“We put a lot of effort into real-time monitoring as these models are working, making sure you have good telemetry into what's happening and can intervene in real time”
♫
Said on Big Technology Podcast
“‘a monitor actually observes in real time an action an agent is doing, it decides to block it… would have helped and prevented the actual incident’”
Alan Chan, GovAI · "The Cognitive Revolution" · Governing AI Agents: Alan Chan of GovAI on Monitoring, Agent IDs, Rogue Agents & Preparing for RSI
Play from 7:40
7:40
Audio streams from the publisher’s own feed. Verbatim does not host or copy it.
What was described
advocated for synchronous real-time monitoring that observes each agent action and blocks it if a classifier flags it as bad
Said nearby on the same subject
“We set each monitor's threshold so that exactly 2% of honest held-out transcripts, pooled across all checkpoints, are flagged.”
♫
Said on LessWrong (Curated & Popular)
“It can do that all live and it even, while you're dictating, it classifies if you've spoken to a point where it can take an action.”
JL
John Lindquist
How I AI ·
“Employee took the phone but wouldn't monitor calls and didn't appear to face any formal disciplinary action as a result.”
AM
Alicia Middleton
ER/IR Unpacked ·
Moments like this, every Monday
The strongest new moments on your topics, in the speakers' own words. Free for three topics.
No card. No password. One-click unsubscribe.
Track a competitor?
Pro watches up to 25 names or phrases and emails you the morning after they are said.
See Pro, $29 a monthVerbatim quotes a short extract and links to the episode; the episode belongs to its publisher. Remove a show.