The creator of Claude Code just revealed his workflow, and developers are losing their minds

Boris Cherny, the creator of the world's most advanced coding agent, has shared his personal terminal setup on X, which has sparked a viral discussion on the future of software development. Cherny's workflow is surprisingly simple, yet it allows a single human to operate with the output capacity of a small engineering department. He achieves this by running five AI agents in parallel in his terminal, utilizing system notifications to manage multiple work streams. Cherny's approach is a shift from traditional coding, where a programmer writes a function, tests it, and moves to the next, to commanding autonomous units. He exclusively uses Anthropic's heaviest and slowest model, Opus 4.5, which he finds to be the best coding model despite its slower speed. Cherny's team also maintains a single file named CLAUDE.md to address the problem of AI amnesia, where they document mistakes made by the AI so it can learn from them. The team uses slash commands and subagents to automate repetitive tasks, making the development process more efficient. Cherny's workflow has been praised for its ability to transform the codebase into a self-correcting organism, where every mistake becomes a permanent lesson. The verification loop is a key aspect of Cherny's workflow, where the AI tests every single change made to the code, improving the quality of the final result. Overall, Cherny's workflow signals a pivotal shift in how developers think about their craft, where AI is no longer just an assistant, but a workforce that can multiply human output by a factor of five.