~/talk

Why Are We Still Governing AI Agents Like Software?

Talkat AI Engineer NYC 2026 · 2026-10-12..10-14· [ai-assisted-development, security]
name:Tim Tyler
links:none on file
slides:none on file

Details of the Agent Accountability framework map exactly to what gets built: a registry that issues a single unique ID per agent, assignment into teams and a named human sponsor (example: Sven owns agent 409), alignment as a role versus runtime mission with least‑privilege authority (agent 203 limited to scopes C and D while 409 holds D E F), and administration via oversight scorecards, guardian agents that act at machine speed, and stepwise recourse (restrict, escalate, suspend) tied to confidence thresholds.

AI-written summary · about

  1. 0:00The job posting nobody would accept
  2. 1:15Agents as a digital workforce
  3. 2:27Assignment: identity, team and sponsor
  4. 6:00Alignment: role and mission
  5. 7:37Authority and lowest scope
  6. 9:34Never assume alignment
  7. 11:17Administration: oversight scorecards
  8. 13:35Agents need micromanagers
  9. 14:21Recourse when agents go wrong
  10. 17:05Recap
~/search
type to search 6,518 conferences, 143,918 speakers and 192,133 talks
~/help
cmdctrl k
search every conference, speaker and series
/
focus the filter search
h
home
c
conferences
s
speakers
e
series
t
light / dark — the status bar button does the same
← →
on a talk: 5 seconds back / forward
0–9
on a talk: jump to 0%–90% of the recording
?
this help
esc
close this, then clear every filter — or just the search, while you are typing in it

Ignored while you are typing. Pages without a filter rail have no search to focus.