> He was holding a presentation on nuclear power and mentioned safety, claiming that according to the best models there would be one large nuclear accident ever

> He was holding a presentation on nuclear power and mentioned safety, claiming that according to the best models there would be one large nuclear accident every million years. This part reminds me of how NASA's calculations for crashes/failures were very low, like 1 in a million officially, but when casually asked most engineers estimated it at 1 in a few/one hundred (my memory is fuzzy) - which was about the actual rate of crashes. Humans, even smart ones, have biases unfortunately. (I personally still think nuclear is the best, even if we end up with ~3 more medium accidents in the 21st century.)