Are AI models getting safer over time, or is it just bullshit?superagent.sh3 points·homanp··0 commentsOpen articleSaveView on HN