Hacking AI: The Jade Puffer Attack, Prompt Injection, and the Dead Internet Theory

In this episode, Sam sits down with Glenn Wilkinson, a long-time ethical hacker, keynote speaker, and CEO of Agger Labs. Glenn shares his journey from a curious teenager in Zimbabwe bypassing game demo protections to snooping on audience phones for a Daniel Radcliffe stage show.

The conversation dives deep into the evolution of ransomware and how Agger Labs is combatting novel threats. Glenn explains their approach of using local, deterministic deception technology—laying traps in the file system for ransomware to trip over—alongside robust anti-tampering measures. They also break down the recent "Jade Puffer" attack, highlighting how AI is now being used to orchestrate the hacker tradecraft of reconnaissance and lateral movement by exploiting unpatched, year-old vulnerabilities.

Addressing the anxiety around artificial intelligence, Glenn offers a grounded perspective: AI will replace tools, not necessarily people. He discusses why the future belongs to the "AI Wrangler" and provides a hacker's insight into prompt engineering. Glenn explains why bypassing AI guardrails is currently a game of "whack-a-mole," comparing it to early SQL injection because data and commands share the same channel.

Finally, they debate the technical realities of banning social media for teenagers, with Glenn arguing that bans simply drive kids to unregulated, darker corners of the internet. The episode wraps up with Glenn's 2030 prediction: a societal pushback against big tech, driven by the "dead internet theory" and a desire for local, offline communities.

Topics covered: the hacker mindset, ransomware defense and deception technology, the Jade Puffer AI attack, cybersecurity fundamentals, AI replacing tools vs. jobs, prompt injection and bypassing LLM guardrails, the proposed UK social media ban for minors, dumb phones, and the dead internet theory.