Gergely Orosz · 6 min read · Gen AI Tools
Building Claude Code with Boris Cherny
Read the original on newsletter.pragmaticengineer.com →
In this Pragmatic Engineer episode, Boris Cherny explains how Claude Code was built at Anthropic, the parallel-agent workflow and plan-first approach that enables high PR throughput, and how AI-assisted coding is shifting engineers’ roles and collaboration with product, design, and infrastructure. The discussion covers deterministic review patterns, context retrieval from large codebases, and the broader implications of AI tooling in software development.
Key Takeaways
- Using multiple parallel Claude instances and a plan-first workflow can drive high throughput, with 20-30 PRs shipped per day.
- Code quality and complete migrations materially boost AI-assisted productivity; partially migrated codebases create confusion for both humans and models.
- Agentic search, driven by glob and grep rather than complex databases, can outperform more sophisticated retrieval approaches in code discovery.
- Automating repetitive code-review patterns with lint rules and data-driven feedback reduces manual review time and improves consistency.
Excerpt
Stream the latest episode Listen and watch now on YouTube , Spotify , and Apple . See the episode transcript at the top of this page, and timestamps for the episode at the bottom. Brought to You by • Statsig — ⁠ The unified platform for flags, analytics, experiments, and more. Stop switching between different tools, and have them all in one place. Check out Statsig. • Sonar – The makers of SonarQube, the industry standard for automated code review. Proactively find and fix issues in real-time with SonarQube MCP Serve r by connecting AI tools like Claude Code, GitHub…