The guy who built Claude Code runs 5 AI agents at once. Here’s how.

When the creator of the world’s most advanced coding agent shares his terminal setup, people pay attention. For the past week, Boris Cherny — the guy who built <a href="https://design.allwinchina.org/ai-tools/claude-code/" title="Claude Code review”>Claude Code at Anthropic — has been the subject of a viral thread on X that’s got developers rethinking how they write software.

Jeff Tang, a well-known voice in the dev community, put it bluntly: “If you’re not reading the Claude Code best practices straight from its creator, you’re behind as a programmer.” Kyle McNease went further, saying Anthropic might be facing “their ChatGPT moment.”

The wild part? Cherny’s workflow isn’t complicated. It’s almost deceptively simple. But it lets one person operate like a small engineering team. One user who tried it said it “feels more like Starcraft” than traditional coding. That’s not hyperbole — it’s a real shift from typing syntax to commanding autonomous units.

Coding as a real-time strategy game

Cherny doesn’t code in a linear fashion. He doesn’t write a function, test it, move to the next. He acts like a fleet commander.

“I run 5 Claudes in parallel in my terminal,” he wrote. “I number my tabs 1-5, and use system notifications to know when a Claude needs input.”

Using iTerm2 system notifications, he manages five simultaneous work streams. One agent runs a test suite, another refactors a legacy module, a third drafts documentation. He also runs 5-10 Claudes on claude.ai in his browser, using a “teleport” command to hand off sessions between the web and his local machine.

This validates the “do more with less” strategy Anthropic President Daniela Amodei talked about earlier this week. While OpenAI is building trillion-dollar infrastructure, Anthropic is proving that better orchestration of existing models can get you exponential productivity gains.

Why he uses the slowest model

This is the part that surprised me. In an industry obsessed with speed, Cherny uses Anthropic’s heaviest, slowest model: Opus 4.5.

“I use Opus 4.5 with thinking for everything,” he explained. “It’s the best coding model I’ve ever used, and even though it’s bigger & slower than Sonnet, since you have to steer it less and it’s better at tool use, it is almost always faster than using a smaller model in the end.”

This is a critical insight for anyone running AI in production. The bottleneck isn’t token generation speed — it’s the human time spent correcting the AI’s mistakes. Cherny’s approach pays the “compute tax” upfront to avoid the “correction tax” later. I’ve seen teams burn through hours fixing Sonnet’s output when Opus would have gotten it right the first time.

One file to rule them all

Cherny also solved a problem that’s been plaguing AI-assisted development since day one: amnesia. LLMs don’t remember your company’s coding style or architectural decisions from one session to the next.

His fix is elegant. His team maintains a single file called CLAUDE.md in their git repository. “Anytime we see Claude do something incorrectly we add it to the CLAUDE.md, so Claude knows not to do it next time,” he wrote.

This turns the codebase into a self-correcting organism. When a human reviews a pull request and spots an error, they don’t just fix the code — they tag the AI to update its own instructions. “Every mistake becomes a rule,” noted Aakash Gupta, a product leader analyzing the thread. The longer the team works together, the smarter the agent gets.

Slash commands for the boring stuff

The “vanilla” workflow people praised is powered by rigorous automation of repetitive tasks. Cherny uses slash commands — custom shortcuts checked into the project’s repository — to handle complex operations with a single keystroke.

He highlighted a command called /commit-push-pr, which he invokes dozens of times daily. Instead of manually typing git commands, writing a commit message, and opening a pull request, the agent handles the whole pipeline. It’s the kind of automation that sounds trivial until you realize how much mental overhead it eliminates.

What this means for the rest of us

Cherny’s workflow isn’t about having the most expensive setup or the latest hardware. It’s about rethinking the role of the developer. You’re not typing code anymore — you’re directing agents, reviewing their output, and teaching them to be better.

This is higher than I expected from a single workflow change. But the reaction from the developer community tells me this isn’t just hype. When people who actually build things say “this feels like Starcraft,” you should pay attention.