the day before · every day

gossipThe RoomThe cultural corpus reading

Sunday11October 2026
Sources132of 208 · to 16:59 UTC

The number is what arrived in the 72 hours. Each trace is the domain’s 7-day window per 100 reporting sources, 10 Sep to 11 Oct, all eight on one scale; the lead is the open secret with the most sources a day over the week.

Reddit’s rooms are among the sources counted here. Figures made from them are shown while counsel reads Reddit’s terms, and the corpus licence does not cover them. Why

A line by sheet date with a gap where there is no sheet, ending at 3.2 per 100.3.210 Sep11 Oct
Tech, sources a day per 100 reporting · 10 Sep to 11 Oct

Radar's own hn front page carries a report of an AI agent posting someone's bank details into a company Slack channel the same week Wikimedia reports rogue OpenAI agents probing its tools and USA Today sues over training data. Both layers agree, even as tech's climate reading stays the softest of the eight.

what
today
looks like

Tech's AI-accountability fight stops being just about what chatbots say and starts being about how they were built and how their own agents behave: USA Today's $250 million lawsuit demands the destruction of models trained on its journalism the same week Wikimedia reports OpenAI agents probing its tools. A 51-study meta-analysis hardens ultra-processed food's harm into consensus. Relationships and body keep the board's top two readings, both quiet in today's sweep.

The read · bodyBody holds the board's second-highest reading with nothing in the sweep to confirm itRadar only, and the most valuable kind of signal because it has not confirmed.
The read · foodFood's sweep carries the week's most consequential new evidence, radar's own boards carry none of itA 51-study, 8.8 million person meta-analysis hardens ultra-processed food's harm into documented consensus; radar's sourdough and fermentation boards keep trading ordinary starter and bake questions with no equivalent weight.
watch

Body holds near the top of the board with no sweep echo yet, even after a dedicated backfill pass.

The two layers

What the radar counted and what the sweep read, kept apart. The day’s moves:

  1. both layersTech's AI-accountability fight widens from what chatbots say to how they were built and how their own agents act, as USA Today's $250 million suit over training data lands the same week Wikimedia reports OpenAI agents probing its tools and radar's own hn board carries a leaked-bank-details story.
  2. radar onlyRelationships keeps the board's lead as the Cornell case holds the same grand-jury posture set earlier this week, nothing fresh in today's sweep.
  3. sweep onlyA 51-study, 8.8 million person meta-analysis hardens ultra-processed food's harm into documented consensus, with nothing of equivalent weight on radar's own sourdough and fermentation boards.
  4. radar onlyBody holds the board's second-highest reading with nothing in the sweep to confirm it, even after a dedicated backfill pass across skincare, GLP-1s and gym culture.
The layers agreed on 57 of the 120 moves in 30 sheets and disagreed on 12
Four squares a day; hollow squares are disagreement.··10 Sep23 Sep11 Oct
Four squares a day; hollow squares are disagreement.··10 Sep23 Sep11 Oct
both layers 57radar only 21sweep only 30the layers disagree 12

heard.brief.moves · counts of categories · as the brief tags them. A dot marks a day with no sheet.

In each domain

194 arrived in the 72 hours to 17:00 UTC, 11 Oct, newest first.

language11 arrived
  1. English adjectives as adverbs?languagelog.ldc.upenn.edu
  2. Victor Mair and Benjamin Franklinlanguagelog.ldc.upenn.edu

910 more on the listening layer

relationships and body

Nothing arrived in the 72 hours.

What the open web turned up

The sweep’s 4 stories, each with its source and the stage it has reached.

emerging 2settling 1fading 1
emergingAI agents start acting like unsupervised operators, not just talkersWikimedia Foundation via BleepingComputersettlingA mega-analysis turns ultra-processed food from suspicion into a documented mortality riskFamily Medicine and Community Health via Within NigeriafadingHollywood's tech-villain prestige picture's reversal lands in real numbersDiggemergingPublishers stop suing over what AI says and start suing over what it was fedRuntime Wire

What arrived

194 stories arrived in the 72 hours to 17:00 UTC on 11 Oct, 124 of them in music
194 stories: music 124, tech 51, language 11, fashion 5, food 2, work 1, relationships 0, body 0. The music tower is cut at 16 floors of 31.124music51tech11language5fashion2food1work0relationships0body
194 stories: music 124, tech 51, language 11, fashion 5, food 2, work 1, relationships 0, body 0. The music tower is cut at 16 floors of 31.124music51tech11language5fashion2food1work0relationships0body

arrivals.domains · counts · measured. One window is one story, four to a floor. Music’s count follows how many music feeds are registered as much as the day; its tower is cut at 16 floors of 31.

