Themes: what the AI conversation is about

Every story the radar keeps is classified by the Jev model into themes — a story can carry several — and the themes form one tree. The numbers below are for Sep 27, 2026: the stories kept that day, their share of the day, the last 7 days and the week-over-week trend.

In the 7 days to Sep 27, 2026, 38,324 stories were classified and 18,782 kept (49%); a parent theme counts every story of its sub-themes, once.

• Models — 573 stories (34% of the day) · 6,757 in 7 days · rising week over week
• Agents — 598 stories (35% of the day) · 5,517 in 7 days · rising week over week
• Usage & adoption — 416 stories (24% of the day) · 3,441 in 7 days · steady week over week
• Protocols & infra — 611 stories (36% of the day) · 5,836 in 7 days · steady week over week
  • Identity & auth — 40 stories (2.3% of the day) · 435 in 7 days · steady week over week
  • Security & safety — 353 stories (21% of the day) · 3,029 in 7 days · steady week over week
  • Agent payments — 24 stories (1.4% of the day) · 341 in 7 days · rising week over week
    • x402 — 8 stories (0.5% of the day) · 58 in 7 days · rising week over week
    • MPP — 0 stories (0% of the day) · 1 in 7 days · fading week over week
    • AP2 — 1 story (0.1% of the day) · 3 in 7 days · steady week over week
    • ACP — 1 story (0.1% of the day) · 2 in 7 days · steady week over week
    • Card networks — 1 story (0.1% of the day) · 15 in 7 days · steady week over week
    • Stablecoins / crypto — 8 stories (0.5% of the day) · 55 in 7 days · steady week over week
    • PSPs & fintechs — 1 story (0.1% of the day) · 43 in 7 days · rising week over week
• Business & policy — 329 stories (19% of the day) · 4,854 in 7 days · steady week over week
  • Companies & business — 230 stories (13% of the day) · 3,690 in 7 days · steady week over week
  • Policy & macro — 191 stories (11% of the day) · 2,124 in 7 days · steady week over week
• SF scene — 4 stories (0.2% of the day) · 80 in 7 days · fading week over week