The Pragmatic Engineer

Building Claude Code with Boris Cherny

1h 37min

Boris Cherny, creator of Claude Code, discusses his journey from a prolific Meta engineer to building one of the fastest-growing AI developer tools at Anthropic. He shares how Claude Code evolved from a side project, driven by the realization that AI models want to use tools, and the internal debate to release it for safety research. Cherny details his workflow of shipping 20-30 PRs daily with zero handwritten code, the AI-driven code review process, and how Anthropic's culture fosters rapid prototyping and a generalist approach to software engineering.

Summarized by Podsumo

🎧 Listen 🎙️ Ask about this episode

✨ Key Takeaways

1

Boris Cherny ships 20-30 pull requests daily with zero handwritten code, leveraging Claude Code's ability to one-shot implementations and even coding on his iOS app using parallel agents.
2

Claude Code originated as a simple chatbot but transformed when Boris realized AI models excel at using tools, leading to its agentic approach and subsequent release for safety research.
3

Every pull request at Anthropic is code-reviewed by Claude Code in CI, catching approximately 80% of bugs, with humans providing final approval, and Claude can even write Lint rules on demand.
4

Anthropic fosters a generalist engineering culture, where all 'Members of Technical Staff' emphasize rapid prototyping (dozens to hundreds of versions) over traditional documentation like PRDs, exemplified by Claude Co-Work being built in just 10 days.
5

Boris likens the current AI transformation to the printing press, where the specialized skill of 'scribes' (software engineers) is democratized, leading to an unpredictable expansion of new roles and making adaptability and multi-disciplinary skills more crucial than ever.

💬 Notable Quotes

"The model is its own thing. You give it tools. You give it programs that it can run. You let it run programs. You let it write programs, but you don't make it a component of this larger system in this way."
"I think one metaphor I have for this moment in time that we're in is the printing press and the 1400s or whatever. Because at that moment, it was actually quite similar. There was a group of scribes that knew how to write."
"I think something that still matters a lot today is being methodical in hypothesis driven... Other skills that I think are more valuable are being curious and being open to doing things beyond your swim lane."