How Claude's Creator Codes with AI
How Anthropic Got to 100% AI-Written Code (And the Moats It Kills)
The creator of Claude Code hasn’t written a single line of code in 2026, and says 100% of the code at Anthropic is now written by the model, with zero manually written code anywhere at the company.
In Sequoia AI Ascent, Boris Cherny laid out exactly how he got there, from shipping 150 PRs in a day off his phone to running dozens of autonomous loops that babysit his codebase.
I pulled the full playbook: the workflow, the moats AI kills, and the hiring shift, so you can run it yourself without watching the whole thing.
In this guide you’ll find:
Why Claude Code was useless for 6 months on purpose (the pre-PMF bet founders should copy)
The exact workflow behind 150 PRs a day: loops, sessions, and hundreds of parallel agents
The 2 startup moats AI just killed, and why disruption is about to go 10x
The org change that is Anthropic’s real edge (not the models)
Who wins the next decade of software
The Claude Code Playbook: 10 Things Founders Can Do Today
Want to learn all about the AI Opportunity before it’s mainstream?
Here’s a good place to start:
↳Alex Karp’s Playbook for Which AI Companies Survive
↳19 great new details about Cursor’s rise
↳Perplexity’s CEO 2026 AI Pitch
↳The First 48 Hours of Claude Fable 5
Try premium free for 7 days. Or get 50% off this week only.
