TLDR: Anthropic CEO Dario Amodei announced that their Claude AI model is now responsible for writing 90% of the company’s code, a significant milestone in AI-driven software development. Despite this high level of automation, Amodei emphasized that human engineers remain crucial for innovation, oversight, and complex problem-solving, highlighting a shift towards a more collaborative human-AI workflow.
In a groundbreaking revelation that underscores the accelerating impact of artificial intelligence on the tech industry, Anthropic CEO Dario Amodei has confirmed that the company’s Claude AI model now generates approximately 90% of its internal code. This announcement, made during the Dreamforce conference, marks a significant fulfillment of a prediction Amodei had made earlier, signaling a profound transformation in software development practices within leading AI firms.
Amodei’s statement, widely reported on October 17, 2025, clarifies that while AI has achieved remarkable efficiency in coding, the role of human engineers is far from diminished. Instead, he stressed their continued indispensability for critical functions such as supervision, editing, creative problem-solving, and navigating the more intricate aspects of software innovation. This perspective positions AI not as a replacement for human talent but as a powerful augmentative tool.
According to Amodei, the synergy between human engineers and AI can lead to a dramatic increase in productivity, potentially making developers up to ten times more efficient. This enhanced productivity stems from AI handling the bulk of repetitive coding tasks, thereby freeing human experts to concentrate on higher-level architectural design and complex logical challenges.
The capabilities of Anthropic’s AI models are rapidly advancing, with the latest iteration, Claude Sonnet 4.5, reportedly capable of generating as many as 11,000 lines of code autonomously. This impressive feat highlights the rapid evolution of generative AI in programming, a trend sometimes referred to as ‘vibe coding’ that is gaining traction across the tech sector.
Addressing concerns about potential job displacement in the tech industry, particularly following recent layoffs at major firms, Amodei adopted a nuanced stance. He predicted that while entry-level coding positions might see a reduction of up to 50% over the next five years, the overall demand for skilled software engineers is expected to grow. This growth, he explained, would be driven by the increased leverage and productivity that AI tools offer, fostering a future of hybrid human-AI teams.
Also Read:
- Anthropic Projects Significant Revenue Growth, Targeting Up to $26 Billion by 2026
- Anthropic Unveils ‘Skills for Claude,’ Empowering AI Agents with Enhanced Task Capabilities
Amodei’s vision aligns with predictions from other prominent figures in the AI and tech landscape. OpenAI CEO Sam Altman previously stated that AI would become the ‘best coder in the world’ by the end of 2025. Similarly, Google CEO Sundar Pichai revealed that AI already contributes to over 25% of the code written at Google, and Instagram co-founder Mike Krieger anticipated a shift where developers would engage more in code reviewing than in actual code writing. These converging views underscore a consensus that the future of software engineering will be defined by a collaborative partnership between human ingenuity and artificial intelligence.
