In a major win for AI companies, a federal judge in San Francisco ruled on Monday that Anthropic’s use of millions of copyrighted books —without writers’ permission— to train its AI models was legal ...