San Francisco
Menu

Kojo Minta

Member, San Francisco Treasure Island Development Authority

We identified Minta’s voice at four meetings.

Next Treasure Island Development Authority meeting: Fri, Nov 6, 10:00 am · Clerk’s calendar

Questions

Minta asked four questions at the four Treasure Island Development Authority meetings where we identified them

At least this many: direct questions only.

Neighborhoods Minta named

One place Minta named in meetings

Lines where Minta named the place, read by two AI readers that both agree it means that place. Each dot plays the newest one; brighter means named more often.

San Francisco
  • Treasure Island 1 time

Places are drawn at the centre of their own boundary in the city’s Analysis Neighborhoods file (DataSF (City and County of San Francisco)). A dot says the place was named, not where Minta was or what was said about it.

Meetings

Every meeting where we identified Minta’s voice

Four, newest first. The date opens the meeting; the time plays the video at Minta’s first remark.

  1. Jul 8, 2026Treasure Island Development AuthorityParks and recreation3.4 min33:28
  2. Jun 10, 2026Treasure Island Development AuthorityPublic transit2.5 min21:57
  3. Jun 5, 2026Treasure Island Development Authority0.5 min31:50
  4. May 13, 2026Treasure Island Development Authority0.9 min4:44
How we identify who is speaking · blind check, pooled across five cities: of 340 commission lines drawn, 260 right, 4 wrong, 1 unclear, 75 couldn’t be judged

A meeting missing here means we couldn’t attribute the voice, not that Minta was absent. Claude and Gemini each read every transcript with that day’s roster. A voice gets Kojo Minta’s name only when both agree and a code check finds the name said on, or just before, their turn. Voices our transcripts merge with someone else’s are never attributed, so every count here undercounts what Minta said. For commissions, of 340 lines drawn, 260 were right, 4 wrong and 1 unclear, and the checkers couldn’t judge 75. We have since removed the voices behind 64 of those lines (3 wrong, 0 unclear, the rest we had already held for other reasons). Of the 276 lines we still publish, the checkers could judge 228: 226 right, 1 wrong and 1 unclear (90% lower bound 0.97). They couldn’t judge the other 48. The check is one sample pooled across the five cities, not of Minta or of San Francisco alone.