
The Return of the Rhetor
There are forty-one cinemas in the world that can project Christopher Nolan’s The Odyssey the way he intended, namely, on 70mm IMAX film. Each print is eleven miles long and weighs about as much as a motorcycle. Some pe…
以下正文同步自 Hacker News 最新,版权归原站所有,已转换为易读排版。
Ivan Yamshchikov
Nov 04, 2026
6
1
Share
There are forty-one cinemas in the world that can project Christopher Nolan’s The Odyssey the way he intended, namely, on 70mm IMAX film. Each print is eleven miles long and weighs about as much as a motorcycle. Some people bought tickets to those forty-one halls a year in advance. It was a sold out. The cinemas had to add screenings at midnight and at three in the morning. Those were sold out too.
The film is a huge hit. It opened to $264 million worldwide and passed $600 million within two weeks. So far it seems to be Nolan’s biggest result ever. However, the three-in-the-morning crowd is the detail I cannot let go of. It’s pretty clear the people were there not for a story. The story is three thousand years old; some of them read it in school. Yet, somehow they wanted to be in a particular room, with other people, at the moment the thing was told the way storyteller intended.
Strugatsky coined a phrase that I love: “people who crave oddities”. In their world those were the people who move the history forward towards civilizational progress.
Craving an IMAX screening of Odyssey at 3 in the morning is an odd thing to want in 2026. It is also, I think, the most important cultural signal of the year or may be a decade to come.
For five thousand years, a well written page cost somebody hours. Now it costs a fraction of a cent. Economists know what happens to a good when its marginal cost goes to zero. It becomes ubiquitous and cheap.
The inflation of text is everywhere: cover letters, LinkedIn posts, product descriptions, executive summaries, you name it. Texts written by nobody for nobody to read. Amazon capped the number of books an author may publish per day at three. Let that sink in.
So here is the question. Is it a coincidence that the story that pulled people out of bed at three in the morning is one of the oldest ones we have?
Homer, whoever he was, did not write The Odyssey. He performed it, or rather generations of singers performed it before someone had it written down. In the 1930s Milman Parry went to Yugoslavia with his team to listen to illiterate bards who could still sing epics of Homeric length. This work created Milman Parry Collection of Oral Literature that remains in Harvard library to this day. They found that no two performances were the same. The singer was playing a glass beads game with the crowd. He had formulas and scenes up his sleeve, and he assembled them live, watching as the room co-created the story with their reactions. In practical terms that meant that a bored room gets a shorter catalogue of ships while an attentive one gets Odysseus lying to his own father in an extra hundred lines.
Nikola Vujnović, pension “Viktorija,” Dubrovnik. Photo by Milman Parry, between 1933 and 1935. View Details. Milman Parry Collection of Oral Literature.
I know a lot of people who identify with the book. They enjoy reading and writing. They self-describe as book people. Who would have thought that academics might have natural affection towards written word! They also tend to look down on the oral tradition. Mostly because they see speech as a worse way of transmitting a text. I would argue that a spoken word should be seen as a different kind of object.
A written text is to the spoken word what taxidermy is to a real animal. Spoken word only makes sense as it lives, pulsates and dies. The yawns and laughs of the audience co-create the habitat of the spoken word. The storyteller reassembles the story live weaving the audience signal into its fabric. The story retold through thousand evenings becomes an apex predator of human attention.
Oral narrative tradition is as close to reinforcement learning from human feedback, as it gets. The difference is that the oral narrator applies RLHF to the story, while in modern ML we run it on the model storyteller.
The Hebrew Bible and the Gospels were spoken, chanted and argued over for generations before they were written down. The Quran is the most explicit case for oral narrative tradition: by its own account the Prophet could not write. God spoke through the Prophet while the listeners remembered and recited what they heard. At least a generation later, somebody collected the recitations into a book. The word itself means “the recitation.” You can see the most fundamental texts of modern civilisation as the survivors of a very long selection process. And even though we tend to see those things as texts, their origin is profoundly different: it is spoken and may be sometimes even performative. The difference between a broadway hit and a failure is simple: does a room full of people want to hear this thing again? That is a brutal filter. Almost nothing passes it. What passes becomes Les Miserables, Hamiltons and Lion Kings.
Ok, so some texts could be oral narratives in disguise. Can we find some more obvious examples of narrative tradition? You bet!
