Anthropic’s ‘secret plan’ to ‘destructively scan all the books in the world' revealed by unredacted files
www.thebookseller.com
A number of newly unredacted (and partially redacted) documents from the Anthropic case have shed light on how executives at the tech giant ramped up plans in early 2024 to scan “all the books in the world” to teach their AI tool “how to write well”.

In the filings, Anthropic states, as reported by the Washington Post: “Project Panama is our effort to destructively scan all the books in the world. We don’t want it to be known that we are working on this.”

  • FauxLivingEnglish
    41·
    5 hours ago

    https://arstechnica.com/features/2025/06/study-metas-llama-3-1-can-recall-42-percent-of-the-first-harry-potter-book/

    The claim was “Yet most AI models can recite entire Harry Potter books if prompted the right way, so that’s all bullshit.”

    In this test they did not get a model to produce an entire book with the right prompt.

    Their measurement was considered successful if it could reproduce 50 tokens (so, less than 50 words) at a time.

    The study authors took 36 books and divided each of them into overlapping 100-token passages. Using the first 50 tokens as a prompt, they calculated the probability that the next 50 tokens would be identical to the original passage. They counted a passage as “memorized” if the model had a greater than 50 percent chance of reproducing it word for word.

    Even then, they didn’t ACTUALLY generate these, they even admit that it would not be feasible to generate some of these 50 token (which is, at most 50 words, by the way) sequences:

    the authors estimated that it would take more than 10 quadrillion samples to exactly reproduce some 50-token sequences from some books. Obviously, it wouldn’t be feasible to actually generate that many outputs.

    • NostraDavid@programming.devEnglish
      2·
      5 hours ago

      The claim was “Yet most AI models can recite entire Harry Potter books if prompted the right way, so that’s all bullshit.”

      In this test they did not get a model to produce an entire book with the right prompt.

      For context: These two sentences are 46 Tokens/210 Characters, as per https://platform.openai.com/tokenizer.

      50 tokens is just about two sentences. This comment is about 42 tokens itself.