Nuclear power generation, as it is practiced today, is: inefficient, dirty, dangerous and water-dependent. It does not have to be that way.
A few weeks back there was a diary by The Anomaly that seemed intriguing but I did not quite get it (there are technical details there which I will not repeat here, so go take a look). Then a few days ago I came across the same subject in a TEDx talk by a guy who used to be a rocket scientist and has taken up the cause of Thorium based nuclear power in the form of Liquid Fluoride Thorium Reactors (LFTR.) The few minutes of the talk were interesting, to say the least, so I sought out more. I admit that I was hooked.
Pretty much everything wrong about nuclear energy today, and there is plenty, can be traced back to the sixties. Pressurized water (heavy or light) reactors extract less than 2% (to be generous) of the energy in the nuclear fuel. High pressure and temperature mean that when something goes wrong (and it will), it goes wrong very badly. They need high volumes of water for cooling which limits their locations and makes them vulnerable. High inefficiency means that there is plenty of nasty radioactive waste. It also means that we will run out of nuclear fuel at some point.
LFTRs operate at normal pressures and extremely high efficiencies, in the high 90s. They generate and recycle the fuel, which is so abundant that we could operate civilization for millennia and not run out. Almost all the fuel is consumed, so there is very little in the way of waste. And they don't need water as the turbines run on gas exchange rather than steam. They are highly scalable and, this is important, throttleable. In the initial design the idea was to use them in aircraft. This will probably not be feasible due to weight and safety constraints but spacecraft do NOT share those constraints.