Post

Surprising: frontier models (Claude, ChatGPT, Deepseek V4) produce the most predictable text of any local AI model or human text I've ever tested. Not surprising: Finnegans Wake is off the charts, by far least predictable—Shannon & Lydia Liu proved right—and Hemingway the most predictable human.

Ryan Heuser@ryanheuser.com

Shannon measured the information rate of English at ~1 bit per character. According to a byte-level LLM measuring next-character predictability in LLM & human text (diaries, abstracts, dreams, fiction), aligned models produce sub-English information rates & LLM text is more predictable than humans'.

Jiggety Jatters. The worst name.

@jessefuchs.bsky.social

29 May

an interesting wrinkle is that I strongly suspect Hemingway’s predictability increased with each draft: i.e., it’s a deliberate design decision

Ryan Heuser

@ryanheuser.com

Definitely, no aesthetic judgment here per se. "Predictability" is part of its power.

29/5/26, 2:42 pm0 reposts3 likes

Jiggety Jatters. The worst name.

@jessefuchs.bsky.social

29 May

Yeah, I think what makes him distinct is his literary genre: more common among postmodern/experimental novelists) to try to create a novel artifact out of carefully assembled clichés. Whereas I imagine reading Hemingway’s drafts is more like watching Nick Offerman plane down the wood for a chair.