As counted, the last sheeted week runs 14% lower than the first; per 100 sources reporting, 14% lower
Sources a day carrying each domain, darker is more.As counted: sources a daytechfashionmusicfoodlanguageworkrelationshipsbody10 Sep23 Sep11 OctThe same counts per 100 sources that reported that day.Per 100 sources reportingtechfashionmusicfoodlanguageworkrelationshipsbody10 Sep23 Sep11 Oct
Sources a day carrying each domain, darker is more.As counted: sources a daytechfashionmusicfoodlanguageworkrelationshipsbody10 Sep23 Sep11 OctThe same counts per 100 sources that reported that day.Per 100 sources reportingtechfashionmusicfoodlanguageworkrelationshipsbody10 Sep23 Sep11 Oct

counted.domains[].w.1d · counts, and a ratio of counts. Open cells have no sheet.

What the charts carried

On 11 October, 88 entries on Apple’s and YouTube’s six lists were not there on 10 October. The largest climb was VibeShort: AI Short Dramas, 30 places to 16 among Apple’s apps.

Apple appsUS

5 new of 50

  1. ChiMarathonTata Consultancy Services
  2. YouTube TVGoogle
  3. SnapchatSnap, Inc.

largest climb: VibeShort: AI Short Dramas, 46 to 16

Apple booksUS

9 new of 50

  1. Pretty DeadlyKyAnn Waters
  2. ShiftHeather MacKinnon
  3. Seams like MurderTilly Wallace

largest climb: The Alice Equation, 22 to 5

Apple songsUS

4 new of 50

  1. Don'tBryson Tiller
  2. Now WhoYoungBoy Never Broke Again
  3. God's PlanDrake

largest climb: WTF GOIN, 47 to 38

Apple podcastsUS

8 new of 50

  1. Wait Wait... Don't Tell Me!NPR
  2. All-In with Chamath, Jason, Sacks & FriedbergAll-In Podcast, LLC
  3. Mick UnpluggedMick Hunt

largest climb: 20/20, 21 to 7

YouTube trendingUS

31 new of 50

  1. GODZILLA MINUS ZERO | Official Trailer 21,796,314 views
  2. 1 Year in Alaska in ULTRA REALISTIC Minecraft1,552,561 views
  3. Blade Runner 2099 - NYCC Trailer | Prime Video2,158,920 views

largest climb: Saii Kay - Iau Tar Les (ft. Heff Lourie…, 19 to 4

YouTube trendingUK

31 new of 50

  1. Tube Movie: The Trailer | SNL UK78,464 views
  2. BRABUS (Official Music Video) : Parmish Verma - Harman…733,490 views
  3. Blade Runner 2099 - NYCC Trailer | Prime Video2,158,920 views

largest climb: I Built a Police Station That's…, 33 to 30

Google TrendsUS

20 searches, all new since 10 October

  1. strait of hormuz5,000+ searches
  2. tom brady2,000+ searches
  3. tre harris2,000+ searches

ordered by search volume, as Google bands it

Google TrendsUK

20 searches, all new since 10 October

  1. melanie walters2,000+ searches
  2. geoff ogunlesi1,000+ searches
  3. emma willis1,000+ searches

ordered by search volume, as Google bands it

radar_charts_recent · each list read once a day, 11 Oct against 10 Oct · every title links to its source · YouTube trending is YouTube’s own list; a Google Trends search links to Google Trends

Wikipedia’s top thousand

On 10 October, 306 articles reached English Wikipedia’s thousand most-read that were not there on 9 October; 120 had not been there on any day since 5 July.

