Anthropic has agreed to a landmark $1.5 billion settlement with authors, resolving claims that it pirated 500,000 books to train its AI models. If approved by the court, each author could receive at least $3,000 per work, in what attorney Justin Nelson called the “first of its kind” AI-era settlement, far surpassing any prior US copyright recovery. Authors’ Guild CEO Mary Rasenberger said the outcome shows “there are serious consequences when companies pirate authors’ works,” while Association of American Publishers president Maria Pallante emphasized it sends a clear message against sourcing from “shadow libraries.” Importantly, the deal compels Anthropic to destroy the pirated texts and does not shield it from future claims over infringing outputs. While Anthropic framed its approach as fair use, the settlement sets a precedent likely to ripple across the AI industry.
-EDITOR·OP_DAILY SHARE TO X
Molly Buckley, writing for the Electronic Frontier Foundation, warns that Mississippi’s age verification mandate, H.B. 1126, is already reshaping the internet in ways that entrench Big Tech while threatening smaller platforms. After the Supreme Court allowed the law to take effect, despite Justice Kavanaugh’s view that it is “likely unconstitutional,” Bluesky and Dreamwidth chose to block all Mississippi users rather than risk massive fines tied to per-user violations. Dreamwidth noted that “even a single $10,000 fine would be rough for us,” calling the structure an existential threat. Buckley argues that the burden of compliance forces smaller communities offline, while companies like Google and Meta can absorb costs and expand surveillance systems. Drawing parallels to the UK’s Online Safety Act, she concludes that these mandates consolidate power in corporate platforms, erode user choice, and jeopardize internet diversity and privacy.
-EDITOR·OP_DAILY SHARE TO X
Benj Edwards reports that OpenAI has introduced conversation branching in ChatGPT, allowing users to fork discussions into multiple parallel threads while preserving the original. The feature, available to all logged-in web users, enables experimentation—such as a marketing team testing different tones, without overwriting prior work. Edwards notes that branching addresses a long-standing limitation of linear dialogue, echoing a 2024 Tsinghua University study that found traditional interfaces increase cognitive load by forcing users to repeatedly revise and copy content. Developers likened the update to Git version control. While Anthropic’s Claude offered branching earlier, OpenAI frames the move as a user-driven update. Edwards emphasizes that the tool highlights a core truth: “You are providing direction to a connection machine—not consulting an oracle with its own agenda.” By enabling non-linear exploration, branching underscores that AI outputs are mutable, not authoritative.
-EDITOR·OP_DAILY SHARE TO X
Core Lightning has released version 25.09, “Hot Wallet Guardian,” with a focus on strengthening payment reliability, node management, and security. A major change brings the Bookkeeper database, previously a plugin, into core code, offering faster, more reliable satoshi-level tracking of balances, fees, and flows. As developer Niftynei noted, “every lightning node needs this…we’re giving it to them.” The update also integrates BIP353 human-readable payment names into xpay, improves payment routing by pruning inefficient paths and refactoring the minimum-cost flow solver, and adds finer control with maxparts for multi-part payments. Developer experience is streamlined by moving to the uv package manager, while protocol compliance is reinforced through mandatory channel type negotiation and enforced BOLT11 payment secrets to prevent probing attacks. Compatibility with Eclair’s splicing enhances channel uptime, further professionalizing Lightning’s reliability as a payments layer.
-EDITOR·OP_DAILY SHARE TO X
