90% | Armin Ronacher's Thoughts and Writings
lucumr.pocoo.org/2025/9/29/90-percent
What Jamie wrote
Very good writeup on how the author is using coding agents to create real solutions.
Today I use Claude Code and Codex. Each has strengths, but the constant is Codex for code review after PRs. It’s very good at that. Claude is indispensable still when debugging and needing a lot of tool access (eg: why do I have a deadlock, why is there corrupted data in the database etc.). The working together of the two is where it’s most magical. Claude might find the data, Codex might understand it better.
Worth reading and seeing how he is getting "north of 90%" of code agentically created. Also notable what he doesn't allow the agent to do.
Same host · 6 links from lucumr.pocoo.org
- Rust for Python Programmers | Armin Ronacher's Thoughts and Writings
- Python's Hidden Regular Expression Gems
- Agent Design Is Still Hard | Armin Ronacher's Thoughts and Writings
- Pi: The Minimal Agent Within OpenClaw | Armin Ronacher's Thoughts and Writings
- The Coming Loop | Armin Ronacher's Thoughts and Writings
Saved the same week · 25 links that week
- 2025 NIST password guidelines: key updates for businesses | Proton
- Claude Sonnet 4.5 is probably the “best coding model in the world” (at least for now)
- Introducing Kagi News | Kagi Blog
- container: A tool for creating and running Linux containers using lightweight virtual machines on a Mac
- Designing agentic loops