> [Nerds] sit in a room and dream up every way the device could fail and/or harm someone

Well, yeah. And as it happened they postulated the failure mode that was actually (allegedly) seen in practice. And they were right. So the "no one could have known" side of Sig's defense seems out the window here. They could have known, and they did, and relevant experts told them.

I really don't get your point here. You seem to be saying that risk analysis, in the abstract, as a whole field of practice, has no value because of the lack of certainty? But... managing uncertainty is the whole point. Do you really live your life like this? "I mean, people say fentanyl is dangerous but you never really know, right?"