I'm not contradicting anything about this, but I would be interested to see a comparison between nuclear power plants and OTHER very large projects (the California Bullet Train, anyone? How about the Sydney Opera House?) It seems like they are mostly all getting worse, not better. It doesn't have to be this way. Bent Flyvbjerg, a Danish professor, has a book on this:
After I wrote that, I wondered if Flyvbjerg had already written on nuclear plants. He has the world's biggest database of large projects, so he must have.
It turns out he has, and they're in the quadrant of "dumb scale-up."
I lost all faith in Flyvbjerg in the 90's when he published his doctoral thesis. It was published as two volumes, one of them being a philosophical background. That volume was 1) mostly cribbed from the Dreyfus brothers (Hubert and Stuart) and 2) contained many very dodgy quotes from Aristotle. He had used '[...]' more liberally than any other author I've ever come across. The quotes ended up being like the Homer Simpson's interview from S06E09:
Hmm. The book doesn't have that, I don't think. It's pretty readable.
On the other hand, he does have some association with Oxford, which is selling a "course" on how to apply his methods of project estimation.
I wouldn't waste my money on the course, but I think the method of "look at similar projects, and take the average" is a much better algorithm than the bottom-up Gantt chart that everyone does. Shit happens, no one thinks it'll happen to them but it does, and the historical data includes all the Black Swan events that no one saw coming.
https://www.amazon.com/How-Big-Things-Get-Done-ebook/dp/B0B3...
It might well be that nuclear plants are even worse than other types of projects, but I'd like to see the proof.