conv.

← The whole story
AI
Claude's AI text watermark sparks backlash, cancellations and a 'perversion of writing' row

Gruber calls the watermark 'a perversion of writing'

7 Aug 16 5:53 PM · 62d ago · 16 articles · 33 posts · 20 comments · 5 sources · development 7 of 8

John Gruber published a widely discussed Daring Fireball essay arguing Anthropic's claim that watermarking doesn't affect quality was false, since biasing word choice for detectability inherently corrupts text generated for the user; it drew a 714-comment Hacker News thread and follow-up Business Insider coverage.

Gruber calls the watermark 'a perversion of writing'
techcrunch.com

“The idea that anything other than *my* needs should factor into the generation of text *for me* is patently offensive.”

John Gruber

Anthropic AI lab, maker of ClaudeJohn Gruber Daring Fireball writerArturo Villarroya Claude Max subscriber, public policy workerVladislav Rajtmajer Freelance developer, Czech RepublicJames Padolsey Author of interactive watermarking explainer

The whole story articlespostscomments the bright band is this development · numbered dots are the others · click one to jump

No items in the last 30 days Sep 17, 7 PM — quietSep 18, 2 AM — quietSep 18, 9 AM — quietSep 18, 4 PM — quietSep 18, 11 PM — quietSep 19, 6 AM — quietSep 19, 1 PM — quietSep 19, 8 PM — quietSep 20, 3 AM — quietSep 20, 10 AM — quietSep 20, 5 PM — quietSep 21, 12 AM — quietSep 21, 7 AM — quietSep 21, 2 PM — quietSep 21, 9 PM — quietSep 22, 4 AM — quietSep 22, 11 AM — quietSep 22, 6 PM — quietSep 23, 1 AM — quietSep 23, 8 AM — quietSep 23, 3 PM — quietSep 23, 10 PM — quietSep 24, 5 AM — quietSep 24, 12 PM — quietSep 24, 7 PM — quietSep 25, 2 AM — quietSep 25, 9 AM — quietSep 25, 4 PM — quietSep 25, 11 PM — quietSep 26, 6 AM — quietSep 26, 1 PM — quietSep 26, 8 PM — quietSep 27, 3 AM — quietSep 27, 10 AM — quietSep 27, 5 PM — quietSep 28, 12 AM — quietSep 28, 7 AM — quietSep 28, 2 PM — quietSep 28, 9 PM — quietSep 29, 4 AM — quietSep 29, 11 AM — quietSep 29, 6 PM — quietSep 30, 1 AM — quietSep 30, 8 AM — quietSep 30, 3 PM — quietSep 30, 10 PM — quietOct 1, 5 AM — quietOct 1, 12 PM — quietOct 1, 7 PM — quietOct 2, 2 AM — quietOct 2, 9 AM — quietOct 2, 4 PM — quietOct 2, 11 PM — quietOct 3, 6 AM — quietOct 3, 1 PM — quietOct 3, 8 PM — quietOct 4, 3 AM — quietOct 4, 10 AM — quietOct 4, 5 PM — quietOct 5, 12 AM — quietOct 5, 7 AM — quietOct 5, 2 PM — quietOct 5, 9 PM — quietOct 6, 4 AM — quietOct 6, 11 AM — quietOct 6, 6 PM — quietOct 7, 1 AM — quietOct 7, 8 AM — quietOct 7, 3 PM — quietOct 7, 10 PM — quietOct 8, 5 AM — quietOct 8, 12 PM — quietOct 8, 7 PM — quietOct 9, 2 AM — quietOct 9, 9 AM — quietOct 9, 4 PM — quietOct 9, 11 PM — quietOct 10, 6 AM — quietOct 10, 1 PM — quietOct 10, 8 PM — quietOct 11, 3 AM — quietOct 11, 10 AM — quietOct 11, 5 PM — quietOct 12, 12 AM — quietOct 12, 7 AM — quietOct 12, 2 PM — quietOct 12, 9 PM — quietOct 13, 4 AM — quietOct 13, 11 AM — quietOct 13, 6 PM — quietOct 14, 1 AM — quietOct 14, 8 AM — quietOct 14, 3 PM — quietOct 14, 10 PM — quietOct 15, 5 AM — quietOct 15, 12 PM — quietOct 15, 7 PM — quietYesterday, 2 AM — quietYesterday, 9 AM — quietYesterday, 4 PM — quietYesterday, 11 PM — quietToday, 6 AM — quietToday, 1 PM — quiet ◂ 1–8 earlier
Sep 20Sep 23Sep 26Sep 29Oct 2Oct 5Oct 8Oct 11now · 7:31 PM ET

What was reported 3 claims about this development

  1. 3 outlets Claude to start watermarking AI-generated text – but will it make quality worse?

    first by The Guardian, 61d ago · also Guardian Media, Guardian Tech

  2. 1 outlet John Gruber calls Claude's AI watermarking 'patently offensive'

    first by Business Insider, 61d ago

  3. 1 outlet Anthropic says text watermarking scheme relies on inconsequential words

    first by The Register, 61d ago

What people said 24 voices · best of 41 · verbatim

All 8 developments of Claude's AI text watermark sparks backlash, cancellations… →

Hacker NewsNewswiresMastodonRedditXBlueskyGoogle NewsLobsters