> Nobody is arguing for unsafe models

Then what are people arguing for? I see only two totally distinct options: unsafe models or someone being the arbiter of safety