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

2026-01-05 · Source: VentureBeat · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Emerging Technologies & Innovation · Depth: Advanced, short

Summary

Boris Cherny, creator and head of Claude Code at Anthropic, recently revealed his personal software development workflow, which has garnered significant attention in the engineering community. His method involves orchestrating five Claude AI agents in parallel within a terminal, using iTerm2 system notifications for input management, and an additional 5-10 agents on claude.ai with a "teleport" command for session handoff. Cherny exclusively uses Anthropic's slowest but smartest model, Opus 4.5, arguing that its superior accuracy reduces human correction time. His team also maintains a CLAUDE.md file in their git repository to document AI mistakes, turning every error into a permanent learning rule for the agents. The workflow heavily relies on slash commands for automating repetitive tasks like /commit-push-pr and employs specialized subagents for architectural cleanup and end-to-end testing. A critical component is the verification loop, where Claude tests its own changes via browser automation and bash commands, reportedly improving code quality by "2-3x" and contributing to Claude Code's reported $1 billion in annual recurring revenue.

Key takeaway

For AI Engineers and CTOs evaluating developer productivity tools, Cherny's workflow demonstrates that treating AI as a parallel workforce, rather than just an assistant, can yield exponential gains. You should explore implementing multi-agent orchestration with self-correction mechanisms and prioritize smarter, slower models like Opus 4.5 to minimize human intervention and maximize output quality, potentially transforming your team's development capacity.

Key insights

Orchestrating multiple smart AI agents with self-correction and verification loops dramatically boosts developer productivity.

Principles

• Prioritize smart models over fast ones to reduce human correction.
• Automate repetitive tasks with custom slash commands.
• Implement verification loops for AI-generated code quality.

Method

Run multiple AI agents in parallel, manage input with system notifications, use a shared knowledge file (CLAUDE.md) for error learning, and deploy specialized subagents for development phases.

In practice

• Configure iTerm2 for multi-agent terminal management.
• Create a CLAUDE.md file for AI error logging.
• Develop custom slash commands for git operations.

Topics

• Claude Code
• Multi-Agent Systems
• Developer Workflow
• LLM Orchestration
• AI-Powered Testing

Best for: AI Engineer, CTO, VP of Engineering/Data, Software Engineer, Machine Learning Engineer, AI Architect

Related on AIssential

• A harness for every task: dynamic workflows in Claude Code — Claude Code's dynamic workflows enable on-the-fly multi-agent orchestration to overcome single-context AI limitations.
• Building Claude Code with Boris Cherny — AI-driven coding tools like Claude Code are transforming engineering workflows, emphasizing planning and rapid iteration over manual coding.
• The Rise of the Workflow Engineer: Why Your Code is Obsolete, But Your Coordination Skills Aren’t — AI's coding capabilities are making code generation a commodity, shifting value to coordination and orchestration skills.
• Conductors to Orchestrators: The Future of Agentic Coding — Software engineering is shifting from direct coding to orchestrating autonomous AI agent fleets for scaled productivity.
• Claude Code now spins up 100s of parallel agents on one task — AI agent capabilities are rapidly expanding through on-device execution, parallel orchestration, and enhanced data interaction.
• Compound Engineering: How Every Codes With Agents — Compound engineering leverages AI agents in a learning loop to make software development progressively easier and more efficient.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by VentureBeat.