01 What we cover
We follow changes in AI models, products, developer tools, and practical uses of AI. We prioritize developments that affect capabilities, cost, availability, terms of use, or work processes.
An announcement alone does not guarantee coverage. We look for what actually changed and who may need to know about it.
02 Sources and claims
We rely primarily on first-party sources, such as official announcements, documentation, and reports from the organizations describing a change. Every story links to its source.
Company statements and results from company-run tests are attributed to the company. A short summary is not a substitute for documentation or independent confirmation.
03 How stories are prepared
AI models and automated checks help select, write, translate, and classify stories. A separate model review compares important claims with the source and looks for material factual discrepancies.
A story may be delayed if its primary source is unavailable, evidence conflicts, or a review cannot be completed. Automated checks can miss nuance and do not guarantee that every story is error-free.
04 Weekly and monthly reviews
Weekly reviews highlight three major developments; monthly reviews highlight five. They draw on published stories and link to sources so readers can check the details.
These are editorial selections, not exhaustive timelines. We aim to explain what is practically available and distinguish a source's claims from limited or independently confirmed results.
05 Review and corrections
Published stories receive recurring automated checks. A confirmed, material error is corrected and checked again; a suggestion without enough evidence is not treated as an established error.
Automated and model reviews are not a complete human fact-check. If you spot an inaccuracy, please email us with the story link and a supporting source.