Reading up on Sentry

With layered quality practices like spec-driven development and AI-assisted testing, coding agents can double output without increasing bugs.

The article explains how to construct a reliable AI‑driven software factory using dual loops, explicit human bottlenecks, and deterministic validation to improve code quality while minimizing manual agent output review.

Building a practical Slack agent required 100k lines of TypeScript and four months of iteration to handle serverless constraints, credential management, and secure tool orchestration.

Linear Agent can now write and review code using Claude Code and Codex, enabling end-to-end issue-to-ship workflows from within Linear.

Claude Managed Agents decouples model reasoning from code execution to let teams deploy production-grade agents without building custom infrastructure.

Sentry cut build times from 14 to under 4 minutes by replacing its headless CMS with Astro, Markdown files, and Claude Code skills for content management.

Coding agents increase senior engineer output but not product velocity, as the bottleneck is product taste and complexity management, not code production speed.

Fair Source licensing remains viable in the AI era by using contract law, not just copyright, to prevent clones from competing, even as AI-generated code challenges traditional open-source enforcement.

Warden is a CLI tool that uses Markdown-defined skills to create AI agents that automatically review every pull request for security and correctness issues.

AI's ability to let enterprises quickly build fit-for-purpose internal tools is undermining bloated, expensive B2B SaaS products that fail to evolve.

The article argues that MCP's value is in steering agents with rich context descriptions, that progressive disclosure is counterproductive, and that subagents may better manage context rot.

Production monitoring and A/B testing offer more reliable evaluation of AI agents than offline evals, which fail to capture real-world performance.

Skills and MCP tools serve different but complementary purposes for coding agent harnesses, and neither replaces the other despite hype cycles claiming otherwise.

Armin Ronacher reflects on 2025 as a year of agentic coding tools like Claude Code, which replaced much of his direct programming, and discusses the cultural, technical, and social challenges this shift presents for software engineering.