The purchase, scanning, and destruction of individual books by Anthropic is causing a public opinion tsunami! In order to obtain "clean" training data without AI generated text, physical books published before 2022 have become a gold mine in the eyes of AI companies. Anthropic has launched an internal project codenamed "Panama Plan" with the goal of acquiring "all the books in the world". They recruited the former head of the Google Books scanning project and spent millions of dollars to purchase physical books in bulk, including a large number of rare and obscure books - then removed the spine, cut open each page, scanned them into digital files at high speed, and finally crushed and destroyed all the paper originals. The US federal judge ruled that this practice constitutes "reasonable use" - the reason being that as long as the physical book is legally purchased, scanned and destroyed, and not distributed to the public, it does not constitute infringement. What is even more disturbing is that this approach has formed a chain of supporting services: A company called ISBNdb publicly offers services for AI buyers to anonymously purchase old books, with single orders ranging from one thousand to one million copies, accompanied by confidentiality agreements, and even suggesting that customers refer to it as "digital storage". Booksellers have discovered that someone is stocking up on unpopular books in large quantities. A second-hand bookstore owner in Houston found that these books are either out of stock or priced at five to twenty times the purchase price on Amazon after picking up orders. If the high price list cannot be sold, Amazon's FBA inventory will be liquidated when it expires - those entire Internet may have only a few surviving copies, and even digital records will be completely evaporated before it is too late. Public opinion has completely exploded. Musk publicly stated that his SpaceX AI will only scan and not open books in the future, and investor Michael Burry only wrote two words on X: "Evil incarnate A large number of netizens and book fans angrily questioned: those lonely books that have survived wars, fires, and hundreds of years to pass down, were actually chopped up and fed to AI in the end? This controversy has long exceeded the scope of copyright. AI companies obtain replicable and trainable data, while humans may lose cultural artifacts that cannot be reconstructed. The law can recognize this type of operation as compliant, but it cannot answer a more fundamental question: Do we have the right to turn the last original document left by civilization into disposable consumables in order to train stronger models?