[bell-cot]

If you have good engineers, using molten sodium as your primary coolant does have a fair number of safety advantages over water. The stuff boils at 1,621 °F - so you lose 99.9% of the traditional (water cooling) problems with ultra-high pressures and boiling dry. There is no risk of hydrogen gas explosions if things go badly wrong. Thermal safety margins are far wider. Etc.

And if your engineers aren't good enough to trust with a mere sodium cooling system, then they certainly aren't good enough to trust with anything nuclear.

[rbanffy]

The history of nuclear power (and pretty much every “difficult” field) is full of examples of engineers and technicians who were simply not good enough.

Coolants will always leak, unexpected things will always happen, and our only safety nets are the ones we build ourselves, so, no. We can’t be trusted and, most importantly, we can never trust ourselves that much.

[bell-cot]

True-ish.

But centuries before the first nuclear power plant fired up, boilers full of pressurized hot water were notorious for killing people. Both those who were "simply not good enough", and innocent bystanders. Leaking low-pressure sodium, in a facility designed for it, is vastly easier to contain than water kept liquid by 2,000+ psi pressure.

Where it has been a very high human social priority - such as in the US Navy - nuclear power's safety record has been very good. Where it has not - well, Dilbert, Inc. can kill people at scale without going anywhere near nuclear power. Modern society offers food processing, mining, chemicals, transportation, and many, many other opportunities.

And you don't need any "difficult" fields to get that effect. Human nature and capitalist optimization will ensure that ill-trained and poorly supervised idiots in "waiting to happen" facilities will cause disasters.

[bell-cot]

One classic , at-scale disaster, in a should-be-easy field:

https://www.michigan.gov/mdhhs/safety-injury-prev/environmen...