The Codex: Where Books Began

The word codex comes from the Latin caudex, meaning “trunk of a tree” or “block.” In ancient Rome, a codex referred to wooden tablets bound together — early notebooks literally made from blocks of wood. By the 1st century CE, Romans began writing on parchment pages bound between covers, replacing the scroll format that had dominated for millennia.

The codex revolutionized information storage. Unlike scrolls, which required linear unwinding, codices allowed readers to flip directly to any page — random access. Both sides of each sheet could be used, doubling the storage density. By the 4th century, the codex had rendered scrolls obsolete. Almost every book you’ve ever read is technically a codex.

The term also took on a legal meaning: a codex became a collection of authoritative laws. The Codex Justinianus and Codex Theodosianus were landmark legal compilations. This sense — a curated, authoritative body of knowledge — would later echo in computing terminology.

···

Code: Rules Written in Symbols

The English word code entered through Old French from the same Latin root, codex. In medieval France, code referred specifically to a systematic collection of laws — essentially a bound legal text.

From there, the meaning expanded. A code became any system of rules, symbols, or conventions for communication. Morse code translated dots and dashes into letters. Genetic code translates sequences of nucleotides into proteins. Legal codes govern behavior. And in computing, code translates human intention into machine instructions.

The evolution moved from physical books to abstract systems — from the container to the content. Today, when we say “source code,” we’re invoking a lineage stretching back to Roman law books and wooden tablets.
···

Claude: From Roman Emperors to Information Theory

Claude began as a French masculine name derived from the Latin Claudius. The name likely traces to claudus, meaning “lame” or “crippled” — a Roman family name (gens Claudia) that produced several emperors, most notably Claudius I, who ruled from 41 to 54 CE.

So how did a Roman emperor give his name to a 21st-century AI assistant?

Anthropic’s Claude AI is named as a tribute to Claude Shannon, the mathematician and engineer called the “father of information theory.” In 1948, Shannon published “A Mathematical Theory of Communication,” laying the groundwork for everything from digital circuits to compression algorithms to artificial intelligence. His work defined the mathematical principles underlying how information is transmitted, compressed, and processed — the very foundation on which modern AI operates.

The Claude lineage
~50 CE
Claudius I rules Rome. The gens Claudia name enters history.
1916
Claude Shannon is born in Petoskey, Michigan.
1948
Shannon publishes “A Mathematical Theory of Communication” — the foundation of information theory.
2023
Anthropic launches Claude AI, named in Shannon’s honor.
Mathematical formulas and geometric diagrams — the language of information theory
The mathematics underpinning modern AI
···

OpenAI Codex: The AI That Writes Code

OpenAI Codex is an AI model designed to translate natural language into code. Announced in 2021, it powers GitHub Copilot, the AI pair programmer that suggests code as developers type. Codex understands context — comments, function names, surrounding code — and generates appropriate snippets in dozens of programming languages.

The name “Codex” here is deliberate. Like the ancient codex that organized and made knowledge accessible, Codex organizes programming knowledge and makes it retrievable through conversation. It democratizes coding by lowering the barrier to entry: describe what you want in plain English, and Codex produces the code.

On February 5th, 2026, OpenAI released GPT-5.3 Codex — the first model in history that was instrumental in creating itself. According to their technical documentation: “GPT-5.3-Codex is our first model that was instrumental in creating itself. The Codex team used early versions to debug its own training, manage its own deployment, and diagnose test results and evaluations.” This recursive loop — AI building AI — closes a circle dating back to those first Roman tablets. The codex that once held laws now helps write the systems that write themselves.

···

Claude Code: Claude at the Command Line

Claude Code, launched by Anthropic in 2025, is a command-line tool that brings Claude AI into the developer’s terminal. Unlike Codex (which focuses on code generation), Claude Code is a full development environment. It can read files, write code, execute shell commands, run tests, and manage git operations — all through natural language conversation.

The name reflects its function: it’s Claude (the AI) operating as code — executing, building, running. Where Codex generates snippets, Claude Code orchestrates workflows.

···

From Papyrus to Terminals

These three words — codex, code, and Claude — trace a thread through three millennia of human knowledge management:

Codex gave us the book, the portable vessel for human thought.
Code abstracted that knowledge into systems of symbols and rules.
Claude (the AI) now helps us navigate, create, and extend those systems.

From Roman legal collections to GitHub Copilot to terminal-based development agents, the impulse remains constant: capture intelligence, structure it, and make it actionable. The medium changes. The mission endures.

Terminal command line glowing on screen — where code meets the modern codex
From papyrus to terminals

Words have histories. The codex that replaced the scroll now lives inside the code that runs on machines named after mathematicians who built on ideas from emperors who ruled two thousand years ago. Language is the oldest technology — and AI is its latest iteration.