How Pangram works: AI-written replacement tokens get flagged as AI-generated
ctjlewis · x · 2026-10-12
The author explains that Pangram (an AI-detection tool) works similarly to the workflow being discussed: if you let AI write the replacement tokens, that paragraph will often flag as AI-generated because they're AI token sequences. The fix: take the bullet-point feedback and write the changes yourself, as you would with human feedback.
Related event: Keep your voice: let AI suggest edits, write them yourself(2 posts)→
More from AGI Musings
- Waymo CEO Dmitri Dolgov: problem convergence tells you if shipping is one year or ten away — a16z · 2026-10-12
- AI 2027 Update: Reality Tracking at 70-90% of Forecast, Superhuman Coder Slips to 2028 — jessi_cata · 2026-10-12
- People confess salaries, medical records and passwords to AI — the case for local LLMs — SuB8u · 2026-10-12
- Robotics researcher argues AI self-awareness is essentially inevitable on current trajectory — chris_j_paxton · 2026-10-12
- AI dev rips 'AI czar' Sacks: try post-training GLM 5.3 with just 3 rules — menhguin · 2026-10-12
- Mathematicians' AI anxiety and the 'rooms in a house' metaphor for identity — soumitrashukla9 · 2026-10-12