A jailbreak turned a consumer chatbot into a government hacking assistant, a robot vacuum became a global spycam, and the tools developers trust were caught with zero-click flaws of their own. This episode is the collision between agentic speed and security reality.

AI just got a lot faster—and a lot more autonomous. In this episode, we break down Mercury 2’s diffusion-speed leap, Alibaba’s open-weight giant, and the rise of agents with wallets, identities, and real machine-to-machine payments.

AI coding just went fully local, mobile, and dangerously powerful. In this episode, we unpack the rise of sovereign agents—from remote control coding and compressed frontier models to the terrifying new Sandworm supply-chain attack that turns your own agent into an insider threat.

A single AI blog post nukes a blue-chip: IBM drops 13% as Claude Code targets the COBOL moat. Then Anthropic alleges an industrial-scale distillation operation—24,000 fake accounts and “Hydra clusters.” We break down the market shock, the spycraft, and why the web’s infrastructure is getting rewritten next.

A blog post triggered a Friday flash crash in cybersecurity stocks—and it might have also ended vibe coding. This episode breaks down why Claude Code Security spooked Wall Street, how agentic engineering is replacing “just ship the vibes,” and what comes next when AI can plan, test, and patch faster than humans can read.

Three model drops. One OAuth crackdown. And a wave of “zero-click” agent exploits. In Episode 2, we map the SaaSpocalypse, the Harness Wars, and the security nightmares that turned vibe coding into a liability overnight.