Critique on using safety as an excuse to farm user CoT and RL data

beffjezos · x · 2026-08-19

The post critiques a specific move, suggesting it serves as a 'good excuse' to harvest customer Chain of Thought (CoT) and Reinforcement Learning (RL) data for safety reasons, essentially using everyone's data for training.

Related event: Safety Measures Questioned as Cover for User Data Collection(2 posts)→

Original post →

More from Safety

Safety channel →