> none of the reactors that have failed have been built using the "newest generation" of reactors, which are supposed to be much safer. At some point there has

> none of the reactors that have failed have been built using the "newest generation" of reactors, which are supposed to be much safer. At some point there has to be a compromise made between cost and safety. You could spend an extra $10 billion making a plant even more safe, but in doing so make the scheme unaffordable, and 20 years late. And even then the station could still have complex failure modes that are very difficult to model. Maybe that is exactly what happened to the nuclear industry since Chernobyl? We ended up with ever more complex and monolithic systems that require thousands of expensive physicists to do basic building and maintenance. Perhaps it would have been better to invest in small modular systems.