I Vibe Coded a Ranking of World History in a Day. Santa Claus Outranks Einstein.

A ranking of five thousand years of human influence used to take a room of historians, economists and engineers months to build. I built one on a Friday, with a few dollars of tokens from a plan that was going to reset the next morning anyway.

It's live at shadowyrealm.com. As I write, it ranks 614 people, organizations and events by how much they changed the world. Ancient Egypt is first. The Maya are second, Hinduism third, the Catholic Church fourth. Jesus is ninth. Santa Claus is 248th and Albert Einstein is 397th, 149 places behind him. Harry Potter is 590th.

I vibe coded it, and I'm using the word the way it's used against me now. Vibe coding started as a description and turned into an insult: you tell a model what you want, it writes the code, and you steer by looking at what comes back. I plead guilty. Claude Opus 5.5 wrote the version that's on the site. It runs on Cloudflare and Supabase, the stack I already pay for, on a domain I've owned for decades. I named shadowyrealm.com after a paint color for my house, and over the years I've pointed it at one thing after another.

The stone beast glaring from the home page is an AI-generated Vaal, the god from the Star Trek original series episode "The Apple."

Standing on Wikipedia

Wikipedia supplies the evidence. It's the best single curated source we have, and the model that wrote the code was trained on it too. Each person, organization or event is split into periods, a reign or a ministry or a war, and each period gets a few paragraphs of Wikipedia text.

Clef, Cloudflare's classification model, reads that text and scores the period on ten numbers: a benefit and a harm in each of five dimensions, which are scientific, economic, human rights, health and cultural. Benefit and harm are never netted against each other. Then Postgres passes credit backward along the links between entities, so the Crusades credit the Church and the Church credits Jesus. The result is 614 entities, 712 periods and 289 links, drawn as a river that widens over five thousand years.

I'm not a historian

My training in history is a few books, some biographies, and a course called Vigilantes in American Society, which I took because a diet of engineering courses makes Geordie a dull boy.

So I can't tell you whether the ranking is right, and neither can anyone else, because ranking influence has no right answer. I told the model to stop writing disclaimers. This is entertainment. The ranking is wrong from a Buddhist point of view and from a Somali point of view, and nothing depends on it.

What I could do was argue. The first version scored each figure on one bounded scale and let influence decay with time. I told it the scale had to be unbounded, and that decay runs backward for the people who matter most: Jesus had little influence while he lived, and it grew for centuries after the early church started promoting him. So time became a run of periods, each scored on its own. Then the score collapsed to a single net figure, and Jesus, who is credited with cathedrals and with conquests, came out as consequential as my uncle Bob. That's why there are ten numbers.

Ten years of news, five thousand years of history

I built this because of the last ten years. Watching the political landscape change made me want perspective, a way to see where a disruption sits beside everything else that has disrupted us.

The ranking supplies it. The Bantu expansion is fifth and the invention of cuneiform is seventh. The events that fill the news land low: World War II is 172nd, and September 11 is 550th of 614. My own intuition said Jesus beats Shakespeare, and he does, 9th to 354th. My intuition also said Jesus would be first. Ancient Egypt took that, and I left it there.

What it cost, and what it's worth

Five years ago this was possible and ruinous. It needed a historian to choose the entities, an economist to argue the dimensions, a data engineer to model the graph, a designer to draw the river and a developer to ship it. A room of experts and months.

The site earns nothing and never will. It has no customers, which is why the room never existed: no one would pay for it. It's the second thing I've built this way that earns nothing. Adam Carolla's two questions, which I wrote about in Money or Happy, are whether it makes me money and whether it makes me happy. This one makes no money and a good deal of happy.

The economic value of an activity and its social value have come apart, and the AI age is full of cases like this one. The AI-narrated videos about the Phoenician empire that people play to fall asleep are the same kind of thing: cheap to make, free to watch, and worth something to the person awake at 2 a.m. This ranking belongs on that shelf. When production gets this cheap, what comes out is a hobby.

A short time ago, ranking history across five thousand years was a serious academic exercise, with committees and footnotes. Now it's something you do on a Friday for the pleasure of it, the way people paint or trace their family trees. I find that fascinating and a little vertiginous, and I don't have the training to say what it means for the historians.

The question of who shaped the world used to cost a career to ask properly. It cost me a Friday.