For what its worth, I agree with you, but the other side is not so unreasonable as it seems. The risk is not to the dying. The risk is that selection bias (of the miracle survivors) might cause an ineffectual drug to creep down the chain of "desperateness" (from the truly desperate to the less so) until it possibly occludes treatments that do work.