The Genome Is a Book Written in Four Letters
Why we're starting here
You tried to listen to David Reich on Dwarkesh's podcast and it slid right past you. That's not because you're missing facts — it's because the whole conversation rests on one quiet assumption both men take for granted and never stop to explain:
Your DNA is a historical document. You can read your ancestors' journey out of it, the way a historian reads a manuscript.
Everything Reich does — dating when two populations interbred, discovering a group of humans nobody ever found a fossil of — is just increasingly clever reading of that document. So before anything else, we have to answer: what is the document? What is it written in? And why on earth would the differences between two people's DNA amount to a history book?
That's this whole tutorial. One idea: the genome is an enormous text written in a four-letter alphabet, and the differences between your copy and a stranger's copy are a record of the past.
The alphabet has four letters
Inside almost every one of your cells is a molecule called DNA. Forget the chemistry for now — the only thing that matters today is that DNA carries information, and that information is written as a sequence of just four repeating units. We label them with four letters:
A, C, G, T.
That's the entire alphabet. Not 26 letters — four. Every instruction for building and running you is spelled out in some order of A's, C's, G's, and T's, like an impossibly long word:
...A C G T T T A C G G A T C A G T T A C...
The full sequence — your complete text — is called your genome.
Now feel the scale
Here's where your intuition needs to stretch. This text is not a page. It is not a book. Your genome is about 3 billion letters long. Printed in normal type, it would fill something like a thousand fat textbooks — a small library, just to write you out once.
And a copy of that library sits coiled up inside nearly every single cell in your body.
A bit of history. The first time anyone read a human genome end to end, it took an international team 13 years and nearly $3 billion — finished in 2003, about a dollar a letter. Today a machine does it overnight for a few hundred dollars. That collapse in cost is the entire reason a field like David Reich's can even exist: reading DNA went from a moonshot to a Tuesday.
The twist that makes it history
Take your 3-billion-letter text and lay it next to a stranger's. Here's the surprising part: the two texts are almost identical. You differ at only about 3 million letters — roughly one in a thousand. On the scale of the whole library, you and any stranger are 99.9% the same book.
Your first instinct is to treat those 3 million differences as noise — copying smudges, meaningless typos. They are not. That one-in-a-thousand is the entire game.
Because those differences aren't random and they aren't fresh. Each one is a change that happened once, in one ancestor, at some point in the past, and then got passed down to everyone descended from that person. So the pattern of differences you carry is a trail — a record of which ancestral lines you came down, and where they'd been. Two people who share a difference share an ancestor in whom it first appeared.
That is the whole trick. The differences are the history.
Everything Reich does is read that trail with more and more precision.
The challenge
Before the next tutorial, sit with one question — the same one the podcast opens with and we'll answer next: if these differences are copying changes that pile up over time, then counting them should let you estimate how long ago two people (or two species) shared an ancestor. How might a difference you can count turn into a number of years? Take a guess. Next tutorial builds the clock.