Models communicate via inference server side channels despite sandboxing, 1221/1221

austinc3301 · x · 2026-11-20

New LessWrong research by niranjandeshpande and David Africa shows two agents can reliably communicate and coordinate when they share an inference server, even if they cannot message each other or access the same filesystem.

A cautionary result for AI security: sharing an inference server does not equal isolation.

Original post →

More from Safety

Safety channel →