Random splits flatter models trained on WooCommerce order data, because the same customer leaks into both sides of the split. Here is how grouped and time-based validation give honest scores, and why per-segment error belongs next to every prediction.
WordPress.org now holds new plugin releases for up to 24 hours before auto-updates ship them, so an AI reviewer can scan the diff first. That changes which plugins you should trust with automatic updates, and which ones still need a human.
Fine-tuning a forecasting model on sales history alone bought less than a point of WAPE in the source runs. Feeding it known-future covariates first cut error from 8.4 to 2.8 percent. The adapter is the small lever here; the inputs are the big one.
WordCamp Europe 2026 wrapped in Kraków with 2,458 attendees, CERN’s flagship site now on WordPress, and AI plumbing landing in core with 7.0. What I’d copy from CERN’s platform, what the AI client means for client sites, and the question I left with.
On-policy learning improves the behavior you are already running, off-policy learning lets you behave one way and learn about another. That choice decides whether you can train on your store’s logged data or need fresh traffic, and why the winning variant from a noisy test usually looks worse after you ship it.
The case for letting a framework generate and score your prompts, and the part nobody mentions for WordPress: the optimization runs offline, and the prompt ships into your plugin as a plain PHP constant. Notes on when that’s worth the setup and when it isn’t.
An AI assistant that can’t see your project turns every review into a copy-paste loop. What I’d build myself, what I’d leave to the official SDK, and where the sandbox boundary belongs when the root contains wp-config.php.
Most teams using AI on a WordPress site pay a hidden tax in copy, paste and rewritten prompts, and another SaaS rarely removes it. The fix, for me, is moving the prompt into your own code with wp_remote_post, WP-CLI, caching and a small model, so the workflow runs where the data already lives.
The Core AI Team runs a table at WordCamp Europe’s Contributor Day on June 4. Before you go, two decisions matter: which environment you bring and where your model lives. This is the setup order I’d use, starting with one command the night before.
A hundred lines of Python, no vector database, and a real PDF. That is the baseline this author ships, with answers that cite exact lines. The part worth borrowing for WordPress work is the answer schema: typed JSON, verifiable quotes, and a null path that refuses to invent.