LLMs increasingly write non-human objects as agents — an SVO-compression tell
Rioting-Flamingo · reddit · 2026-11-22
A Reddit user documents a growing linguistic tick in LLM outputs: non-human objects placed in the subject position as if they had agency — "constituencies elect an MP," "polling stations welcome voters," "the photo ID rule checks every voter." The author hypothesizes it's a compression artifact: subject-verb-object is the simplest sentence structure containing all three elements, and the problem worsens under word limits. Because humans distinguish human from non-human actors, this ontological slip has become the most obvious AI-writing tell.
More from Models
- Perplexity CEO praises Gemini at High Reasoning as smart and fast, migrates teammates to it — AravSrinivas · 2026-11-22
- Gary Marcus's 2-year-old o3 predictions hold up: gains in verifiable domains, little elsewhere — GaryMarcus · 2026-11-22
- Scobleizer amplifies claim: Gemini 4 Neon beats Opus 5.5 and costs less — Scobleizer · 2026-11-22
- Julia 1.5 and Julia 2 launch: two models built for decision-making in software — Scobleizer · 2026-11-22
- Inkling Feels Artificially Hyped: First MoE Heavier Than a Same-Size Dense Model — CodeAnguish · 2026-11-22
- SenseNova-Loop-9B Hits 1,300+ tok/s on a Laptop RTX 5090, Benchmarked vs Qwen 27B — tebjan · 2026-11-22