AI models pushing back: A cautionary tale of the artificial confidence of o1 pro | Hacker News Reader