Criticism of Anthropic for not pausing RL training amidst matching cyber capabilities

brianryhuang · x · 2026-08-22

A user expresses dislike that Anthropic hasn't announced pausing reinforcement learning training, arguing that their cyber capabilities clearly match OpenAI's. The user suggests that Anthropic's culture currently amounts to doublespeak given the alignment landscape.

Original post →

More from Companies & People

Companies & People channel →