⚡ 67 h early
Anthropic’s ‘Watermark’ Text Adulteration in Claude Is a Perversion of Writing
6 experts across 4 network communities independently surfaced this.
6 experts
4 communities
1 sources clustered
“I’m convinced, upon rereading, that this simply misunderstands how either or both of how LLMs or watermarking works daringfireball.net/2026/08/anth...”
“John Gruber dislikes Anthropic’s watermarking because Claude may choose words that make AI text easier to detect instead of simply writing the best answer. I personally don’t care. Claude’s writing was already too clunky to send without editing. If editing …”
6 experts discussed this · 22 posts
Alex Hern: I’m convinced, upon rereading, that this simply misunderstands how either or both of how LLMs or watermarking works daringfireball.net/2026/08/anth...
Alex Hern: If you use Claude right now, then it doesn’t pick the “best” word for a sentence. It creates a distribution of possible words, assesses the probability of any given word fitting the slot, and rando…
Alex Hern: The watermarking technology randomly shifts that final selection, imperceptibly shifting the virtual dice roll. A word which may have been picked to fill a particular slot 40% of the time may, post…
Open the full discussion →