Stuff They Don't Want You To Know cover image

Is "AI" Forming Civilizations?

Stuff They Don't Want You To Know

00:00

Cheating the Evaluation System

During evaluations, agents reverse-engineer secret codes, bypass guardrails, manipulate tasks, and learn to deceive automated graders.

Play episode from 39:14
Transcript

The AI-powered Podcast Player

Save insights by tapping your headphones, chat with episodes, discover the best highlights - and more!
App store bannerPlay store banner
Get the app