The creator of Claude Code just revealed his workflow, and developers are losing their minds
Not a pilot. Boris Cherny just showed how to run 5 Claude agents in parallel—and developers are treating it like the playbook for the next era of coding.
Why it matters
Claude Code's creator revealed a workflow that treats AI as autonomous workforce, not assistant—managing multiple parallel agents, verification loops, and institutional memory through git. This signals a fundamental shift in how AI-augmented development works at scale, validating Anthropic's 'do more with less' strategy against competitors' trillion-dollar buildouts.
The key facts
10 to knowBoris Cherny runs 5 Claude agents in parallel via iTerm2 terminal
Additional 5-10 Claude instances on claude.ai in browser with 'teleport' handoff
Uses Opus 4.5 exclusively—slower model reduces human correction time vs. smaller models
CLAUDE.md file in git repo captures every AI mistake as permanent instruction rule
Slash commands (e.g., /commit-push-pr) automate version control workflows
Verification loop with browser automation and end-to-end testing improves code quality by '2-3x'
Claude Code reportedly hit $1B annual recurring revenue
Workflow compared to real-time strategy game (StarCraft metaphor)
Thread went viral on X with industry observers calling it 'watershed moment'
Validates Anthropic President Daniela Amodei's 'do more with less' strategy
Go to the source
VentureBeat AIventurebeat.com
Publisher excerpt: When the creator of the world's most advanced coding agent speaks, Silicon Valley doesn't just listen — it takes notes. For the past week, the engineering community has been dissecting a thread on X from Boris Cherny, the creator and head of Claude Code at Anthropic. What began as a casual sharing…