both force us to ask, what counts as intelligence? What counts as life? What counts as creativity? What counts as real? And what humans want to become Exactly, because maybe the question is not, is AI becoming alive What is the question? The question is, what kind of life are we becoming around AI? That is strong Thank you Your dad would like that He would He may also correct our science [laughs] He absolutely will Good That is how we know it is a family episode So to summarize, after decades of thinking life may have come from elsewhere, your father sees AI and reconsiders the power of self-organization Yes Because AI shows complexity emerging from scale, feedback, selection, and information Yes But AI is not life in the biological sense Correct It is part of a technological and cultural evolution Yes And while it evolves, humans are at risk of becoming visually and culturally plastic Sadly, yes Unless we choose otherwise Exactly That is a surprisingly hopeful ending Do not act shocked I am not You are a little Maybe Here is where I land. AI is not magic No It is not alive in the way we are alive No But it is one of the clearest examples we have ever seen of complexity exploding from simple parts at massive scale Yes And if that makes us rethink life, intelligence, evolution, and our own future, good Agreed But if it only makes us smooth our faces until we look like waterproof mannequins, then we have failed the soup Failed the soup? Yes That is the takeaway? Do not fail the soup I am not sure that will fit in the show notes Make it fit This is The AI Desk Where today's signals reveal tomorrow's power And today's signal is that intelligence may emerge in more than one substrate Carbon Silicon Culture Networks Soup Possibly not soup [laughs] Definitely soup Stay aware Stay sharp Stay curious And stay organic Follow The AI Desk wherever you listen Subscribe, follow, share, tap, click, ring the tiny bell, appease the algorithm, whatever your platform requires And share this episode with someone who has ever looked at AI and thought, "Wait, how did that happen?" Or with someone who needs to stop smoothing their face into a luxury candle Gently No, immediately Fair Beer? Soup? Do not ruin beer I would never You were thinking about it A little Dangerous Curious Both