Alex Stamos: AI-automated hacking has plunged cybersecurity into a 'dark age'

The Verge AI · rss · 2026-11-07

On The Verge's Decoder podcast, Cognition CSO and Stanford lecturer Alex Stamos argues AI has upended cybersecurity.

The industry has been at "DEFCON 2" for a month: major PHP/Node bugs, successive breaches of AI neoclouds, and the Binance hack — the largest heist in human history.
Attacks now chain together: Fireworks breached → stolen API keys → Cursor compromised → malicious plugin pushed to 5,000 companies.
AI has inverted the kill chain: the open-source Chinese tool ARTEX, powered by a jailbroken GLM model, finds bugs and writes exploits in real time — hitting three South Korean banks first, then 200 more companies this week.
He compares it to the '90s Wild West: downloadable tools can now break into multibillion-dollar companies, but attackers are more numerous, often beyond law enforcement's reach, and no longer dependent on guarded frontier models.
Not a long-term doomer, but he expects a "dark age" while defenders adapt; the episode also touches on managing AI misinformation and slop on social networks.

More from AGI Musings