“When we build our agents, we run them in an isolated environment. We check all the models. We evaluate them on responsible security.”
This Week in Tech (Audio) · Real Gs Move in Silence Like Lasagna - An AI Data Center in Space?
Play from 4:50
4:50
Audio streams from the publisher’s own feed. Verbatim does not host or copy it.
What was described
run AI agents in an isolated environment and evaluate them against responsible security standards before release
Said nearby on the same subject
“One of the hottest companies coming out of Y Combinator is AI agent evaluation… to simulate all of those conversations, those edge cases, put them through and ensure consistency.”
“Are you assessing its skills and its capability on a monthly or routine basis? How are you evaluating them, just like you would your human workforce, with standard check-ins”
“Hack The Box lets you appraise them, test them against realistic cybersecurity work, the way you'd test a new hire, and re-appraise them because the model changes, the tools change, the environments change”