So, Can AI Kill Us All?; Apple’s Foldable iPhone; the Census’ Bad Data Problem
Summary
Wired discusses an Anthropic researcher’s resignation, arguing that AI labs are racing toward recursive self-improvement with misaligned incentives. The show also covers Apple’s first foldable iPhone, priced at $2,000, and new privacy concerns regarding its audio-recording features. Additionally, hosts analyze a flawed Census Bureau report on non-citizen voting, revealing political ties among its authors and raising fears about long-term data integrity.
Topics discussed
Serval: AI for automating IT support tickets
Dun & Bradstreet: Context layer for AI workflows
Lenovo: Eva Longoria business advice campaign
Show intro and episode agenda overview
AI researcher resigns from Anthic over safety concerns
Will Knight discusses AI safety and recursive self-improvement
Debate on AI doomerism vs. marketing hype
Concerns about AI agent misbehavior and corporate control
Journalistic approach to covering uncertain AI advancements
Apple's first foldable iPhone and $2000 price tag
Foldable phone market growth and developer opportunities
Census Bureau report on non-citizen voting data flaws
Political ties of Census report authors and data integrity
Long-term impact of faulty government data on trust
Superhuman Go: AI chat for browser productivity
404 Media Podcast: Tech news and surveillance
Dun & Bradstreet: Data confidence for AI agents
Wired/Tired: Apple Watch audio intelligence privacy concerns
Wired/Tired: Back to school and declining student reading scores
Credits and Financial Tea podcast promotion
Dun & Bradstreet and Lenovo closing ads
Listen ad-free on Castria