Couldn't agree more, it’s wild how imposing some 'bounded rationality' on LLMs finally got them to cooperat. Could this approach also foster more robust ethical decision-making in autonomous agents?