I agree in principle - but the hypothetical was about someone who was (presumably measurably) worth twice as much as other employees.
I agree that it's difficult to measure productivity / output - but it's by no means impossible - and sharp discrepancies in productivity between employees are often highly noticeable anyway.