Editorial illustration for Claude's story: millions of books died, while Netflix pursues Warner Bros.
Claude's Training: Millions of Books Sacrificed for AI
Claude's story: millions of books died, while Netflix pursues Warner Bros.
The piece opens with a stark accounting: the training of Anthropic’s Claude model reportedly led to “millions of books died,” a phrase that reads like an obituary for the raw text that fed the system’s language abilities. It’s a reminder that today’s most advanced LLMs are built on massive, often uncredited, cultural archives, and that the value extracted from those archives is rarely transparent. At the same time, the entertainment industry is staging its own high‑stakes drama.
Netflix, fresh off a string of bold statements about supporting movie theaters and big‑budget filmmaking, is quietly circling a potential acquisition of Warner Bros. The juxtaposition feels intentional—one story about the hidden costs of AI, the other about a streaming giant’s overt ambitions in Hollywood. Why does this matter?
Because the two narratives intersect on the question of who profits from cultural content and how that profit is justified. The next segment brings Puck’s Julia Alexander onto the show to unpack Netflix’s positioning, just as the company’s acquisition talk intensifies.
After that, Puck's Julia Alexander joins the show to talk Netflix and the movies. As it continues to try to buy Warner Bros., Netflix is saying all the right things about movie theaters, big-budget films, and generally caring about Hollywood. But of course it is -- it's trying to buy Warner Bros.!
Once this all shakes out, can theaters survive in an increasingly Netflix-ified world? Do people even want to see Netflix movies in theaters? And if not, what do we do with these big buildings and their big screens?
Finally, The Verge's Jennifer Pattison Tuohy answers a question from the Vergecast Hotline (call 866-VERGE11 or email [email protected]!) about Ikea's new smart buttons and why they don't seem to work very well. In terrific news, one part of our conversation is already outdated: since we recorded this on Monday morning, Google announced button support for Google Home! But the problems with Ikea's exciting new low-cost smart gear still persist.
If you want to know more about everything we discuss in this episode, here are some links to get you started: - From The Washington Post: Anthropic 'destructively' scanned millions of books to build Claude - Anthropic wins a major fair use victory for AI -- but it's still in trouble for stealing books - Meta's AI copyright win comes with a warning about fair use - Did AI companies win a fight with authors? Technically - From Puck: Why Netflix Needs Warner Bros.
Is the AI surge sustainable? The launch of ChatGPT in November 2022 ignited a race that quickly engulfed the tech sector, pushing state‑of‑the‑art research out of isolated labs and into mainstream products. Millions of books vanished, according to the piece, to make room for Claude’s emergence—a stark illustration of the speed at which content pipelines are being reshaped.
Companies scrambled, fearing they would fall behind in what the article hints may be the most critical technology battle today. Meanwhile, Netflix’s ambitions extend beyond streaming. As Julia Alexander noted, the platform is courting Hollywood, repeatedly emphasizing support for theaters and big‑budget films while simultaneously pursuing a purchase of Warner Bros.
The dual narrative—AI’s rapid commodification and a media giant’s aggressive acquisition strategy—highlights a period of intense competition. It is unclear whether these moves will stabilize or further destabilize their respective markets. What remains certain is that both arenas are undergoing profound, fast‑paced change, and the long‑term implications are still unknown.
Further Reading
- What to know about Netflix's landmark acquisition of Warner Bros. - TechCrunch
- Netflix agrees to buy Warner Bros. in an $82.7-billion deal that will ... - Los Angeles Times
- Netflix to Acquire Warner Bros. Following the Separation of ... - Netflix
- Netflix Makes $83 Billion All-Cash Bid for Warner Bros Discovery ... - AlphaSpread
Related Reading
Browse all Anthropic news →
LLMs & Generative AI Ant Group unveils Ring-1T, first open-source trillion-parameter reasoning model
LLMs & Generative AI ChatGPT Health Event Shows AI Modernizing Dev Workflows, GitLab Unveils Plans
LLMs & Generative AI Gen AI app sessions up fivefold, downloads jump 778% as ChatGPT leads traffic
Related Topic Google launches AI chips with 4× boost, lands Anthropic multibillion deal
Business & Startups
Related Topic Anthropic finds strict anti-hacking prompts increase AI sabotage and lying
Research & Benchmarks
From Archives Five Emerging Time‑Series Foundation Models Challenging Classic Methods
From Archives OpenClaw users prompt bots to explore Moltbook and verify accounts
From Archives Anthropic's Claude also citing Elon Musk's Grokipedia, reports say
From Archives Claude code migrates nearly 200 HA devices and builds tap‑control dashboard
Common Questions Answered
What new capabilities does Claude Opus 4.5 introduce?
[deeplearning.ai](https://www.deeplearning.ai/the-batch/claude-opus-4-5-retakes-the-coding-crown-at-one-third-the-price-of-its-predecessor/) reports that Claude Opus 4.5 offers adjustable effort levels, extended thinking capabilities, and improved tool use including web search and computer use. The model can process up to 200,000 tokens of input and generate up to 64,000 tokens of output, while performing at one-third the previous price point.
How does Anthropic describe Claude Opus 4.5's reasoning approach?
Anthropic describes Claude Opus 4.5 as a hybrid reasoning model that can respond rapidly in its default mode or take more time to process reasoning tokens when extended thinking is enabled. The model was trained on a combination of public web data and non-public data from various sources, and was fine-tuned to be helpful using reinforcement learning from human and AI feedback.
What performance benchmarks has Claude Opus 4.5 achieved?
[deeplearning.ai](https://www.deeplearning.ai/the-batch/claude-opus-4-5-retakes-the-coding-crown-at-one-third-the-price-of-its-predecessor/) indicates that in independent tests by Artificial Analysis, Claude Opus 4.5 scored second place on the Intelligence Index with a score of 70, matching OpenAI GPT-5.1 and trailing Google Gemini 3 Pro. In non-reasoning mode, the model scored 60, which was the highest among non-reasoning models tested.