4 ms·

As a PSA, one additional reason to seed is because Anna accidentally doxxed herself via GitHub. So it’s worth preserving the archive on the basis that we should

As a PSA, one additional reason to seed is because Anna accidentally doxxed herself via GitHub. So it’s worth preserving the archive on the basis that we should expect {the centralized portion of} it to disappear within the next couple years. I was sad to see that happen, but it’s important to be objective and plan future actions accordingly. (And sure, there’s always the chance that some random person on GitHub just so happens to be named Anna and is an archival enthusiast, but a jury of one’s peers may find that it passes the reasonable doubt threshold.) My legal troubles with books3 weighed on me pretty heavily, and I wasn’t even the target. Yet. I can only imagine what it feels like to be waiting for an indictment. There ought to be some sort of protection for preserving books in bulk. No one is going to read two million books. But of course, one could also argue that having a readily available archive is harming the economic profitability of the works, on the basis that content licensing for AI is now a multimillion industry. It’s weird, because it feels like important work, rather than criminal — someone should put into words exactly what the distinction is.

iwontberude 2y ago
I think you’ve stumbled upon a paradox where the data is worth preserving because of its scarcity, but once it no longer scarce, the value is diminished along with the priority to preserve and make it discoverable. Similar in some way to the antique/collectibles market and the U shaped curve. The issue is that the data can become so devalued that we create a scarcity in the future. It’s one of those self-regulating systems.
jprete 2y ago
I think the Library of Congress already does preservation; it seems to be legally required to give the LoC two copies of every published work in the US (https://www.copyright.gov/circs/circ7d.pdf https://www.copyright.gov/circs/circ7d.pdf). Preservation of works is very obviously not why Anna's Archive is asking for torrent seeders. Seeding is for distribution and availability, not preservation. It would be more honest to say "preservation of ML training fair use access".
- FabHK 2y ago
  > Preservation of works is very obviously not why Anna's Archive is asking for torrent seeders. Can you elaborate on that? They write: "These torrents are not meant for downloading individual books. They are meant for long-term preservation. With these torrents you can set up a full mirror of Anna’s Archive, using our source code and metadata (which can be generated or downloaded as ElasticSearch and MariaDB databases)."
  - jprete 2y ago
    I'm no longer convinced I'm correct.
- dredmorbius 2y ago
  Preservation is, unfortunately, not access for not only the vast majority of US citizens, but of people worldwide locked out of the proprietary document system. Even such half-aborted systems such as Hathi Trust permit downloads only by the page, even from out-of-copyright works, which is absolutely infuriating. Full access to the Hathi archives (that is, in-copyright works, which cannot otherwise be viewed at all) is restricted to college libraries only, not public libraries generally. The Internet Archive, LibGen, Z-Library, Sci-Hub, Anna's Archive, and other similar efforts are really the only viable means of access to much of the world's published information, whether inside, outside, or straddling extant copyright law. Which of course was written and lobbied for by existing copyright holders. I'd much rather we burnt down that law than our true libraries. As to Anna's Archive and the question of preservation vs. access: AFAIU the full archive is already available and stored in multiple fully-independent copies. As such it will all but assuredly survive even extreme legal attacks, let alone other threats. Access to those works is what torrent seeding provides, and as a means of making the archive available and useful is key to its function and service, but (probably) not its survival.
esdf 2y ago
They did deny involvement so it'll be interesting to see what happens (you probably know this though) https://torrentfreak.com/key-defendant-in-annas-archive-lawsuit-denies-any-involvement-with-the-site-240416/ https://torrentfreak.com/key-defendant-in-annas-archive-laws... >It’s weird, because it feels like important work, rather than criminal — someone should put into words exactly what the distinction is. Important work can be criminal
- ffsm8 2y ago
  > Important work can be criminal Yeah, illegal/criminal merely means that it's against the law and will be prosecuted if it can be pinned to you. As an outlandishly severe example: helping Jews in Nazi Germany was criminal.
dredmorbius 2y ago
Link to the doxxing incident? When / how did this happen? Edit: I'm asking as I find no mention of this elsewhere. Notably not on Torrentfreak, which tends to keep on top of such things, on Anna's subreddit, blog, or her Telegram channel
- cbsks 2y ago
  Maybe this is it? https://torrentfreak.com/key-defendant-in-annas-archive-lawsuit-denies-any-involvement-with-the-site-240416/ https://torrentfreak.com/key-defendant-in-annas-archive-laws... > On information and belief, in addition to her extensive online presence, she has a GitHub (a software code hosting platform) account called, "anarchivist," and she developed a repository for a python module for interacting with OCLC's WorldCat® Affiliate web services.
  - sillysaurusx 2y ago
    Yeah. There was some HN discussion about it when the story broke too; maybe someone can dig it up on HN Search. I think the civil suit will probably be dismissed due to lack of evidence, but it’s likely the police have probable cause to start a surveillance warrant. At that point, either she’ll stop all activity (which it doesn’t seem like she’s doing) or it’s inevitable they’ll catch her in the act. The slim hope is that maybe it’s not her, or that police won’t pour a lot of resources into surveilling her. Maybe she’ll get lucky. But given the aggressive ways law enforcement has gone after people for e.g. scraping AT&T’s website, I wouldn’t bet on it.
    - dredmorbius 2y ago
      Thanks, and I'm hoping likewise. Here's the 2-month-old discussion on the Torrentfreak article: <https://news.ycombinator.com/item?id=40143549 https://news.ycombinator.com/item?id=40143549> And that seems to be the only HN submission with > 5 points that matches this description since 1-Apr-2024: <https://hn.algolia.com/?dateEnd=1718327852&dateRange=custom&dateStart=1711929600&page=0&prefix=true&query=anna%27s%20archive%20points%3E5&sort=byDate&type=story https://hn.algolia.com/?dateEnd=1718327852&dateRange=custom&...> A similar search for "OCLC" turns up nothing.

