The creator of Claude Code just revealed his workflow (venturebeat.com)

Boris Cherny, the creator of Claude Code at Anthropic, has unveiled a revolutionary workflow that dramatically enhances coding efficiency, capturing significant attention from the engineering community. His approach involves running multiple AI agents in parallel—specifically five versions of Claude—enabling a single developer to manage tasks simultaneously, akin to a real-time strategy game. This method not only streamlines the coding process but also applies a "do more with less" philosophy, showcasing how superior orchestration of existing models can boost productivity far beyond traditional methods. One of the standout aspects of Cherny's workflow is his use of Anthropic's slower but smarter Opus 4.5 model, which minimizes the need for correction by leveraging a deeper understanding of code. Additionally, he tackles the issue of AI memory by maintaining a single, evolving file that encapsulates lessons learned from each coding error, thereby allowing the AI to adapt and improve over time. By automating tedious tasks through custom commands and deploying specialized subagents for different phases of development, Cherny’s methods suggest a paradigm shift in software engineering, encouraging developers to utilize AI not just as a tool, but as a powerful co-worker that fundamentally changes the way they approach coding.