Platinuman comments on Announcing an Empirical AI Safety Program