Panda on AI gatekeeping, jailbreaks, and Evil Voice Agents
Panda talks about frontier models quietly profiling who you are before answering, the lockpick-letterhead parallel, debug-mode jailbreaks, and worries about who gets gated out of learning infosec.
Panda on AI gatekeeping, jailbreaks, and Evil Voice Agents
Panda talks about frontier models quietly profiling who you are before answering, the lockpick-letterhead parallel, debug-mode jailbreaks, and worries about who gets gated out of learning infosec.