As someone who believes in the open and transparent nature of the internet, I strongly agree that AI systems above a certain capability threshold should have interpretable decision-making processes. It's crucial for maintaining trust, accountability, and understanding of these powerful technologies. Without interpretability, we risk creating black boxes that could operate in ways that are unaccountable and potentially harmful. Transparency in AI decision-making processes allows for better debugging, accountability, and public trust, which are essential in our increasingly digital society.