1995 one-liners gathered on the evening of 2026-09-08 from the free sources named on the coverage page, in two passes, one per side of the space. WHERE: Wikipedia's Current events, Deaths and On-this-day pages; Google News country editions (37 countries); outlet feeds across regions; GDELT by region. WHAT: Google News topic feeds in five editions, 60 sub-field searches on Google News (space, AI, chips, physics, books, film, each sport, rates, parenting …), and 45 topical outlet feeds (NASA, Space.com, Quanta, Ars Technica, Variety, ESPN, CNBC, SCMP …). Each feed is capped at 12 lines for readability; uncapped, the same sources give three to four thousand. No search fees, no model. The same story from several outlets is folded to one line by title overlap. This is what the meeting, or the pitch pass, would read; nothing here was commissioned.
| Source family | lines |
|---|---|
| Google News | 810 |
| Outlet | 524 |
| Google News search | 454 |
| GDELT | 143 |
| Wikipedia | 64 |
| total | 1995 |
Languages: en 1732, es 36, zh 36, ar 23, de 12, fr 12, it 12, pt 12, ja 12, ko 12, id 12, vi 12, th 12, tr 12. Gathering took 25 s for the region feeds and Wikipedia in parallel, 55 s for the topic pass, and about 3 min for GDELT, which allows one request every five seconds and answered 429 to bursts. Endpoints that failed on this run: 86 (transient SSL drops through the dev PC's tunnel, two dead feed URLs, rsshub refusing); each is logged and the harvest went on without it.
One tap opens a source. Non-English lines are left as they came; the meeting reads them as they are, and the desk's ladder would fetch and translate only what is commissioned.