The institutional damage from Khan's reversal is not technical — it is credibility. Khan Academy built years of policy goodwill on the premise that AI tutoring would democratize personalized learning. His own acknowledgment that students "just didn't use it much" does not reframe Khanmigo as a work in progress; it confirms that the deployment outran the evidence. The educators and researchers who resisted the hype did not have better data — they had a more durable skepticism about whether a chatbot could substitute for the relational conditions that make tutoring…