Re: Joined up thinking....
" it doesn't ramp up and down very efficiently."
Nuclear using fuel rods doesn't ramp up/down very well (ie, can't load follow)
This is because of xenon buildup in the rods when you ramp down that has to decay away before it can ramp back up again (this also builds up a shitload of pressure inside the rods and the gassing causes the ceramic fuel pellets in the rods to break down to powder over time - both are undesirable, so ramping down is best avoided in a fuel-rod based design.
Molten salt fuel reactors don't have this problem, because the xenon (insoluble in Fl/Li salts) gasses out in the circulating pump headspace and can be extracted or left in situ to decay before being removed. This was amply demonstrated at Oak Ridge in 1968
Conventional MSRs aren't pressurised and can run far hotter than water-cooled systems, which in turn means more efficient turbines or process heat supply and no risk of radioactive steam explosions.
There's a UK fuelrod design variant which substitutes circulating molten salts inside tubes for the fuel rods. Whilst it's a simpler engineering change vs current civil reactor designs, it still puts superhot (400C), pressurised (20-40 atmospheres), acidic water (boric acid is dissolved in the water as part of the moderation process) in close proximity with radioactives which is a "very bad idea" in the overall scheme of things and means you still need a huge containment building with all the associated gubbins. Water isn't known as "the universal solvent" for nothing, and almost all the nuke incidents in the last 70 years have been because of water-related issues (corrosion of piping or rods) or compounded by the release of radionucleide-contaminated water/steam.
If a nuke plant is producing hydrogen for fuel (with further carbon tacking to the molecules to make 'em more easily transportable - hydrogen is a bitch to store and transport), then you do not want it as a load-following electrical generation system as well - large scale hydrogen generators or processors don't take kindly to variable inputs. In any case, a water-moderated nuke plant doesn't run hot enough to directly drive the water-cracking process (electrolytic generation is supremely inefficient, never mind the inefficiencies of uranium plants (mined vs fuel vs waste) and water-moderated electricity generation.)
Bear in mind that a molten salt system doesn't need to dump heat to water bodies (ocean or river), so you're not location-constrained to vulnerable areas, nor by hot days (dumping to atmosphere is sufficient and it's entirely possible you can scavenge more energy by using a vortex generator, bringing the overall thermal efficiency up from 35% to something like 45%, vs a water-cooled system's absolute best thermal efficiency of 28%)