The Core AI team put out a Call for Testing for the AI Experiments plugin. Here is what is in it: MCP server support, markdown feeds, extra providers and AI request logging, plus the $400 API bill that convinced me logging is the part you cannot skip.
A million-token context window does not mean a model can reason across a million tokens. Recursive Language Models hand the context to a sandboxed Python REPL and let the model write code to partition and query it, which is how DSPy now implements the strategy.
Cron jobs and hope are not a data pipeline. Ahmad Wael walks through an ETL setup for crime trend data with Python, Prefect and PostgreSQL, including schema validation, automatic retries and a Metabase dashboard running in Docker.
WordPress 6.9 shipped the Abilities API, and the MCP Adapter is what lets AI agents find those abilities and call them. Ahmad Wael covers flagging an ability for MCP, the two transport methods, custom servers, and keeping permission callbacks tight.
A custom VS Code theme takes hours rather than weeks once you move past CSS variables and into TextMate tokens. Ahmad Wael on scaffolding the first pass, debugging scopes in the Extension Host, and picking contrast that does not fry your eyes by lunchtime.
Basic GNNs treat every SKU connection the same. Heterogeneous Graph Transformers weight each edge type separately, which cut WAPE from 0.86 to 0.58 on an FMCG dataset. The post also shows the typed-edge table I use instead of flat post meta.
The January 2026 Towards Data Science pieces I actually used: Ryan Pegoud cutting LLM memory 84% with Triton fused kernels, ACE context engineering, and Hugo Lu on the ceilings at Databricks and Snowflake.
Shipping a theme to the VS Code Marketplace and Open VSX took longer than writing it. Notes on Azure DevOps PAT tokens, uploading a .vsix by hand when the CLI login stalls, Eclipse Foundation namespaces, and why relative README image paths break in both registries.
RoPE rotates query and key vectors instead of adding an index, so the distance between two tokens stays the same wherever the sequence sits. Covers the rotation intuition, why dimension pairs rotate at different speeds, and a PyTorch implementation for a custom transformer.
Which control to reach for once a dropdown of 500 product variations stops being usable: combobox, multiselect, listbox or dual listbox. With a Gutenberg ComboboxControl refactor and the keyboard focus rules a real combobox has to meet.