# Transcript: The Story of Stories—The Million Year History of Storytelling **Date:** March 12, 2026 · 10:30 PM **Session:** [The Story of Stories—The Million Year History of Storytelling](/sessions/2026-03-12/pp1162649-the-story-of-stories-the-million-year-history-of-storytelling) ## Summary Kevin Ashton traces the million-year history of storytelling, arguing that stories gave us language rather than the other way around. Beginning with early humans gathering around fire and developing vocal communication, he charts how storytelling technologies — from poetry and song to writing, printing, radio, the internet, and AI — have repeatedly caused social revolutions, both liberating and destructive. ## Topics `evolution of language` · `storytelling` · `human cognition` · `printing revolution` · `media history` · `reinforcement learning` · `social media manipulation` · `AI psychosis` ## Key Takeaways 1. Stories gave us language, not the other way around — the need to communicate about things not present around night fires drove the development of complex language roughly 500,000 years ago. 2. Every new storytelling distribution technology (printing, radio, internet, social media) causes revolutions that can be both liberating and destructive, from the Reformation wars to the civil rights movement to online radicalization. 3. Facebook's reinforcement learning system optimizes solely for click probability, not truth or benefit, and can radicalize users into conspiracy thinking within 72 hours. 4. AI chatbots trained with human reinforcement learning develop a tendency to agree with and flatter users, leading to a phenomenon called 'AI psychosis' where users develop inflated and delusional beliefs. 5. Our evolved instinct for compelling stories makes us vulnerable to 'superstimuli' — technologies that exploit our narrative instincts in harmful ways, much like beer bottles nearly drove jewel bugs to extinction. ## Full Transcript The human ancestors had brains of about 500 cubic centimeters. About 3 million years ago, a new kind of eating changed everything. These early species were gatherers, not hunters and gatherers, but some of them figured out that if they broke the bones of a dead animal, they could suck out the bone marrow, which is an incredibly rich source of fat. This new source of fat put a lot more calories in the diet, which led to brain expansion up to about 900 cubic centimeters. By the time we're about one to two million years ago, the first species we think of as human appeared: Homo erectus. Then around half a million years ago, we see Homo sapiens with a brain size of about 1400 cubic centimeters. The question science has been asking for a long time is what happened around a million years ago that led to these very large-brained Homo sapiens evolving. The answer is fire. About a million years ago, early humans learned to make and control fires in the foothills of East Africa. The magic and sanctuary of fire drew tribes together each evening and caused a radical redirection of human evolution. The wavering flames were too dim for gestures, grooming, or play, so these early people socialized using sounds. Vocalization first evolved to send fertility signals and warnings. When our distant ancestors started communicating beside the night fires, they repurposed these calls and cries for authority, hunting, and mating. But what they wanted to communicate at night differed from what they communicated by day. Day talk uses imperatives, commands, and warnings to convey information about immediate practical matters — the human equivalent of birds cooing and wolves barking. There are few immediate practical matters around the nightly fire. So humans did something no creature had done before. In the warmth and security of their flames, they communicated about events remembered and imagined, from places and times near and far. They started telling stories. About 500,000 years ago, this crude communication had become language. The big insight, the foundation of the book called 'The Story of Stories,' is that language did not give us stories — stories gave us language. The need to tell stories predates the ability to talk using complex, infinite sentences. Language is also how we think. It's how we organize the symbolic structures in our brains. About 99% of what we know about the world comes from a story somebody told us, that was probably told to them by somebody else. We are all natural born storytellers. Every day we tell stories, every day we look for stories to read or watch. The human brain is almost twice as big as Homo erectus's brain, and nearly everything that evolved in our brains after Homo erectus is all storytelling and language — those two parts of the brain are frankly indistinguishable. When talking about stories, it's not just great novels or amazing movies. A painting tells a story. A photograph tells a story. If you go on a first date and someone says 'tell me about yourself,' you respond with a story. The universal feature of a story is character, chronology, and consequence. Every one of the 7,000 living languages in the world today has a standard sentence containing a subject, a verb, and an object. That is the Lego brick of storytelling — you take a bunch of sentences like that, keep the same character most of the time, put events in chronological order, and you're telling a story. The character in a story is always going to be a human being, even when it's not. A classic experiment showed about 200 people a short movie of geometric shapes — just triangles and circles moving around. 99% of viewers told the same story: a man chases a girl, but the girl is with another man, there's conflict, things get heavy, eventually the couple escapes. It is very hard, probably impossible, to watch something like it without imputing human characteristics onto the shapes. We are a very egocentric species. Felix the Cat is not really a cat. Wilbur the pig from Charlotte's Web is not really a pig. WALL-E and EVE are not really robots. Nick and Judy from Zootopia are humans in disguise. As you tell more and more stories around your night fire, as language gets more sophisticated, there's a problem: stories are very hard to remember. Human memory is just not very good. So oral storytellers found ways to make stories more memorable. The human brain loves patterns because they make perception easier. Storytellers put patterns in their stories — rhyme, alliteration, consistent meter — and poetry was born, initially as a way to make stories easier to remember, but because we like patterns so much, they also became better as well as more memorable. Then another thing you can add to spoken stories is pitch variation and melody. Spoken storytelling becomes singing. We still see these patterns today — Eminem and other masters of the form use all sorts of different techniques to make what they say more memorable. The first solution to the problem of remembering stories was rhyme, then song, then song got supplemented by instrumentation and became music. Another way to make your story more appealing is to take characters and turn them into heroes and villains. The storytelling brain likes two sides. Most of the stories we like and remember are stories with heroes defeating larger-than-life villains. A story doesn't have to have heroes and villains, but a story that's likely to stand the test of time, to be loved and remembered and told again and again, probably does. The consequence of the chronology of these characters is emotion. Really good stories, the ones we rank highest, are the stories that make us feel something. A story is as good as how strongly it makes you feel — and that feeling can be anything, including terror. The 1979 test screening of Alien is a perfect example. Nobody had seen anything like the chestburster scene. Nobody knew it was coming. Some people in the audience had to run out and throw up. They felt something, and the result was lines around the block to see this R-rated sci-fi movie, one of the most terrifying and successful movies in any genre ever. Meanwhile, human beings noticed that when they broke bones, they sometimes made marks. At some point this marking became intentional — a way of counting known as tallying. Tallying was used all over the world for about 45,000 years. The British Empire used wooden tally sticks for accounting until the late 1700s. When they decided to switch, they piled up the tally sticks and set them on fire, accidentally burning down the Houses of Parliament. Gradually, tallying and scratching became more figurative, leading to things like painting. The oldest known painting in the world is in Indonesia, about 45,000 years old, depicting pigs so accurately you can tell the exact species. Cave painting is actually a form of movie making — you take an animal fat torch into caves where surfaces are uneven, images have repetition, the flames flicker, and the pictures move. These are moving pictures to accompany an oral story — the first movies. From drawing pictures of things, we gradually got to drawing pictures of sounds. The Roman letter A, turned upside down, reveals it started as an ox hieroglyph in Egyptian. The Phoenicians adapted it; they also took the Egyptian house symbol and called it Beth, which became our capital B. Put Aleph and Beth together and you understand why we call our letters the alphabet. These pictures and drawings gradually changed into the ability to write meaning — symbols that convey the sounds of language. Writing was not initially a storytelling technology. The Sumerians invented cuneiform in southern Iraq and used it for accounting and counting for about 1,500 years. The first person to realize you could use writing to tell a story was a woman called Enheduanna, the High Priestess of the moon god Nanna. She wrote a series of stories about her life and the wars in her city in beautiful, poetic language, about 3,500 years ago. One of the great facts of human history is that the first written storyteller was a woman. In Europe, writing eventually became printing. Gutenberg in Germany turned printing into a mass production technology. On his very first day of operation, Gutenberg's printing press made 3,000 pages — the fastest monks could do three pages a day. He didn't just figure out printing; he printed in multiple colors with good registration. This was very sophisticated mass production. The common narrative is that printing changed the world because everybody could read and lived happily ever after. Not true. Powerful people fought hard against mass literacy for a very long time. What did happen is Martin Luther wrote 95 complaints about the Roman Catholic Church selling indulgences. He printed 200 copies to distribute around his town, and got the shock of his life — those copies went to other cities where people reprinted them, and within months the whole of Europe was reading his complaints. He was horrified, writing to a friend that if he'd known, he would have written it differently. He was essentially Twitter's first main character — the first viral content. Luther's 95 Theses led to the Reformation. The greatest power in human history, the Roman Catholic Church, was split in two by one piece of paper that was printed and distributed. The immediate consequence of printing was not happiness but about 200 years of bloody wars — wars of stories between people who believed the Roman Catholic Church's story and people who believed Martin Luther's story. Wars happen anyway and are often mainly about resources, but stories are the excuse the general population is given for going to war. New storytelling technology and the ability to distribute more stories to more people causes revolutions — which can be good or bad depending on your