Jacobian Conjecture Refutation Reveals a Structural Limit of AI Interpretability | Hacker News Reader