16,000 Databases Expose Personal Data Online
Summary
This episode covers the US government's new AI chatbot for citizen services, which aims to simplify bureaucratic tasks like passport applications. It also discusses Nvidia's "Open Shell" platform for containing rogue AI agents and the use of LLMs to crack unsolved WWII Enigma codes. Additionally, the hosts analyze Anthropic's push for super-voting shares ahead of its IPO, ongoing legal battles with the Pentagon, and Tesla workers resisting the training of Optimus robots.
Topics discussed
Intro and AI Chat podcast promotion
Episode overview and AI Box AI sponsorship
Launch of America.gov AI agent for federal services
Nvidia OpenShell and industry self-regulation of AI
AI models cracking unsolved WWII Enigma codes
Anthropic founders seek super-voting shares pre-IPO
Appeals court upholds Pentagon ban on Anthropic
Tesla workers resist Optimus robot training
Supabase data leaks and vibe coding security PSA
Outro and final AI Box AI promotion
Listen ad-free on Castria