Anna's Archive hit with $19.5M default judgment and global domain takedown order

Original Article ↗ Hacker News Discussion ↗

A $19.5M default judgment and global domain takedown order against shadow library Anna’s Archive has reignited debates over U.S. courts’ extraterritorial reach and the leverage they exert via ICANN, DNS, and international treaties. Commenters weigh the ethics of large-scale book piracy against the public-interest role of open access to knowledge, especially as traditional libraries struggle with restrictive digital licensing. The case also highlights perceived double standards, with several noting that major AI companies allegedly trained on the same pirated collections yet face very different legal and practical consequences.

Jurisdiction & Global Reach

  • Many debate how a New York court can order domain takedowns worldwide.
  • Some call it “performance art,” but others note:
    • ICANN and root DNS are US-based, giving leverage even over country TLDs.
    • Mutual legal assistance treaties and trade agreements can pressure foreign entities.
    • Historic examples: pressure on Sweden over The Pirate Bay; Assange extradition.
  • Concerns that stronger enforcement (e.g., via RIPE, US transit providers) could fragment the global internet if IP ranges/ASNs become political tools.

Effectiveness of Takedowns

  • Consensus that new domains and mirrors will appear; compared to The Pirate Bay “hydra.”
  • DNS-focused injunctions break links (e.g., from Wikipedia) but don’t erase content.
  • Suggestions: Tor/onion services for stronger censorship resistance; prediction that enforcement may push AA in that direction.

AI Companies vs Shadow Libraries

  • Thread highlights that publishers cited AA as an AI training hub (Meta, NVIDIA).
  • AA reportedly offered high-speed bulk access for large donations; at least one major US AI company allegedly paid for it.
  • Contrast drawn with big AI firms:
    • They face lawsuits and large settlements, but keep domains and operations.
    • Argument that rich companies can “pay to proceed,” while AA operators risk prison and therefore avoid court.
  • Debate over legality:
    • Some say training is treated differently from distribution; AA directly distributes copies, AI companies mostly don’t.
    • Others argue models are effectively “IP laundromats” and should be retrained without infringing data.

Piracy, Authors, and Access

  • Sharp split:
    • One side: AA harms authors, publishers, booksellers, and even libraries; undermines future work.
    • Other side: AA provides vital access and preservation; law and current copyright terms (~life+decades) are seen as unjust.
  • Piracy framed by some as a “service problem”: if DRM-free, reasonably priced ebooks were easily available, demand for AA would drop.

Libraries, Digital Goods & Control

  • Discussion of how digital licensing breaks the traditional “first sale” model:
    • Libraries pay per-loan or time-limited licenses instead of owning ebooks.
    • Digital licenses are expensive; publishers keep control and can limit or stop lending.
  • Worries about centralized censorship: far easier to choke off access by not renewing digital licenses than by pulling physical books.
  • Some advocate personal archives and decentralized, censorship-resistant “shadow library” designs; others stress the broader social role of public libraries.