Anton now doubt a nice person but not as knowledgeable about molten salt reactors as he could be. First, on the safety mechanism that drains the reactor tank into much larger drain tanks that separate the fuel enough so that it becomes non-critical, it is a melt plug, a plug of reactor salt/fuel that is frozen and kept that way by active cooling, if that cooling fails or if the reactor overheats (essentially impossible because of the self-regulating nature), it melts, and drains the salt / fuel into a much larger drain tank or tanks where the fuel is too spread out to remain critical. Since it is entirely passive, mechanical failure of this device is impossible, except for the possible failure of active cooling that under normal operations keeps it from melting, and if that happens the reactor shuts down. Another thing Anton did not mention is that the fuel cycle of a liquid salt reactor involves continuous reprocessing so fission products are continuously removed and thus when the melt plug melts, the fuel does not generate sustained heat like it does in a light water reactor even after being shut down. Third these reactors can burn uranium or plutonium and in fact need it to get started, there after they breed thorium-232 which is a fertile but NOT fissile material into uranium-233 which IS fissile, further they also breed any of the actinides that are not fissile into fissile isotopes so they burn all the actinide products, which are the portion of normal nuclear waste that is normally very long lived taking up to a million years to cool off to the point where they equal the radioactivity of the ore they were mined from, and leaves only fission products which are at the same level as the ore they were mined from within 300 years, transforming a million year problem into a 300 year problem. And there are industrial uses for many of those so even not all of that need become waste. Another reason these reactors make particular sense to China is that Thorium is a byproduct of rare-earth refining and China is the worlds largest supplier of rare Earth's thus they have a lot of thorium trailings from these operations so this fuel is essentially a free byproduct of another lucrative operation. Regarding the corrosive issues, the corrosion issues actually were more with another related technology, there were molten sodium reactors, sodium IS highly reactive both with metals containing it and air. Japan has operated these reactors and had fires result from leaks because sodium will spontaneously combust when exposed to air. Lastly, Anton is WRONG about not needing water, water may used in the secondary cooling loop or a secondary salt may be and a tertiary loop with water is used to turn turbines, at al rate, water is needed at some point, and is turned to steam to turn generators, it just is not used in the reactor vessel itself which makes the whole thing infinitely safer. Also, Anton is wrong about the purpose of the graphite rods, they are not control rods, they are moderator rods, the nuclear cross-section of most fissile isotopes is much larger to slow neutrons than fast ones, so by slowing the neutrons the graphite rods increase the reaction rate, but some fast neutrons are needed for breeding which is necessary to burn thorium or u-238 (u-238 is bread to pl-239, thorium-232 is bread to u-233). They do have control rods but they contain a neutron absorber rather than moderator, such as beryllium. However, they largely self-regulate and adapt their power output to the load automatically by thermal expansion of the salt without operator input. Conventional light water reactors can only burn the U-235 isotope of Uranium which is only .7% of the natural uranium and because of the accumulation of fission products they can't even burn that bit completely. Molten salt reactors can also burn the U-238 which is 99.3% of natural uranium and can achieve a complete burn as fission products are continuously removed. One product that is produced that is radioactive and ultimately vented to the atmosphere is tritium. Tritium has a half life of 12.3 years and therefore is highly radioactive, but because it has a short biological half life, that is to say it does not stay in your body long, it is not as biologically dangerous as that half-life would indicate, further being extremely light, it rises high in the atmosphere. Now, if we had controlled hydrogen fusion also operational, we'd have a demand for that tritium so it would no longer be a waste product.

@Billy Wild One word, bombs.