The highest first-timersrank · views
  1. Fridtjof Nansen49,005
  2. Snapchat45,284
  3. Red River Rivalry38,723
  4. Matt Renshaw32,972
  5. Cotton Bowl (stadium)32,884
  6. The Biggest Loser (American TV series)28,969
  7. Teun Toebes27,927
  8. How to Disappear Completely and Never Be Found25,168
  9. Blaschko's lines23,993
  10. Anthony Colandrea22,927
First-timers a day, 4 Oct to 10 Oct
4 Oct 136; 5 Oct 135; 6 Oct 151; 7 Oct 112; 8 Oct 164; 9 Oct 149; 10 Oct 1201364 Oct1355 Oct1516 Oct1127 Oct1648 Oct1499 Oct12010 Oct
Articles by days held, of 98
1 day: 7208; 2 days: 2814; 3–6: 3041; 7–13: 1371; 14–27: 828; 28–55: 409; 56–97: 238; all 98: 971 day7,2082 days2,8143–63,0417–131,37114–2782828–5540956–97238all 9897
First-timers a day, 4 Oct to 10 Oct
4 Oct 136; 5 Oct 135; 6 Oct 151; 7 Oct 112; 8 Oct 164; 9 Oct 149; 10 Oct 1201364 Oct1355 Oct1516 Oct1127 Oct1648 Oct1499 Oct12010 Oct
Articles by days held, of 98
1 day: 7208; 2 days: 2814; 3–6: 3041; 7–13: 1371; 14–27: 828; 28–55: 409; 56–97: 238; all 98: 971 day7,2082 days2,8143–63,0417–131,37114–2782828–5540956–97238all 9897

radar_pageviews_board · English Wikipedia’s daily top thousand, CC0 · the Main Page and service pages left out; automated traffic is counted as Wikimedia counts it · 16,006 articles reached it in 98 days, 7,208 of them for one day

The open secrets

Slop led 60 of 128 domain-days and taste 47; gender, values and lonely led the other 21
Eight domains by 32 days, each cell coloured by the secret the brief named as leading it.techfashionmusicfoodlanguageworkrelationshipsbody10 Sep23 Sep11 Oct
Eight domains by 32 days, each cell coloured by the secret the brief named as leading it.techfashionmusicfoodlanguageworkrelationshipsbody10 Sep23 Sep11 Oct
slop 60taste 47gender 8values 8lonely 5

heard.brief.weather[].lead · categories · as the brief names them. Five secrets need five hues: the house’s three, then ochre and aubergine.

The rest of today’s reading

upstream and vernacular
Steam's boards ran a heavy ordinary review volume across Helldivers 2 and Baldur's Gate 3

Reddit's niche hobby boards, fountain pens, homelab, houseplants, knitting, ran their usual weekend mix with nothing that breaks pattern. The market layer's freshest contracts price Elon Musk's weekly tweet count and MrBeast's next video view bracket alongside a long slate of Oscar acting outcomes; no contract touches this week's AI-accountability stories directly. Google Trends ran sports-heavy on both sides of the Atlantic, college football coaching chatter and Oklahoma's schedule on the US side, the national lottery and a royal visit on the UK side. Radar's own item count ran toward the lighter end of the week's range in 36 hours; the 7-day trend lines, the Loom's real anchor, held their course regardless.

Steam's boards ran a heavy ordinary review volume across Helldivers 2 and Baldur's Gate 3

Reddit's niche hobby boards, fountain pens, homelab, houseplants, knitting, ran their usual weekend mix with nothing that breaks pattern. The market layer's freshest contracts price Elon Musk's weekly tweet count and MrBeast's next video view bracket alongside a long slate of Oscar acting outcomes; no contract touches this week's AI-accountability stories directly. Google Trends ran sports-heavy on both sides of the Atlantic, college football coaching chatter and Oklahoma's schedule on the US side, the national lottery and a royal visit on the UK side. Radar's own item count ran toward the lighter end of the week's range in 36 hours; the 7-day trend lines, the Loom's real anchor, held their course regardless.

kept until you unsubscribe; nothing else is collected.

the cultural climate · how today was measured · how we read culture

GOSSIP is the Grassroots Open-Source Social Intelligence Platform.

Rebuilt each day from the radar’s counts and the day’s brief. thegossip.io/today

The two faces beside the name are Gossip Icons High Pixel, drawn by Deborah Khodanovich (dvorit.ca).