I ran the WordPress.com AI engine at 21 site types, from portfolios to membership hubs, to see whether the blocks it generates survive a mobile viewport and a Core Web Vitals pass. Here is what ships as is, and the parts I still fix in functions.php.
The WordPress handbook now has a section on AI. Five rules cover who stays accountable, when you disclose model help, GPL compatibility, which contributions are in scope, and the quality bar reviewers can enforce.
Tammy Silva’s LinkedIn side project turned into 50 messages a day. She used an AI website builder to put up a WordPress hub in an afternoon, moved lead capture off DMs, and filled 10 coaching slots in two days. Includes the hook I use to route inbound leads.
Siloed training data produces models that fail confidently. This walks through the Federated Learning Flower framework on a biased MNIST setup, the pyproject.toml config, and the jump from 65% to 96% accuracy.
Prompting an LLM for plugin and WooCommerce code ships fast, then leaves you debugging plumbing you never wrote. Where vibe coding falls apart in WordPress, and what the vetted version of the same function looks like.
WordPress now has a repository for AI agent skills, starting with wp-playground. It uses the Playground CLI so an agent can mount a plugin or theme, wait for WordPress to be ready, log in, and check its own work in seconds.
Raw JSON in an embedding model spends your token budget on braces and quotes. Flattening product data into plain sentences gives the model the kind of text it was trained to read.
A notebook demo is not production. This is what sits between the two: an API contract, a container, infrastructure that suits the workload, and monitoring that catches drift before the predictions quietly go bad.
WordPress Studio 1.7.0 ships a CLI that can create, start, stop and delete local sites, and run WP-CLI inside the Studio environment. What the studio site and studio wp commands do, plus how to let a coding agent use them without losing a site.
Metadata filters only take music discovery so far. How CNNs read audio as Mel-spectrograms, how contrastive learning with InfoNCE loss shapes the embeddings, and why the inference belongs outside WordPress.