Notes from the WordPress 7.0 product review meeting with Matt Mullenweg: a CSS-level admin refresh, a customizable dashboard, real-time co-editing over a pluggable transport, the AI Client and its settings UI, and PHP-only block registration.
Nine things I check when a WordPress site is not turning up in AI search: answer-first writing, topical clusters, JSON-LD schema, explicit authorship, and a robots.txt that is not quietly blocking GPTBot.
Lily Burton moved from biochemistry into science journalism and used the WordPress AI website builder to put a portfolio online in sixty minutes. Why shipping the prototype beat building the perfect site, and how to adjust the generated layout afterwards.
Linear RAG breaks down as soon as the solution path is not known up front. LangGraph swaps the chain for a stateful graph: a shared state object, nodes that update it, and conditional edges that pick the next step at runtime.
ML research runs on a cycle of downtime, flow and hard deadlines. After 14 years of WordPress work I started copying that rhythm, and it did more for code quality than any productivity hack I had tried.
Static thresholds flag every promo spike and still miss the slow climbs. This post walks through a hybrid setup: Z-scores and growth rates to catch candidates, then a GroqCloud agent that decides whether to fix, keep or escalate each point.
AI tools like Cursor and Claude Code prototype fast, then leave race conditions and unbounded queries behind. A look at what vibe coded WordPress code misses, with a before and after refactor of a WooCommerce loop.
Automattic shipped a developer preview of a Claude Cowork plugin and a set of Skills that build block themes in WordPress Studio. Here is what the workflow does, where it breaks, and why the theme.json output matters.
Central pathfinding blows up on sparse graphs. This covers distributed Q-Learning routing: one agent per node, the update rule, a QNode class, and what happened when it replaced Dijkstra on a real traffic project.
Most WooCommerce stores do not need a deep learning recommender. Catalog stability and how subjective taste is in your category decide the architecture, and GBDTs with cached scores usually win.