iwontberude 2y ago

I think you’ve stumbled upon a paradox where the data is worth preserving because of its scarcity, but once it no longer scarce, the value is diminished along with the priority to preserve and make it discoverable. Similar in some way to the antique/collectibles market and the U shaped curve. The issue is that the data can become so devalued that we create a scarcity in the future. It’s one of those self-regulating systems.

jprete 2y ago

I think the Library of Congress already does preservation; it seems to be legally required to give the LoC two copies of every published work in the US (https://www.copyright.gov/circs/circ7d.pdf https://www.copyright.gov/circs/circ7d.pdf). Preservation of works is very obviously not why Anna's Archive is asking for torrent seeders. Seeding is for distribution and availability, not preservation. It would be more honest to say "preservation of ML training fair use access".

FabHK 2y ago
> Preservation of works is very obviously not why Anna's Archive is asking for torrent seeders. Can you elaborate on that? They write: "These torrents are not meant for downloading individual books. They are meant for long-term preservation. With these torrents you can set up a full mirror of Anna’s Archive, using our source code and metadata (which can be generated or downloaded as ElasticSearch and MariaDB databases)."
- jprete 2y ago
  I'm no longer convinced I'm correct.
dredmorbius 2y ago
Preservation is, unfortunately, not access for not only the vast majority of US citizens, but of people worldwide locked out of the proprietary document system. Even such half-aborted systems such as Hathi Trust permit downloads only by the page, even from out-of-copyright works, which is absolutely infuriating. Full access to the Hathi archives (that is, in-copyright works, which cannot otherwise be viewed at all) is restricted to college libraries only, not public libraries generally. The Internet Archive, LibGen, Z-Library, Sci-Hub, Anna's Archive, and other similar efforts are really the only viable means of access to much of the world's published information, whether inside, outside, or straddling extant copyright law. Which of course was written and lobbied for by existing copyright holders. I'd much rather we burnt down that law than our true libraries. As to Anna's Archive and the question of preservation vs. access: AFAIU the full archive is already available and stored in multiple fully-independent copies. As such it will all but assuredly survive even extreme legal attacks, let alone other threats. Access to those works is what torrent seeding provides, and as a means of making the archive available and useful is key to its function and service, but (probably) not its survival.

esdf 2y ago

They did deny involvement so it'll be interesting to see what happens (you probably know this though) https://torrentfreak.com/key-defendant-in-annas-archive-lawsuit-denies-any-involvement-with-the-site-240416/ https://torrentfreak.com/key-defendant-in-annas-archive-laws... >It’s weird, because it feels like important work, rather than criminal — someone should put into words exactly what the distinction is. Important work can be criminal

ffsm8 2y ago
> Important work can be criminal Yeah, illegal/criminal merely means that it's against the law and will be prosecuted if it can be pinned to you. As an outlandishly severe example: helping Jews in Nazi Germany was criminal.

dredmorbius 2y ago

Link to the doxxing incident? When / how did this happen? Edit: I'm asking as I find no mention of this elsewhere. Notably not on Torrentfreak, which tends to keep on top of such things, on Anna's subreddit, blog, or her Telegram channel

cbsks 2y ago
Maybe this is it? https://torrentfreak.com/key-defendant-in-annas-archive-lawsuit-denies-any-involvement-with-the-site-240416/ https://torrentfreak.com/key-defendant-in-annas-archive-laws... > On information and belief, in addition to her extensive online presence, she has a GitHub (a software code hosting platform) account called, "anarchivist," and she developed a repository for a python module for interacting with OCLC's WorldCat® Affiliate web services.
- sillysaurusx 2y ago
  Yeah. There was some HN discussion about it when the story broke too; maybe someone can dig it up on HN Search. I think the civil suit will probably be dismissed due to lack of evidence, but it’s likely the police have probable cause to start a surveillance warrant. At that point, either she’ll stop all activity (which it doesn’t seem like she’s doing) or it’s inevitable they’ll catch her in the act. The slim hope is that maybe it’s not her, or that police won’t pour a lot of resources into surveilling her. Maybe she’ll get lucky. But given the aggressive ways law enforcement has gone after people for e.g. scraping AT&T’s website, I wouldn’t bet on it.
  - dredmorbius 2y ago
    Thanks, and I'm hoping likewise. Here's the 2-month-old discussion on the Torrentfreak article: <https://news.ycombinator.com/item?id=40143549 https://news.ycombinator.com/item?id=40143549> And that seems to be the only HN submission with > 5 points that matches this description since 1-Apr-2024: <https://hn.algolia.com/?dateEnd=1718327852&dateRange=custom&dateStart=1711929600&page=0&prefix=true&query=anna%27s%20archive%20points%3E5&sort=byDate&type=story https://hn.algolia.com/?dateEnd=1718327852&dateRange=custom&...> A similar search for "OCLC" turns up nothing.