A mix of practical tooling, performance engineering, and testing philosophy.
-
Hooking Claude Code Up to WordPress: This walkthrough is written for a non-developer audience: marketers and content people who want to use Claude Code to manage their WordPress sites. It covers MCP server setup, connecting to the WordPress REST API, and the kinds of tasks that work well (bulk content updates, metadata management, template tweaks) versus the ones that don’t (visual design decisions, complex plugin interactions). If you work with non-technical people who’ve heard about Claude Code and want to try it, this is a useful reference to share.
-
Laravel blog post on Nightwatch: This official write-up explains how the feature cut load times by 60% in Laravel Cloud. Nightwatch handles background job monitoring and scheduling visibility with a different architecture than the traditional approach. The performance improvements come from how it processes and displays monitoring data, not from caching tricks. It’s a good window into where Laravel Cloud is heading as a platform.
-
Local Code Embeds: Aleksandr Hovhannisyan describes a self-hosted alternative to GitHub Gists for embedding code snippets in blog posts. Instead of depending on an external service (which can change its API, styling, or availability), he built a system that reads code from local files at build time and renders them as syntax-highlighted code. The result is faster loading, no external dependencies, and complete control over styling. A nice approach if you run a technical blog and want your code samples to be as reliable as the rest of your content.
-
5 Agent Skills I Use Every Day: This AI Hero roundup covers Claude Code skills that have become daily habits: code review with specific rubrics, commit message generation, test scaffolding, refactoring with constraints, and documentation generation from code comments. The useful part is the specific prompting patterns and skill configurations that make each one reliable enough to use without constant babysitting.
-
Faster Than Understanding: Sebastian Bergmann’s short, uncomfortable essay is about the risk of writing tests that pass before you’ve understood what you’re actually testing. Modern tooling, and especially AI-assisted development, makes it trivially easy to iterate until green. But green without understanding is a false signal. The test says “this works” but you can’t explain why it works, or what it would take to break it. The essay applies well beyond PHPUnit, to the relationship between speed and comprehension in any craft. I keep returning to it.