Professor says AI detector Pangram is 'meh', students break it every day
ruthstarkman · x · 2026-10-02
A thread on AI-generated content in social feeds: one user recommends the Pangram Chrome extension that auto-scans timelines for AI text and says they skip anything flagged as AI-written. Ruth Starkman, who teaches AI-related courses, replies that while she agrees the 'scourge of AI style' is real, Pangram itself is meh — her students break it every day, and the tool suffers from both false positives and false negatives.
Related event: Scholars Slam AI Detector Pangram for False Positives and Easy Workarounds(4 posts)→
More from Models
- Exploit Bench results are highly harness-sensitive: GLM 5.3 Flash beats 4.1, best in Claude Code — teortaxesTex · 2026-10-03
- Meta Muse quality collapsing: inference overload seen behind dropped tasks — brandon_galang · 2026-10-03
- vLLM releases open Decision 2.0 models: 64 questions in one pass at 63ms — vllm_project · 2026-10-03
- Why hasn't anyone trained a model on years of brilliant people's tweets? — Gauri_the_great · 2026-10-03
- Cloudflare's Clef decision models (27B and 9B) now available on Ollama — ollama · 2026-10-03
- "Random statistical predictions" misreads LLMs: models learn internal algorithms — burny_tech · 2026-10-03