The idea that some safety costs are entirely frivolous, obviously so, with no serious argument against that notion, just doesn't strike me as believable. Yes, government projects (and private industry projects) tend to run over budgets. The saying is "work expands to meet the budget provided", but that's intended to be cynical snark, not descriptive of reality. There are plenty of governments that care less for safety than the nations of the global north, yet they haven't found this holy sweet spot of safe efficiency, as far as I can tell.
The reason we still have to run coal and natural gas plants is to cover the times that wind and solar aren't producing. Nuclear is a substitute for coal and natural gas, not wind and solar. Because we're shutting down old nuclear plants and not building new ones, coal plants are still going up in many parts of the world. Even if nuclear was ten times as dangerous, it would still be a net win on safety compared to natural gas and especially coal.
But also, my understanding is that a whole bunch of these deaths were at Chernobyl. From the article I linked:
The deaths from Chernobyl, 35 years ago, were due to unforgivably bad reactor design that we‘ve advanced far beyond now. There were zero deaths from radiation at Three Mile Island or at Fukushima. (The only deaths from the Fukushima disaster were caused by the unnecessary evacuation of 160,000 people, including seniors in nursing homes.)
Did you check out the article from my original comment? It does explain things much better than I can. And about your last point of the sweet spot of safe efficiency -- Jack Devaney, the author of the book that the article is about, is currently partnering with the government of Indonesia to produce nuclear power that's cheaper than coal: https://thorconpower.com/ The website is pretty generous with info, and doesn't set off my bullshit/vaporware detector. Maybe it's already happening.
That said, it's likely we can't fix our regulation AND build new plants in time to fix climate change, so I agree with you there. I don't have any agenda about what tech we use to solve this problem -- I'm perfectly happy with solar, wind, or flower power if that's what it takes :)