Anthropic, the US-based AI startup, has revealed that its AI model Claude now writes 90% of the company’s code. CFO Krishna Rao disclosed the productivity gains in an interview, highlighting how the AI-driven approach is reshaping software development at the firm. Rao, an Indian-origin executive, emphasized the efficiency improvements but did not specify the timeline for the shift.
Krishna Rao, Anthropic’s CFO, stated that Claude’s ability to generate 90% of the company’s code has led to measurable productivity gains. While the exact metrics were not disclosed, Rao indicated that the AI’s role in coding has streamlined development workflows. Anthropic, known for its safety-focused AI models, has been expanding its capabilities beyond chatbots to include advanced software development tools. The company’s reliance on Claude for coding reflects broader industry trends where AI is increasingly automating repetitive tasks.
Rao, who joined Anthropic in 2023, is a key figure in the company’s financial and operational strategy. Prior to Anthropic, he held leadership roles at tech firms, including a stint at an Indian unicorn. His comments underscore the growing integration of AI in core business functions, particularly in software development. Anthropic, founded in 2021, has raised over $7 billion in funding and is valued at $18.4 billion, positioning it as a major competitor to OpenAI and Google’s DeepMind.
The revelation about Claude’s coding capabilities comes amid a surge in AI adoption across industries. Anthropic’s models, including Claude 3.5 Sonnet, are designed to handle complex tasks such as coding, data analysis, and content generation. The company’s focus on AI safety and alignment has attracted significant investor interest, including backing from tech giants like Amazon and Google. Rao’s disclosure highlights how AI is not just augmenting but actively replacing human tasks in software development.
Anthropic’s use of Claude for coding aligns with its broader mission to develop AI systems that are both powerful and safe. The company’s models are trained using a technique called constitutional AI, which aims to ensure ethical and reliable outputs. Rao’s comments suggest that Claude’s coding proficiency is a result of this approach, enabling it to generate functional and secure code. However, the company has not disclosed the specific programming languages or frameworks Claude is most proficient in.
The productivity gains from Claude’s coding capabilities could have significant implications for the software industry. If AI models like Claude can reliably write 90% of code, companies may reduce their reliance on human developers for routine tasks, shifting their focus to higher-level design and problem-solving. Rao’s remarks also raise questions about the future of software engineering roles, particularly in markets like India, where coding is a major employment sector.
Anthropic’s rapid growth has been fueled by its ability to attract top talent and funding. The company’s valuation of $18.4 billion reflects investor confidence in its AI models and safety-focused approach. Rao’s background, including his experience in Indian tech, adds a unique perspective to Anthropic’s leadership team. His insights into Claude’s coding capabilities provide a glimpse into how AI is transforming software development at scale.
While Anthropic has not released detailed benchmarks for Claude’s coding performance, Rao’s statement suggests the AI’s capabilities are already being deployed in production environments. The company’s emphasis on safety and alignment could make Claude a preferred choice for enterprises looking to integrate AI into their development workflows. However, challenges remain, including ensuring the AI’s code is free from biases or vulnerabilities.