As AI Writes Anthropic’s Code, Engineers Redefine Their Role in the Development Process
At Anthropic, AI now writes nearly all internal code, but engineers remain central in guiding, reviewing, and shaping innovation.
Artificial intelligence is no longer just assisting developers at Anthropic — it is writing nearly all of the company’s internal code. The revelation has sparked fresh debate about the evolving role of software engineers in an AI-driven world.
Speaking earlier at the 2026 Cisco AI Summit, Anthropic’s chief product officer Mike Krieger shared that the company’s AI systems, powered by Claude, now generate almost 100 per cent of its internal codebase. In other words, Claude is effectively helping build itself. The development aligns with earlier projections by CEO Dario Amodei, who had predicted that AI would soon handle the bulk of coding tasks inside the company.
The announcement immediately triggered questions across the tech community. If AI is writing nearly all the code, what are human engineers doing? One user on X pointed out the apparent contradiction: “Claude Code is writing 100% of Claude code now. But Anthropic has 100+ open dev positions on their jobs page?”
Responding to the query, Boris Cherny, head of Claude Code, clarified that coding represents just one dimension of engineering work. “Someone has to prompt the Claudes, talk to customers, coordinate with other teams, decide what to build next. Engineering is changing and great engineers are more important than ever,” Cherny wrote.
According to Anthropic’s leadership, AI tools are now capable of generating massive pull requests spanning thousands of lines. These systems can draft new features, refactor existing components, and even produce documentation. However, the process does not operate unchecked. Human engineers carefully review, validate and approve AI-generated output before it is merged or deployed. Strict guardrails and internal workflows are in place to maintain reliability and quality.
Beyond reviewing code, engineers remain deeply involved in architectural decisions, long-term product planning and system design. Their focus has shifted from writing routine code to steering AI systems and shaping broader strategic direction.
The transformation at Anthropic reflects a wider industry trend. Across technology firms, AI is increasingly taking over repetitive and time-consuming programming tasks. Developers are gradually transitioning toward higher-level responsibilities such as design thinking, cross-functional collaboration and customer engagement. Coding itself, once the defining skill of software engineering, may no longer dominate daily workflows.
Still, not everyone sees this transition as purely evolutionary. Some tech leaders believe the disruption could go even further. Elon Musk, CEO of xAI, recently predicted that traditional coding might soon become obsolete. “I think (..) by the end of this year you don't even bother doing coding. The AI will just create the binary directly. And the AI can create a much more efficient binary than can be done by any compiler,” Musk said in a widely circulated video.
Whether that vision materialises remains uncertain. For now, Anthropic’s experience suggests that while AI can generate code at remarkable scale, human engineers continue to play a vital role — not as typists of syntax, but as architects, reviewers and decision-makers shaping what technology should build next.
The future of software engineering may look different, but at least for now, it still firmly includes humans at the helm.
