Anthropic is Introducing Text Watermarking… But How? - DTNS 5329
Summary
Hosts Jason Howell and Tom Merritt discuss Anthropic’s global rollout of text watermarking for Claude to comply with the EU AI Act, debating its technical implementation and potential impact on user adoption. The episode also covers a Lenovo patent for a horizontally expanding laptop screen, leaked iOS 27 code hinting at a split iPhone 18 launch, and OpenAI’s new security-focused GPT models. Briefs include Google’s buttonless homepage test, Meta smart glasses banned from UK courts, and Spotify’s upcoming AI artist badges.
Topics discussed
Sponsors and Intro
Anthropic's AI Text Watermarking
Sponsors: Capital One, Sephora, Bombas
Lenovo's Expandable Laptop Screens
iPhone 18 Leaks and Split Launch
OpenAI's Daybreak Cybersecurity Program
Sponsors: Capital One, Cologuard, Morning Brew
Google Homepage Test
Bumble Allows Men to Message First
Smart Glasses Recording Concerns
Claude Solves Math Problem
Section 230 Court Ruling
Spotify AI Artist Badges
Bluetooth Tracking Privacy
Listener Perspective on Smart Glasses
Outro and Sponsor
Listen ad-free on Castria