~/talk

We Trust AI Agents That Break Every Day

Talkat AI Engineer NYC 2026 · 2026-10-12..10-14· [ai-assisted-development]
slides:none on file

Survey finds 80.8% daily use of agents but only 49.1% in production, with 85.5% saying they trust output while 41.1% see issues daily and 22% lose progress to crashes. A Temporal workflow demo moves a $2,500 payout using a local Quen 36B 9B model, records every step in durable execution, survives intentionally killed workers, preserves state, avoids rerunning LLM tool calls and therefore no re-spending of tokens, and resumes cleanly.

AI-written summary · about

  1. 0:00Intro
  2. 0:34Finance runs on long-running workflows
  3. 2:05The State of Development survey
  4. 3:16Gap 1: Usage vs. production
  5. 4:33Gap 2: Trust vs. reality
  6. 7:18Gap 3: Prototype vs. enterprise
  7. 8:55Gap 4: Seeing vs. recovering
  8. 10:53Demo: a payout that survives a crash
  9. 13:06Durable execution explained
  10. 18:12What the demo solved
  11. 20:10Who runs this in production
  12. 20:50Wrap-up
~/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.