Newer LLMs Are Converging on a Dense, Gibson-like Writing Style, Redditor Observes
AnticitizenPrime · reddit · 2026-10-03
A Redditor spots a cross-vendor trend: latest models, open and closed alike, are converging on a terse, information-dense writing style with an expanded vocabulary.
- Testing GLM 5.3 Flash with a silly prompt about naming a Chinese restaurant, the output included lines like "Baudrillardian flourish at the failure point" — six words packed with meaning that many readers would find hard to parse.
- The style resembles William Gibson's prose: maximum meaning per word, choosing richer vocabulary over longer sentences — "Gibson-speak."
- Possible cause: token-efficiency-oriented training targets.
- Tradeoff: reduced readability for average users; rare in human writing but increasingly common across new models.
More from Models
- Qwen's Justin Lin: DeepSeek v4.1 Flash Paper With Only 191 Upvotes Is 'Ridiculous' — JustinLin610 · 2026-10-03
- Open-closed model gap widest at 12 points; observer predicts Chinese labs close it in a year — teortaxesTex · 2026-10-03
- Indie project Auro, fine-tuned for GPT-4o-style personality, preps V6 release — TheMoonMidas · 2026-10-03
- GPT-6 Astra clears StarCraft BW Tier A bots after nearly 8 hours of play — philipvollet · 2026-10-03
- Redditor finds GPT 6.1 Sol runs for hours without usage caps, matching Opus — topdaytrader1 · 2026-10-03
- Anthropic engineer investigating reports that Pro 500 plans didn't get expected usage reset — soumitrashukla9 · 2026-10-03