Anthropic Executives Confirm Claude AI Now Writes All Internal Code
In a groundbreaking revelation, executives at Anthropic have confirmed that Claude Code, the company's advanced AI system, is now responsible for writing nearly 100% of the firm's internal code. This announcement comes amidst a curious observation: Anthropic's job page still lists over 100 open engineering positions, sparking questions about the evolving role of human engineers in an AI-driven world.
Addressing the Contradiction: Engineers' New Responsibilities
Boris Cherny, the head of Claude Code, directly addressed this apparent contradiction when an X user highlighted it. He explained that the nature of engineering is undergoing a significant transformation. "Someone has to prompt the Claudes, talk to customers, coordinate with other teams, decide what to build next," Cherny wrote. He emphasized that while AI handles the coding, great engineers are more crucial than ever, shifting from hands-on programming to strategic oversight.
The Backdrop: AI Summit Insights and Internal Shifts
The context for this shift was detailed at the 2026 Cisco AI Summit, where Anthropic's Chief Product Officer, Mike Krieger, confirmed that AI tools now generate almost all internal code. "Claude products and Claude code are being entirely written by Claude," he stated, building on CEO Dario Amodei's earlier prediction that AI would manage 90% of coding—a target that has been surpassed within Anthropic itself.
In a Fortune interview, Boris Cherny added more depth to this narrative, revealing that he hasn't written a single line of code in over two months. Instead, he ships 22 to 27 pull requests daily, all 100% generated by Claude, showcasing how AI is taking over repetitive tasks while humans focus on higher-level decision-making.
Practical Implications: Engineers Move Up the Stack
In practice, this means engineers are moving up the value chain. Repetitive tasks such as drafting features, refactoring code, and writing documentation are now delegated to Claude. Human engineers concentrate on system design, architectural decisions, customer interactions, and strategic planning for future projects. Importantly, all AI-generated output is still rigorously reviewed and validated by humans before deployment, ensuring quality and alignment with company goals.
Broader Industry Trends: AI's Growing Role in Coding
This trend is not isolated to Anthropic. Companies like Microsoft and Salesforce report that approximately 30% of their code is AI-generated, while a recent study in the Science journal estimated that around 29% of GitHub Python functions in the U.S. are created by AI. However, within frontier AI labs like Anthropic, the adoption curve is much steeper. Cherny believes the rest of the industry will soon catch up, as AI continues to automate coding tasks and redefine engineering roles globally.
The Human Element: Running the Show Amidst AI Advancements
Despite Claude writing Claude, humans remain firmly in control, orchestrating the broader vision and ensuring ethical and effective implementation. This shift highlights a pivotal moment in technology, where AI handles the grunt work, freeing engineers to make bigger, more impactful decisions that drive innovation forward.
