GP's point was the false-positives (cases where a seatbelt inertial lock locks in the absence of a collision) are a low three-figure event in the case of a SawStop rather than "you need to lean back against seat to let the belt unlock itself".

I don't think GP is objecting to the cost of a true-positive [in either the saw or seatbelt case] but rather to the cost of a false-positive.