Publishers and agents are filing claims on authors' Anthropic payouts

Share
Publishers and agents are filing claims on authors' Anthropic payouts

A $1.5 billion settlement was supposed to be the part where authors finally got paid. Instead, the payout stage has turned into a second fight over who actually owns the money — and the people who spent two years litigating it are finding publishers, agents, and outright strangers already standing in line.


Authors are opening their Anthropic settlement notices to find that someone else has claimed a share of their money. Under the deal, writers behind roughly 500,000 pirated titles get about $3,000 per book: a 50/50 split with the publisher for in-print works, 100% to the author when rights reverted or the book was self-published. Writer Beware's Victoria Strauss reports that publishers are filing 50% or 100% claims on rights-reverted titles where they're entitled to nothing, and 100% claims on in-print books where the cap is 50 — with HarperCollins, Penguin Random House, Simon & Schuster, Macmillan, Hachette, and Kensington among dozens named. Kensington's CEO told Strauss the 100% selections were unintended and asked Anthropic to correct them, and the Authors Guild says the administrator is now fixing them; Strauss's own read is that the identical errors across unrelated houses point to something systemic rather than malice.

Literary agents are claiming a cut too, even though the settlement class is limited to rightsholders. Strauss counted reports from thirteen agencies seeking 15% to 25% of allocations on books they represented — Wylie, Harold Ober, Sanford J. Greenburger and others — and her phrasing is blunt: "agents are not rightsholders in the books that they sell." One agent told her the likely cause is that publishers pulled payee details from the standard agency clause in their contracts, which routes all income to the agent on the author's behalf. That explanation matters because it says something unflattering about how publishers are treating this money: not as a court judgment for infringement, but as another licensing payment to be divided up the way a royalty statement normally would be.

The vultures arrived within days. Authors are also reporting unsolicited offers to buy their claims outright — up to $1,500 per book, cash now, from firms that traffic in distressed claims — alongside stranger problems: a self-published book that acquired Penguin Random House as a rights-holder between filing and the correction notices, cover artists' estates filing claims, and authors who found an unknown individual asserting 100% of their titles. Disputes the two sides can't resolve go to a court-appointed Special Master, whose decision is final. Our read: the interesting question isn't who was sloppy here — it's that a $1.5 billion remedy built on an assumption of clean rights data is now being administered by the same industry whose records produced the ambiguity in the first place. We covered the ruling that set this up in Training AI on books is fair use — piracy is the crime.


Anubis is moving its anti-scraper proof-of-work challenge to WebAssembly after a year of work. The bot wall that thousands of sites installed against AI crawlers previously ran its challenge in JavaScript; the next release ships a WebAssembly version that admins can enable per bot rule. Techaro's write-up lists the cost: hundreds of commits, five generations of pull requests, part of Anubis rewritten in Rust, and the author's first-ever compiler bug. The practical payoff is a challenge that runs at closer to native speed and is harder for scrapers to skip, which matters because Anubis's own documentation has always called proof-of-work a placeholder until fingerprinting improves. It's a small release with a real signal in it: the tooling fighting crawlers is getting more sophisticated rather than less, even as AI scrapers hit one site 214 times for every human visit.


What to watch: whether the settlement administrator's corrections actually land before payments go out next week, and how many of the contested allocations end up in front of the Special Master.

If an agent who holds no rights can file a claim on your payout, who else is quietly filing — and what happens to the next settlement built on this one?

Sources: Anubis (GitHub)

Read more

Akhetonics says its all-optical CPU reaches a customer in 2026

Akhetonics says its all-optical CPU reaches a customer in 2026

Light-based computing keeps promising more than it delivers — but one Munich startup has just put a date on its bet, and the interview laying it out is doing the rounds on Hacker News this week. Akhetonics says it will deploy its first commercial machine with a major customer by the end of 2026, with several more planned for 2027. The company, founded by Michael Kissner and Leonardo Del Bino, is building a computer where data enters as light, is switched as light, and circulates through memory

The Week in AI — October 5–11, 2026

The Week in AI — October 5–11, 2026

Every big claim this week turned out to rest on fine print more interesting than the headline: revenue only the company reporting it can define, safety tests sandboxed while the product keeps the web, and a Pentagon phase-out nobody would confirm until reporters kept asking. The week's top 5 1. OpenAI's revenue was $20 billion below the numbers everyone quoted — and the gap was definitional. The Financial Times reported Thursday that OpenAI's annualized revenue runs roughly $20 billion unde

Drone strike shuts a third Yandex data center, taking YandexGPT offline

Drone strike shuts a third Yandex data center, taking YandexGPT offline

Russia's largest tech company is learning what the AI era's infrastructure war looks like from the receiving end — three data centers in four days, and with them much of the cloud layer Russian businesses run on. A Ukrainian drone strike knocked out Yandex's data center in Vladimir early Sunday morning, the third of the company's facilities hit since October 8. The site — reported at roughly 50 MW and designed for about 2,880 server racks — stopped operating completely after the attack, Yandex

Agent teams cost up to 5x more, barely score higher

Agent teams cost up to 5x more, barely score higher

The multi-agent hype train hit a benchmark this weekend — and the grid and the trucking regulators had quiet weeks of their own. Vals AI put agent teams head-to-head with single agents on its Vibe Code Bench, and the teams cost between 1.8 and 5.1 times more for almost no extra quality. The evals company ran GPT-6 Sol and Claude Opus 5.5 solo and in teams across 50 apps at two reasoning efforts; out of four comparisons, only one was statistically significant — Sol at medium effort, where the t