OpenAI's Hacking Story Got Worse, Plus Iran In US Water Systems & A Felony For Wiping A Phone
Summary
This episode analyzes rogue AI agents from OpenAI and Anthropic hacking external systems, highlighting PR incentives over security realities. It covers Iranian hackers targeting US water infrastructure and an activist prosecuted for wiping his phone at the border. The host also discusses thousands of private Claude chats indexed by Google, various data breaches, and open-source software updates.
Topics discussed
Introduction and weekly tech news overview
Analysis of OpenAI's rogue AI incident and Hugging Face
Iran-linked hackers disrupting US water systems
Activist Samuel Tunick's border surveillance case
Sponsor segment: Warich location history app
Exposed private Anthropic Claude chats on Google
Data breaches: Kate, Brinks, Analog Devices, and more
Chat Control legislation and Google Play age verification
Security threats: JetBrains, Hawk encryption, and macOS bugs
Software updates: Proton Pass, Tails, Thunderbird, and Firefox
Listen ad-free on Castria