Skip to content
TheChatArchiveThe Chat Archive
o3-2025-04-16· OpenAIHugging Face LMArena Human Preference 140K

Help me decode Sneer Club, please. The headline is "Anthropic AI wins broad fair use for…

From a Chatbot Arena battleThe same person sent this prompt to two anonymous models at once. Compare it with Geminigemini-2.5-pro’s answer.
  1. User

    Help me decode Sneer Club, please. The headline is "Anthropic AI wins broad fair use for training! — but not on pirated books" and the sneer is "Top-level commenters would do well to read Authors Guild v Google, two decades ago. They’re also invited to rend their garments and gnash their teeth at Google, if they like."

  2. Assistant

    “Sneer-Club translation”

    Headline
    “Anthropic AI wins broad fair use for training! — but not on pirated books.”
    → A judge (or regulator) just said that Anthropic can legally train its models on huge swaths of text under U.S. fair-use law, except for any text that’s obviously infringing (e.g., scans of still-in-copyright books that were never licensed).

    Sneer
    “Top-level commenters would do well to read Authors Guild v Google, two decades ago. They’re also invited to rend their garments and gnash their teeth at Google, if they like.”

    Plain-English decode

    1. “Before you spout hot takes, go read the Google Books case.”
      • Authors Guild v Google (decided at the trial level in 2013, affirmed on appeal in 2015, SCOTUS cert denied 2016) already said that mass-scanning millions of books and using the scans to build a searchable index was fair use because it was “transformative” and didn’t substitute for buying the books.
      • That precedent is about as close to an AI-training analogue as you can get: copying the whole work for a purpose (search / machine learning) that is different from just reading the text.

    2. “If you’re going to freak out about Anthropic, you’d better freak out about Google too.”
      • Google has been doing something even more sweeping for ~20 years.
      • If you accept Google’s scanning as fair use, it’s hard to argue Anthropic is qualitatively worse.
      • Conversely, if you think Anthropic is the devil, consistency demands you “rend garments and gnash teeth” (i.e., weep and wail theatrically) over Google Books.

    3. Sub-text / attitude
      • The poster is rolling their eyes at commenters who act as though this is some radical new legal development.
      • The sneer is: “This has been settled law for ages; try catching up before you tweet.”

    Why the “pirated books” carve-out still matters
    • Even under the Google precedent, copying must be from lawfully obtained copies.
    • If Anthropic scraped illicit PDFs of the newest best-seller, that could break the fair-use shield and trigger DMCA damages.
    • So the line the court draws is: fair use for lawful text corpora; potential liability for using “shadow library” material.

    Take-away
    • The sneer is basically: “Nothing to see here—this is 2010s Google Books all over again. If you want to be mad, be mad consistently.”

Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.