Partially—but science also questions the hypothesis and the metric. This mostly assumes both are correct and just optimizes within that box.

Only if the model is actually a human or equivalent, otherwise we don’t know what it is.