How these stories get written

Grade One Press reports on technology through primary source documents. Most of what is published in News and the Brief begins with an automated watch on filings, dockets and official registries, followed by model drafting and human review.

How the stories are drafted and checked

Most of what’s published in News and the Brief starts the same way: a script built for this site watches a fixed list of primary sources — company filings, court dockets, regulatory releases, rule changes — for anything new. When one changes, that document is the input.

From there, an AI model drafts the story from that document. Not from a press release summary, not from another outlet’s coverage of the same event — the primary document itself. Before anything publishes, the draft is checked against that document: does the summary match what it actually says, does the grade on the source match the Standard page’s definitions. A story does not go live until that check happens.

This is not AI-assisted editing in the sense of a spell-check pass on something a person wrote start to finish. The draft comes from the model. Describing it any more softly than that would be exactly the kind of thing this site’s own grading system exists to catch. The byline says Grade One Newsroom because that is literally what produced the sentences. The grade next to the source is still the thing worth trusting — that part doesn’t come from AI, it comes from checking the primary document, same as always.

What this means in practice:

• Every News story and Brief item links to the one primary document it’s traced to.
• AI drafts it; a person checks the draft against that document before it publishes.
• Reports and Findings go through more source-gathering than a single-document News story — see the standard.
• A story found to misstate a source gets corrected in place, dated, visible — the same as every other correction here (see corrections).

Editorial scope and volume

Stories are published when verified, not to satisfy an arbitrary daily target.

We do not stop at four stories a day simply because a quota was reached, nor do we publish routine marketing announcements to keep a feed active. If an event is supported by a verified, high-grade primary source (Grade 1 or Grade 2) and represents a material, verifiable event that matters to the technology community — statutory workforce filings, regulatory antitrust orders, critical vulnerability directives, core infrastructure failures, or empirical model benchmarks — it gets published. If nothing clears that evidentiary bar on a given day, we do not pad the record with commentary.