80+ sources. 15 minutes. What matters in AI today. Here is exactly how each episode is built.
01 · How episodes are structured
Six sections, one continuous narrative
Each daily episode runs 14 to 18 minutes as a single continuous narrative across the six editorial sections. No segments, no jingles between sections, no filler. The pillars themselves are listed on the home page.
Saturday weekly recap.The Saturday episode steps back from the daily news to identify the five trends that defined the week and flag what is likely to matter next week.
02 · How sources are selected and weighted
80+ sources, weighted by editorial standards
40+ is a minimum threshold, not a target. Each episode draws from 80+ monitored global sources. A minimum of 40 articles — scored and filtered — are selected per episode. Sources are not treated equally. Each carries a weight score based on editorial standards, domain authority, and primary-source proximity.
| Tier | Weight | What it covers |
|---|---|---|
| Primary / Official | 2.0× | Official announcements from AI labs and companies; peer-reviewed research and academic preprint archives. |
| Tier-1 Press | 1.5× | Major international newswires, financial press, and peer-reviewed science journals with original reporting. |
| Specialist Tech Press | 1.0× | Dedicated industry and technology publications with editorial oversight. |
Each episode's source manifest, with the actual domains and articles used, is published alongside the episode page.
03 · What we exclude, and why
- Regulatory speculation and policy rumors. Unverified regulatory moves produce noise that is almost always revised. We wait for official filings.
- Unverified announcements without independent corroboration. If only one source carries a claim, it doesn't air.
- Opinion and editorial commentary. AI Press Review reports what happened, not what commentators think about it. Interpretation is yours.
- Funding rounds below $50M and pre-seed announcements. Capital events dominate AI coverage and crowd out operational signal.
- Engagement-optimized content. Social media posts, thought-leadership, and content farms are excluded regardless of virality.
04 · About the host
DP05 · Feedback
Feedback is welcome
If you spot a factual error, a source that should be excluded, or a domain underrepresented in coverage, write to david.perron@aequitus.net.