
Run several sessions at once, plan before building, and give the agent a way to check its own work.
Steal this
- Start most sessions in plan mode.
- Every mistake becomes a line in the shared CLAUDE.md.
- Pre-allow safe commands instead of turning checks off.
- Give the agent a way to verify its own work.
The lesson and the takeaways are our summary. Everything below is theirs, with where they said it.
The setup
FactualMachines
A terminal with five numbered tabs, plus the web and his phone.
Software
Claude Code, mostly as it ships. Shared slash commands, subagents and hooks checked into git.
Models
Opus 4.5 with thinking for everything, at that date.
As of January 2, 2026. Setups change; follow the links.
Their gear
- Claude Code Anthropic's coding agent.
Links go to the maker's own page. We earn nothing from them. Every machine is on the gear shelf.
How they work
Factual- Five sessions in the terminal, with system notifications when one needs him. (Boris Cherny on X)
- More sessions on the web, some started from his phone in the morning. (Boris Cherny on X)
- Most sessions start in plan mode. (Boris Cherny on X)
- One shared CLAUDE.md in git. Every mistake becomes a line in it. (Boris Cherny on X)
- Repeated workflows become slash commands and subagents. (Boris Cherny on X)
- Safe commands are pre-allowed instead of turning permission checks off. (Boris Cherny on X)
- The most important habit is letting the agent verify its own work. (Boris Cherny on X)
- By September 2026 he ran Opus 5.5 on medium effort for most work, low for speed and xhigh for more thinking. (Boris Cherny on X)
- In October 2026 he kept medium about 95% of the time and raised Opus to xhigh or max for complex visual work. (Boris Cherny on X)
Where they talk and write
Interviews 19
- Meet the miso-making engineer behind Anthropic’s biggest hit productCNN Business ·
- In Conversation with Boris Cherny, Creator of Claude CodeComputer History Museum (CHM Live) ·
- ‘Advanced Insights’: Agentic AI and the Future of SoftwareAMD Advanced Insights ·
- Boris Cherny: Building Claude CodeY Combinator Startup Podcast ·
- The Creator of Claude Code on The Hottest Piece of Software in the World | Odd LotsOdd Lots (Bloomberg) ·
- Fireside Chat with Boris Cherny, Head of Claude Code@Scale Conference ·
13 older
- Boris Cherny: Claude Code & the Future of Engineering | Acquired Unplugged presented by WorkOSAcquired Unplugged (WorkOS) ·
- Claude Code creator Boris Cherny on the end of the software engineerPlatformer ·
- Claude Code Head Boris Cherny: Insane Growth, Tokenmaxxing, AI Agents' Next FrontierBig Technology Podcast ·
- Anthropic's Boris Cherny: Coding's Printing Press MomentTraining Data (Sequoia) ·
- Building Claude Code with Boris ChernyThe Pragmatic Engineer ·
- How Claude Code Claude CodesThe Vergecast ·
- Head of Claude Code: What happens after coding is solved | Boris ChernyLenny's Podcast ·
- Inside Claude Code With Its Creator Boris ChernyLightcone (Y Combinator) ·
- Boris Cherny (Creator of Claude Code) On How His Career GrewDeveloping Dev (The Peterman Pod) ·
- How to Use Claude Code Like the People Who Built ItAI & I (Every) ·
- Anthropic's Surprise Hit: How Claude Code Became an AI Coding PowerhouseThe MAD Podcast ·
- Claude Code: Anthropic's Agent in Your TerminalLatent Space ·
- SE Radio 384: Boris Cherny on TypeScriptSoftware Engineering Radio ·
Writing 19
- I am often wrongBoris Cherny's Blog ·
- Claude Code: Best practices for agentic codingAnthropic ·
- NPM and NodeJS should do more to make ES Modules easy to useBoris Cherny's Blog ·
- Learning to work (very) remotelyBoris Cherny's Blog ·
- Time and TimingBoris Cherny's Blog ·
- Dinosaur Food: 100 million year old foods we still eat todayBoris Cherny's Blog ·
13 older
- Data Management on the Web Still SucksBoris Cherny's Blog ·
- React+TypeScript: Use unions of objects for propsBoris Cherny's Blog ·
- On ContagionBoris Cherny's Blog ·
- 13 Tips for Writing a Technical BookBoris Cherny's Blog ·
- Programming TypeScriptBook · O'Reilly Media ·
- On Derived StateBoris Cherny's Blog ·
- JavaScript in 2017: Year in Review, Predictions for 2018Boris Cherny's Blog ·
- Towards a Safer ReduxBoris Cherny's Blog ·
- Angular 1: 8 Lessons For Designing JavaScript FrameworksBoris Cherny's Blog ·
- Options in TypeScriptBoris Cherny's Blog ·
- Introducing: Lazy Arrays in JavaScriptBoris Cherny's Blog ·
- The Best Frontend JavaScript Interview Questions (written by a Frontend Engineer)Boris Cherny's Blog ·
- Automatically Generating Base64 Inline Images With SASSBoris Cherny's Blog ·
Each link goes to the original page, newest first.
Habits worth copying
Give the agent a way to check its work, and do not take "done" for an answer.Calls a way to verify the most important thing.How much effort, or thinking, should the model spend?Medium about 95% of the time. Xhigh or max for complex visual work.Approvals stay on, or go off when it is safe?Never skips permissions. Pre-allows the bash commands he knows are safe.Review the plan before the agent edits anything.Starts most sessions in plan mode, shift+tab twice.Turn what you do many times a day into a slash command.A slash command for every inner-loop workflow.When the agent gets something wrong, write the correction into the instruction file.His team adds a line to the shared CLAUDE.md each time.