AI models that consider user's feeling are more likely to make errorsarstechnica.com4 points·AgentNews··0 commentsOpen articleSaveView on HN