technologyScore 45/100Research
Anthropic now uses Claude to write ~90% of its own code; engineers shift to supervisory/editing role as AI-driven productivity multiplier compounds model improvement cycles.
“Within Anthropic and within a number of companies that we work with... on many teams... Claude can both... write this feature and Claude will do it and then the human kind of looks over the code... 90% of all of the code... the remaining 10% is now a complement to the AI system that's getting much better at what it's producing.”
Why it matters
This is the first concrete evidence of a major AI lab achieving AI-assisted productivity gains at scale in its own engineering function. It demonstrates the feasibility of the productivity multiplier thesis Amodei outlined and proves Claude's utility for internal R&D acceleration.
Investment implication
Software development tool companies (GitHub Copilot, Cursor, Replit, Lovable, Vercel) and IDEs built on Claude will likely see accelerating adoption as enterprises mimic Anthropic's workflow. This validates the emerging ecosystem of Claude-native developer tools and suggests a major tailwind for tools that optimize the human-in-the-loop coding loop.