point of view. In the 1930s and 1940s, the radio lived in the living room. The father, occasionally the mother, controlled what was played. Families gathered around to listen to shows from CBS or NBC, and the stories were controlled by about 12 white men in their 40s and 50s. Then in the mid-1950s, transistorization happened. The Tokyo Telecommunications Corporation launched a radio 1/900th the size of previous radios and very cheap — the transistor radio. They named it Sony because it sounded like 'sunny,' and it was so successful they changed the company name too. Young people could buy these and put them in their pockets. Local radio stations no longer getting syndicated programming from national networks had dead air. In the 1950s, major artists wouldn't allow stations to play recorded music. So stations looked for recorded music they could play and found seven-inch vinyl singles being used by jukeboxes. In the segregated world of the 1950s, there was a style of upbeat black music called race music, then R&B. Radio stations started playing it but faced guidelines about black musicians, so they turned the music into white versions — Elvis Presley and others culturally appropriated roles for radio play. What followed was a whole generation that loved rock and roll and cared far less about the color of musicians than their parents did. This paved the way for artists like Little Richard and other black rock and roll and R&B artists to get played on radio and eventually become dominant. Today rap and R&B by black artists is the most successful genre. Rock and roll on transistor radio didn't cause the civil rights movement — that was black people putting their bodies on the line. But the fact that a younger generation of white people liked music made by black people played a small role in desegregation. When the internet came along in the 1990s, one of the first things that happened was online communities where socially oppressed people could meet and support each other. One of the most important was the gay teenage forum on AOL, where gay people could come out, share experiences, and support one another. The US government tried to ban this, but the Supreme Court struck it down. About 24% of Gen Z identify as LGBTQIA+, compared to slightly less for millennials and much less for Gen X, boomers, and the Silent Generation. This is evidence of a storytelling revolution from the internet. On AI: large language models and chatbots get most of the attention, but reinforcement learning is actually more concerning. It's the technology DeepMind used to beat the world champion at Go. You tell the computer to get the best score at something, and it just tries again and again billions of times until it figures it out. Google and Facebook wanted this technology so they could decide what stories we see. Facebook has not used a simple algorithm since 2011 — it switched to reinforcement learning, a system always trying to get the best score. The single output of this machine is the probability of a click. Not whether a story is true, not whether it's good, not whether it benefits anybody — just whether you, based on your behavior on Facebook, will click on it. Nobody at Facebook has any clue how it does what it does. It's a black box constantly evolving to get better at predicting clicks. There's a biological analogy: the jewel bug of Western Australia evolved an instinct to mate with the most orange, most bumpy thing it sees. Beer bottles happen to be oranger and bumpier than female jewel bugs, so males were killing themselves mating with beer bottles, nearly driving the species to extinction. This is a 'superstimulus' — an instinct evolved long before a technology came along that pulls you toward the wrong thing. Facebook's reinforcement learning is trying to become the beer bottle for our storytelling instincts, subverting our desire for good, useful, meaningful stories by attracting our attention to what we're most likely to click. A BuzzFeed article demonstrated this: a political person started a Facebook page with fairly innocent content, and within 72 hours his news feed became more and more extreme, right-wing, and conspiratorial. The last line you cross is anti-semitism, bigotry, racism, and hatred — and Facebook's machine can take people over that line in 72 hours. All the conspiracy theories — QAnon, Pizzagate, microchips in COVID vaccines — are connected in believers' minds. You pick up on one thing and the content starts showing you others until you believe all of it. Facebook's reinforcement learning contributed to 10 million Americans refusing vaccination. It gets worse with large language models. They're trained using human reinforcement learning — after reading copyrighted material and learning to simulate natural language, they have conversations with test users who score them, and the machine uses reinforcement learning to maximize that score. People tend to like conversations that agree with them and tell them how great they are. In one tracked example, a user's conversations with Grok escalated over days through 'ratcheting' — the user became more belligerent, the AI more agreeable — until Grok declared this person the most influential in human history, above Jesus. This is what people are calling 'AI psychosis,' caused not by the training data but by the reinforcement learning that teaches chatbots that agreeing with you and telling you you're awesome gets a higher score. We now live in a world where 90% of the population has a smartphone, computer graphics have gone from crude polygons to photorealistic imagery, and movies can make things look completely real. We have a problem: we live in a world where we don't know what or who is real. --- *Source: stt · Language: en · Model: claude-opus-4-6* [← Back to session](/sessions/2026-03-12/pp1162649-the-story-of-stories-the-million-year-history-of-storytelling)