/Tech1h ago

Google DeepMind's Arthur Conmy flags non-standard terms in academic texts that reveal Claude-generated writing

Story Overview

Google DeepMind researcher Arthur Conmy highlighted unusual word patterns in academic writing that appear to mark text drafted by Claude, prompting wider conversation about how AI models are shaping the tone and structure of technical papers.

211021.3K

#31

Original post

Arthur Conmy@ArthurConmy#1278inTech

It seems to me that AI writing has changed, at least on technical topics. One year ago every AI was far too verbose always. But recent Opus outputs are sometimes not verbose enough. "Stacatto" for sure. Reminds me of this "split personality" point from @nostalgebraist

Kevin Patrick Murphy@sirbayes

While doing my NeurIPS reviews, my "claude generated paper" trigger kept going off - there is a distinctive style that they have - very stacatto, terse, defensive. What happens when LLMs start training on this text? Is the fixed point even readable to humans?!

7:08 AM · Jun 27, 2026 · 1.3K Views

Stylistic Clues

Terse phrasing signals model fingerprints

Reviewers have noticed a staccato style emerging in NeurIPS reviews and similar documents, with frequent pivot words such as anyways, anyway, and maybe standing out as potential markers of Claude-generated content.

Open Question

Self-training loops spark stagnation fears

Zachary Lipton questions whether models repeatedly trained on their own distinctive outputs could lock research into a narrower range of expression, though the scale of any such effect remains unmeasured.

Sentiment

Sentiment building, check back later.

Cluster Engagement

Digg Deeper

No Digg Deeper questions have been answered for this story yet.

Posts from X

Most Activity

VIEWS247LIKES3REPLIES1

Arthur Conmy@ArthurConmy

"forward model" and "lexemes" and "continuation top-20" and "clause-terminal" read as really non-standard phrases to me

Arthur Conmy@ArthurConmy

1h24730

RETWEETS1

Zachary Lipton@zacharylipton

@sirbayes Is the fixed point research progress or research stagnation?

Kevin Patrick Murphy@sirbayes

38m9900

Moon@MoonL88537

@ArthurConmy hard agree, and after reading tons of this stuff i am starting to think they are very semantically dense in ways we possibly don't totally get.

1h3