I found this article frustratingly vague on how prosecraft.io actually worked. As far as I can tell, the author scraped the web for books, including in-copyright books. Then he analyzed it with techniques based on "classical" natural language processing techniques, rather than transformers or deep learning. He appears to have retained the books he scraped for future analysis. The site itself seems to use only snippets.
However, the apology [0] says that the creator did not "intend" to participate in AI that can "create zero-effort impersonations of artists." I'm not sure if the wording is unintentionally vague, or if there is some way his project could be used in that way.
For what it's worth, the Computational Story Lab's hendometer [1] seems to have largely out-of-copyright books from Project Gutenberg, plus the Harry Potter series.
[0]: https://blog.shaxpir.com/taking-down-prosecraft-io-37e189797...
[1]: https://hedonometer.org/books/v3/863/
Edit: Apparently he was working on an LLM project. https://twitter.com/stealcase/status/1688721685585809408. It's unclear whether he was planning to use the books he scraped (although as @stealcase points out, GPT-Neox itself was trained on books that were pirated).
However, the apology [0] says that the creator did not "intend" to participate in AI that can "create zero-effort impersonations of artists." I'm not sure if the wording is unintentionally vague, or if there is some way his project could be used in that way.
For what it's worth, the Computational Story Lab's hendometer [1] seems to have largely out-of-copyright books from Project Gutenberg, plus the Harry Potter series.
[0]: https://blog.shaxpir.com/taking-down-prosecraft-io-37e189797...
[1]: https://hedonometer.org/books/v3/863/
Edit: Apparently he was working on an LLM project. https://twitter.com/stealcase/status/1688721685585809408. It's unclear whether he was planning to use the books he scraped (although as @stealcase points out, GPT-Neox itself was trained on books that were pirated).