AI coding agents in WordPress Playground, wired with MCP
The @wp-playground/mcp package lets an AI client drive a Playground instance over MCP: run PHP, write files, query the database and check the front end, all inside the browser.
Category: Development
Modal vs separate page UX, and how to choose
Modals are interruptive by design. Four checks for deciding between a modal, a side drawer and a dedicated page, plus the tasks each one is actually good for.
Where to cache in a RAG pipeline besides the prompt
Prompt caching gets all the attention while the expensive work happens before the model sees a token. Five places to cache in a RAG pipeline, from query embeddings to full query-response pairs, with a transients example for WordPress.
What the plugin directory MCP server gives your IDE
WordPress.org now runs an MCP server for the plugin directory, so an AI assistant can validate readme.txt, check where a submission sits in the queue and read the guidelines from the source. The review team still holds you responsible for the code.
WordPress 7.0 RC1 slips to March 24 over performance
RC1 for WordPress 7.0 was due today and now lands on March 24, 2026. The holdups are real-time collaboration performance and client-side media optimization, plus a fresh look at release package size. Both are worth watching before you update.
You can now run the WordPress Plugin Directory locally
The Plugin Directory now ships a wp-env setup, so you can boot a local copy instead of guessing how its backend behaves. Here is what works inside it, and where it stops short of the live site.
How Neuro-Symbolic AI makes fraud rules auditable
A run at differentiable rule induction: the model wrote its own IF-THEN fraud rules, rediscovered V14 without help, and went dark in three of five runs. Notes on the consistency loss and when the architecture is worth shipping.
WordPress 7.0 Beta 6: 132 fixes and one big revert
Beta 6 lands 132 fixes, drops Client-side Media Processing to shrink the package, and stretches Real-Time Collaboration polling by 4x. Collaboration is opt-in behind a constant now, and RC1 is due March 24th.
What AI coding assistants still get wrong in production
AI coding assistants write decent functions and miss the system around them. A naive REST endpoint next to a reviewed one, two outage stories, and the workflow I use to keep an agent on a leash.
Linear regression is a projection problem
Linear regression is a projection: you look for the point where the error vector meets the feature space at a right angle. Vectors, dot products, a forest-and-highway analogy, and the scikit-learn code that matches it.