no, it did better because it was trained on a specific data point and the doctors were asked the same questions without meeting a real patient.