Back to Daily Feed 
Anthropic's Claude Code Team: Inside Their AI-Driven Workflows
Editor's Pick
Originally published on Simon Willison's Weblog by Simon Willison
View Original Article
Share this article:

Summary & Key Takeaways
- Claude Tag now handles 65% of product engineering PRs for the Claude Code team.
- Automated code review is increasingly used for "outer layers" of the product.
- Best practices for prompt engineering have evolved, with smaller, less restrictive prompts.
- Anthropic's "ant fooding" culture drives internal tool adoption and feature validation.
- Fable, their AI, is competent at video editing, even for its own launch video.
- The team emphasizes "auto mode" as an enabling technology for collaboration.
- Working in public internally is key to their success.
Our Commentary
This fireside chat is an absolute goldmine. The insights into how Anthropic uses its own AI tools are mind-blowing. 65% of PRs by an AI? Prompt engineering becoming simpler? This isn't just theory; it's real-world application from a leading AI lab. We're seeing a genuine paradigm shift in how software is built, and this article pulls back the curtain.
View Original Article
Share this article: