Millions of books died so Claude could live - The Vergecast Recap

Podcast: The Vergecast

Published: 2026-02-03

Duration: 1 hr 29 min

Guests: Will Arimus, Julia Alexander

Summary

AI companies, including Anthropic, are acquiring vast amounts of book data, sometimes through questionable means, to train models. Legal challenges and copyright concerns are mounting as these practices come under scrutiny.

What Happened

Anthropic has been at the center of controversy for its Project Panama, which involved destructively scanning millions of books to train its AI model, Claude. Books, seen as high-quality content, were sourced from shadow libraries like LibGen and purchased in bulk from warehouses. The process involved slicing spines off books, a method that destroyed them but allowed efficient scanning. Legal challenges have arisen, resulting in Anthropic settling a lawsuit for $1.5 billion over its acquisition methods, though the legality of using book content for AI training remains unresolved.

Anthropic's actions are part of a larger trend among AI companies like OpenAI and Meta, who are pursuing vast data sets to improve their models. The pressure to stay competitive in AI development has led companies to take shortcuts, often raising ethical and legal questions. These practices are scrutinized through the lens of fair use, with court rulings offering mixed outcomes.

David Pierce is considering switching from iOS to Android and invites listener suggestions for phones to try. This personal anecdote opens up a broader discussion about the challenges of changing mobile ecosystems, especially from a consumer perspective.

Julia Alexander from Puck joins to discuss Netflix's potential interest in Warner Brothers and the implications for movie theaters. As Netflix considers theatrical releases to boost streaming performance, the conversation explores the changing dynamics of the film industry and the role of theaters in an increasingly digital age.

The episode also touches on the significance of IMAX screens and how movie studios are competing for these limited resources. With the average moviegoer attending fewer films, theaters are exploring alternative content and nostalgia-driven screenings to draw audiences back.

Smart home technology faces its own set of challenges, as discussed by Jen2E. Google Home's lack of support for certain Matter standards and the connectivity issues with IKEA's new smart home products highlight the growing pains in the smart home ecosystem.

David Pierce invites listeners to engage with the show, suggesting phones he should try and sharing their thoughts. This interactive element reflects the show's commitment to community engagement and audience involvement.

Key Insights

• Anthropic's Project Panama involved destructively scanning millions of books sourced from shadow libraries and bulk purchases, leading to a $1.5 billion lawsuit settlement over acquisition methods.
• AI companies like OpenAI and Meta are under pressure to acquire vast data sets for model improvement, often leading to legal and ethical challenges regarding fair use of content.
• Netflix's potential interest in Warner Brothers could lead to more theatrical releases as a strategy to enhance streaming performance, reflecting a shift in film industry dynamics.
• Smart home technology faces challenges with compatibility and connectivity, as seen with Google Home's lack of support for certain Matter standards and IKEA's new product issues.