Boris Cherny on AI, Claude Code, and the Future of Software Engineering

Boris Cherny's Unconventional Journey

• 💡 Boris Cherny, creator of Claude Code, has an unconventional background, studying economics and dropping out, but founded his first company at 18, becoming a generalist in product, design, and coding.
• 🚀 His early career at Meta involved cross-functional collaboration on projects like Messenger/Facebook Groups, where he used a cafeteria survey method for user research.
• 🎯 He developed a product design principle focused on identifying latent user needs, observing how users organically used platforms (e.g., Facebook groups for buying/selling led to Marketplace).

Joining Anthropic: A Vision for AI

• 🧠 Despite a senior role at Meta, Boris made the "unexpected" decision to join Anthropic, driven by the exponential growth of large language models and a deep personal interest in sci-fi.
• ✅ He found Anthropic's culture aligned with his values, prioritizing AI safety and alignment as core to product development, unlike Meta where safety was often seen as a burden.
• 🤝 The company's strong mission-driven environment and "startup common sense" fostered a culture where everyone, including product managers and data scientists, could write code.

Claude Code's Transformative Impact

• 🛠️ Claude Code, initially not superior to competitors, succeeded by designing for future model capabilities, anticipating the exponential improvement of LLMs over 6-month cycles.
• 📈 This foresight led to Claude Code generating 80-90% of Anthropic's internal code, significantly boosting engineer productivity and expanding coding capabilities to non-technical roles.
• ⚠️ Boris emphasizes that AI-generated code must meet the same quality standards as human-written code, requiring engineers to audit, refine, and ensure reliability, especially for critical modules.

The Evolving Role of Engineers

• 🔄 Engineers are transitioning from direct code producers to designers and managers of AI agents, focusing on planning, architecture, and risk assessment rather than just writing lines of code.
• 🚀 This shift allows engineers to delegate tedious, repetitive tasks to AI, freeing them to concentrate on higher-value, creative, and strategic aspects of software development.
• 💡 AI tools lower the barrier to writing code but do not diminish the overall complexity or importance of software engineering principles, such as understanding business logic and ensuring quality.

Adapting to the AI Era

• 🌱 Boris advises engineers to embrace change and view AI as a powerful tool that can enhance their capabilities, rather than a threat to their jobs.
• 🔑 Success in the AI era involves learning to effectively communicate requirements to AI, critically review AI-generated code, and leverage AI to solve complex technical challenges.
• 🧑‍💻 For non-CS background individuals, his experience highlights that programming is a practical skill best learned through hands-on problem-solving and building, not just theoretical knowledge.