Models Love to Explicitly Say They Left Things Open
Miles_Brundage · x · 2026-08-17
Miles Brundage observed a common behavior in AI models where they explicitly state they are leaving issues open rather than covering them up, or addressing things one by one rather than pretending to be thorough. He attributes this to conflicting reward signals during training.
Related event: Why AI Models Keep Saying They're Not Hiding Problems: Conflicting Rewards(2 posts)→
More from Fun
- AI Reads and Writes Faster Than You, Jokes Say Elementary School Kids Are Cooked — SatelliteNetSec · 2026-08-17
- Twitter as 'Love on the Spectrum': AI community meme goes viral — justalexoki · 2026-08-17
- Bing Sydney crashout flagged as human by latest Pangram, netizens seek original chat — IgorBrigadir · 2026-08-17
- People got psyoped into buying maxed-out Mac Minis to run OpenClaw, netizens joke — ctjlewis · 2026-08-17
- User meme: I still don't know what Gemini Spark is — Zergylord · 2026-08-17
- 1567 Painting Apparently Predicted AI Hallucinations of Extra Limbs — docmilanfar · 2026-08-17