Andrew Tandy— /lab/docket
PRJ-03 · Docket[pilot]
The County AI Policy Board's masthead and the "Where each board stands" table, showing all seven governing bodies and how much of the AI-policy record has surfaced for each.
1 / 2
The board's masthead and where-things-stand table.2026-09-20

Docket runs a weekly sweep of San Diego County K-12 sources — district and county board agendas, meeting video, state guidance — looking for anything touching AI policy. Matches get deduplicated against what it’s already seen and folded into one digest email, so the same document never surfaces twice.

Docket itself doesn’t publish anything; it’s a private collection system. What it feeds is the County AI Policy Board: a public record of where each governing body stands, pulled from documents that have been read in full, with what the record doesn’t yet answer listed openly rather than guessed at.

  • 2026-09-20 Published the County AI Policy Board at /ai-in-education/board: a public snapshot of where each of the county's 7 governing bodies stands, the policy threads in motion, the calendar, and open questions the record doesn't yet answer. Not a live mirror — it reflects a point in time and gets republished, not auto-updated.
  • 2026-09-16 First digest built end-to-end — fetch, keyword match, dedup, and a rendered HTML digest archived to disk. Tier 1 covers 9 of 11 planned sources; two (a BoardDocs district and the CA Legislature bill search) stay manual-only because their platforms disallow automated access.
  • 2026-09-16 Dedup verified — running the sweep twice in a row produces zero new hits on the second pass. Email sending itself is still unconfirmed against a real mailbox.