If it makes you feel any better, many types of reactors produce helpful medical isotopes as a side effect of operation. So when looking at lives-per-megajoule nuclear is one of the absolute best, even accounting for some of the accidents.
Coal is incredibly dangerous directly - mining and operation are generally unsafe. Hydro has had some terrible accidents like the Chinese Banqiao dam rupture in 1975 which killed tens or hundreds of thousands. Solar and wind I'm not too sure about, but nuclear power has tens of deaths total attributed to it.
The nuclear-relate deaths that occurred have been from using light-water reactors, which are essentially very-high pressure high-temperature radioactive containers. Modern designs like molten salt reactors (like LFTR) have the potential to be safer, but don't have regulatory support and are mired in the process [2].
I made a whole infographic about the safety of nuclear power, but I'm not an engineer or journalist [3]
[1]: https://www.forbes.com/sites/jamesconca/2012/06/10/energys-d...
[2]: https://www.aps.org/units/fps/newsletters/201101/hargraves.c...
[3]: https://create.piktochart.com/output/37686680-safe-nuclear-p...
Hydro's mortality is largey attributable to a single, though immensely bad incident, the Chinese Banqiao Dam disaster of 1975. That itself was a consequence of poor engineering, planning, management, response, and an unanticipated weather event (parked typhoon/cold-front dumping > 1,000mm rain in 24h), with most issues being common to any high-profile technical engineering project.
Hydro otherwise is largely safe (there are a few other exceptions, again higghligghting planning, management, response, and institutional integrity, especially in developing regions or periods), problems not specific to the energy-harnessing mechanism) and the Banqiao site is presently home to > 15 millions, rather than being a multi-century technically-induced wildlife park as is the case with other technologies.
Absolutely the same can be said about Chernobyl (well, apart from the weather) and Fukusima.
Though to clarify: as non-engineering failures, there is no technical engineering fix to this problem.
There are literally tens of thousand of dams, if not > 100k, worldwide (over 22k in China alone), with well over a century of major utilisation. Granted 100 or fewer major (2+GWe) installations.
There are fewer than 400 nuclear power stations worldwide (3911 by my count from Wikipedia: https://en.wikipedia.org/wiki/List_of_nuclear_power_stations...), the bulk in operation for less than 50 years. We've had three major disasters and numerous close calls, all with massive long-term consequences. Building out infrastructure to the scale of present or future (industrialised-world electric provisioning to 8-12 billion inhabitants) would see even at much better safety records* a major accident every few years. And that's only one disadvantage.
Notably, once floodwaters recede, even a catastrophic dam failure returns to livable condition rapidly: weeks to years, rather than centuries.
And, as noted, it's human, social, business, and organisational factor common to any major high-value asset technical project generally to blame.
(This was not a large value of 400.)
This is very similar to nuclear.
> Per IPCC, wind and solar are very nearly on par with (possibly better) than nuclear.
Mind sourcing this? every place I've looked shows nuclear beating solar by 5x and wind by 2x.
Analysis unit was deaths per GWh or similar.
I'll see if I can find it though I'm not optimistic.
Our World In Data posts similar stats, premature deaths/yr at 1 TWe/yr generation:
Coal: 25
Oil: 18
Gas: 3
Nuclear: 14--100
Wind: 29 years
Hydropower or solar: 42 years
Solar: 53
I'm skeptical of this figure - Chernobyl alone is expected to cause ~4k+ deaths due to radiation exposure.
I'm sure nuclear is still significantly lower than combustion fuels, but I'd expect solar is somewhat safer.