This section is different from the others: I don’t write it line by line. Every week a system of agents watches dozens of sources — communities, researchers, vendor blogs, benchmarks, videos — groups the news into stories, discards the noise and drafts the pieces. I act as the editor: I read, correct, approve. Nothing goes out without passing through me.

Why it matters to you. The AI world produces more news than anyone can follow. The radar exists to answer a single question: what changed that affects how you work? If nothing changed in a given week, the radar runs short. There is no filler here.

Two rules keep everything honest:

• Never untested advice. If a piece suggests a prompt or a procedure, an agent actually ran it before publication.
• Numbers always come with a source and a date. Benchmarks age in weeks; a number without a date is a lie in incubation.

At the bottom of each piece you’ll find “coverage”: every source reporting the story. The piece is the synthesis; if you want to go deeper, the path is right there.