Obviously zero deaths is a great target, yes. Obviously zero deaths is impossible. No energy production method causes zero deaths; even wind and solar have a death toll (eg, from workers falling off things, construction accidents, manufacturing processes, etc), and yes, at the scale we're talking about that adds up.
Everything causes deaths when you're talking about thousands of TWhs. At most you can minimise the deaths.
> Deaths are not made acceptable because of energy production.
Nor are they acceptable because of food production, and yet food production causes deaths. Fishing boats are famously dangerous places; even farms have a lot of hazards (chemicals, large animals, heavy machinery). We could say the same about minerals, lumber (a friend of mine lost her father at a young age; he worked in forestry and a tree came down the wrong way), and more. And yet, we need all these things.
> Deaths per TWh was not a metric we measured. [...] nobody ever said deaths would be OK if only the numbers went up.
I don't doubt it. That is, after all, how we ended up in this mess.
Japan is currently building a large fleet of coal fired power plants (https://www.nytimes.com/2020/02/03/climate/japan-coal-fukush...). Those plants will kill a horrifying number of people; many hundreds of times more than died from the Fukushima disaster.
That seems fairly clearly like the sort of decision you'd made if you weren't thinking about the obvious results.