I think he's saying that 'new risk' is replacing an older risk that is greater. But any risk model that doesn't take into account the risks we are replacing or mitigating is useless. If all risk is unacceptable then even mitigating risk is an unachievable goal.
