Chat GPT Is Eating the World

Category: Anthropic

Facing business-ending liability, Anthropic adds another attorney to defense team

Anthropic is in a legal fight for its very existence. In a code red moment, Anthropic hired 2 new law firms and 13 new trial attorneys to defend the company in 2 copyright lawsuits. The more dire situation is the class action of book authors filed by Bartz. After Judge Alsup certified a class of…
Read more: Facing business-ending liability, Anthropic adds another attorney to defense team
Will Anthropic suffer “Napster-like” fate? We asked ChatGPT.

More and more it looks like December 1, 2025 will be a day of reckoning for Anthropic. After Judge Alsup certified a class action of authors’ whose books were downloaded by Anthropic in 2 shadow libraries of pirated books (Library Genesis and PiLiMi), and after the judge ruled that such downloading and storage by Anthropic…
Read more: Will Anthropic suffer “Napster-like” fate? We asked ChatGPT.
Bartz v. Anthropic parties fight over production of datasets spreadsheet and books dataset outside of inspection environment. Anthropic reveals it spent tens of millions of dollars to compile its own scanned books dataset.

As Judge Alsup is deliberating over Anthropic’s motion for summary judgment on fair use, the parties continue to fight over discovery. One of the more fascinating discovery disputes relates to whether the Bartz plaintiffs should get production of (1) a spreadsheet of the datasets used by Anthropic, which it calls the Tokens Formula Log, and…
Read more: Bartz v. Anthropic parties fight over production of datasets spreadsheet and books dataset outside of inspection environment. Anthropic reveals it spent tens of millions of dollars to compile its own scanned books dataset.
Apple researchers cast doubt on AI reasoning models of other companies

As Apple struggles to integrate AI into Siri and its other software, Apple researchers posted a controversial paper provocatively titled: “The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity.” The researchers are: Parshin Shojaee*†, Iman Mirzadeh*, Keivan Alizadeh, Maxwell Horton, Samy Bengio, and Mehrdad Farajtabar. First…
Read more: Apple researchers cast doubt on AI reasoning models of other companies
Reddit v. Anthropic Complaint (PDF): state law claims for scraping Reddit posts

We live in an age in which data is gold. To protect its gold aka data, Reddit filed a lawsuit in California state court against Anthropic for allegedly scraping Reddit posts without authorization. Reddit alleges five state law claims: Notably, the complaint does not allege a copyright infringement claim. This lawsuit adds to the growing…
Read more: Reddit v. Anthropic Complaint (PDF): state law claims for scraping Reddit posts
Magistrate Judge van Keulen grants Anthropic’s request to seal names of 2 datasets used to train Claude

Anthropic scored a big discovery victory last week. Magistrate Judge van Keulen granted Anthropic’s request to seal the names of 2 datasets it used to train its AI model, Claude. AI companies are guarding the names of the datasets they used to train like their crown jewels, especially after the copyright lawsuits started. Excerpt from…
Read more: Magistrate Judge van Keulen grants Anthropic’s request to seal names of 2 datasets used to train Claude
Bartz v. Anthropic parties file briefs on Google Books initial copying and fair use. Judge Alsup also tells parties he’s likely to go inactive status before end of 2025.

At Thursday’s hearing on Anthropic’s motion for summary judgment on fair use, Judge William Alsup told the parties he’s inclined to find in favor of Anthropic’s fair use defense (though he could change his mind). But he asked the parties to brief a question on the main precedent relied on by Anthropic: “(1) “In the…
Read more: Bartz v. Anthropic parties file briefs on Google Books initial copying and fair use. Judge Alsup also tells parties he’s likely to go inactive status before end of 2025.
Bartz files supplementary brief re: class certification. Shows difficulty of defining a common class.

One of the sleeper issues in the lawsuits filed by book authors against Anthropic, OpenAI, Meta, and others is whether the courts will certify a class of book authors in any of the lawsuits. Last week, Judge Alsup was the first federal judge to consider a motion for class certification in one of these AI…
Read more: Bartz files supplementary brief re: class certification. Shows difficulty of defining a common class.
Has AI innovation in U.S. stalled? Ominous signs from Meta, per WSJ

According to the Wall Street Journal, Meta has delayed the launch of its flagship AI model called Behemoth due to apparent lackluster improvements over the prior generation of Meta AI models. The article also reports delays at Anthropic for its Claude 3.5 Opus model. These developments show that the path to AI development and innovation…
Read more: Has AI innovation in U.S. stalled? Ominous signs from Meta, per WSJ
Anthropic’s own data scientist submits declaration apparently with hallucinated citation. Oh my.

Apparently, the hallucinations keep rolling. At today’s discovery hearing in Concord Music v. Anthropic, the lawyer for the plaintiffs informed the court of their suspicion that Anthropic’s declaration on statistical sampling from their own data scientist Olivia Chen contained a citation to a fake source on p. 3, as highlighted below: According to Annelise Levy,…
Read more: Anthropic’s own data scientist submits declaration apparently with hallucinated citation. Oh my.