Stand-up comedy is the purest case of oral narrative form that plays ever growing role in modern societies. A comedian builds an hour the way Parry’s singers built an epic: in small rooms, night after night, keeping the lines that land and cutting the ones that bomb. The one hour special is the “taxidermy” of the exceptionally good evening closer to the end of the year, not the living organism of a daily concert. Every hour you see on your streaming service bakes in a year of edits performed in front of a few thousand drunk strangers.
The long-form podcast is another form of non-written word. Three hours of two people talking should be unlistenable. The form by design defies every rule of written composition. However, it is the dominant format of the decade. Why? First, it is visibly unscripted. The bones of written structure are completely removed. The listener enjoys juicy boneless fillet of the thought process. Dialog comes from Greek “through speech/reason”. The intellectual joy you get when the speaker picks up a thought where the interviewer left or then the host interjects with a tangent is very similar to the feeling you get on a good live performance.
The talking-head clip on Instagram and TikTok is the degenerate form of the same thing. One face, one take, straight to camera. If it’s scripted, people feel it and scroll further. If it’s raw they watch till the end and copy the link to send it to a friend or two.
Those content forms are profoundly oral and they happen to command the most important assets of the attention economy: eyeballs and ears.
The usual explanation stops at “people crave authenticity.” I think the mechanism is more specific than that: people try reinventing an old almost forgotten intellectual tradition as text-based structures collapse under the inflation of the written word.
Homer’s feedback loop was slow. A singer might perform for twenty people, then forty, then a wedding party of a hundred. The signal was rich, because he could see every face, but it took centuries to converge. The modern “bard” runs very similar loop on a different hardware budget. A comedian tests a line in front of three hundred people on Tuesday and (if they are lucky) in front of three million in the end of the year. A podcaster sees the retention curve of yesterday’s episode before recording today’s. The signal is thinner per listener, a like is not a laugh, but it comes from millions, and it comes within hours. Technically, similar feedback loop could be run on texts, but human attention does not allow such luxury. We rarely re-read texts. And the text we re-read are already exceptional. Oral tradition allows to sample feedback multiple times from multiple audiences. Anyone who has trained a model knows what that does. Faster, stronger reinforcement does not give you the same result sooner. It gives you a qualitatively different result. The oral tradition is back, and it is back on steroids.
The models are in a loop too. Every thumbs-up, every regenerated answer, every conversation that ends early is a signal, and it comes from hundreds of millions of people. On paper that is the richest feedback loop ever built. In practice it is a loop with the bandwidth of a telegraph. A thumbs-up is one bit. The singer in front of forty people sees the room lean in, hears the breath before the laugh, notices who stopped listening, and all of that arrives at once, through a channel evolution spent a few million years tuning. Until a model can stand in a room and feel it turn, it is running the oral loop on a written channel. This helps with many tasks, but it converges on what the written channel rewards. So, for sometime the models will drive text inflation with minimal spillover to the spoken word.
Which means we should expect the world around us will keep on changing rapidly. For several centuries the path to influence ran through the page: the pamphlet, the book, the op-ed, the paper. Rhetoric, the art of saying it to a room, was demoted to a party trick. The rhetor returns. The twenty-five-year-olds who are winning attention right now are not better writers than their parents. They are far better speakers. We have to reinvent oral tradition that formed the foundation of modern civilisation.
The Athens we like to remember was a city of writers. The Athens that really existed was a city of talkers. Policy was made by whoever could hold the Assembly; lawsuits were won by whoever could hold a jury; philosophy was more a market performance than a disciplined study. The written word was the inferior medium. Plato puts the argument against writing into Socrates’s mouth: a text keeps saying the same thing no matter what you ask, no matter who asks.
I do not think we will see a return to our ancient past. The loop is too fast, the reach too large, the feedback too cheap. But I think it is the end of a parenthesis: the five centuries between Gutenberg and the transformer in which being good on paper was the only way to be taken seriously. The parenthesis is closing. Cinemas, at three in the morning, full of people who wanted to hear the oldest story retold the way the director wanted, stand as a testimony to that. The oral tradition is not dominating our culture yet. It returns at a new scale. And the author, the central figure of the written age, packs his things on the way out.
It seems to me, when the viewers were leaving the cinema after the 3 a.m. Odyssey screenings this summer they were in the hour between dog and wolf in more senses than they could have imagined.
正文由 探How 从来源站点 RSS 同步,内容未经改写;遇到排版缺失或需要图片、视频时请以原文为准。