"At QCon San Francisco 2025, Adam Wolff described how Claude Code at Anthropic is built with an AI coding assistant at the center of the workflow. He reported that about ninety percent of the production code for the tool is written with or by Claude Code. The team ships continuously to internal users and targets weekday releases for external users."
"In the first story, they introduced a virtual Cursor class that models the text buffer and cursor position as an immutable value. The initial implementation was a few hundred lines of TypeScript supported by a substantial test suite. Later, another engineer added Vim mode on top in a single pull request, with hundreds of lines of logic and tests generated with Claude Code."
"As adoption grew across languages, Unicode related issues began to surface. The team added grapheme clustering and a later refactor reduced worst-case latency from several seconds per keystroke to a few milliseconds by deferring work and using more efficient search strategies. Wolff treats this story as an example of a successful experiment in which the pain of additional complexity decreased over time and the architecture continued to support fast changes."
Claude Code at Anthropic places an AI coding assistant at the center of the development workflow, producing about ninety percent of production code and enabling continuous internal shipping with weekday releases for external users. Rapid generation and refactoring of code and tests shifts emphasis from upfront planning to fast shipping, production observation, and iterative requirement updates. Rich terminal input with slash commands, file mentions, and keystroke-specific behavior was implemented intentionally to control every keystroke despite conventional advice. A virtual Cursor class modeled the text buffer immutably, enabled later Vim mode additions, and subsequent grapheme clustering and refactors reduced keystroke latency dramatically.
Read at InfoQ
Unable to calculate read time
Collection
[
|
...
]