Anthropic's 'Watermark' Text Adulteration in Claude Is a Perversion of Writing
Anthropic plans to embed a semantic watermark in all Claude-generated text, altering word choices to comply with an EU regulation. The author argues this adulterates writing quality, sacrifices precision for provenance, and is ineffective against determined cheaters. The piece critiques the EU's impractical regulation and Anthropic's misleading claims, offering a technical explanation of the watermarking scheme and its flaws.
The idea that anything other than my needs should factor into the generation of text for me is patently offensive.