There are a whole bunch of linked issues here that have to be dealt with as what they are -- linked issues.
We've spent a literal 40 years screwing around with energy policy and tying it to defense, which has led to terrible distortions in both areas. That has to stop but it's not easy to stop, because there are very-entrenched interests that like it the way it is.
Second, it takes a LOT of energy to produce a vehicle. Simply scrapping the existing ones means you pay twice -- you throw away the investment in the existing vehicle's production (in terms of energy) that you never amortize out and then you spend again to replace the transport capacity. That's actually pretty dumb when you add it all up.
Fukushima happened for a number of reasons; one of the more-serious problems was that their switchgear got flooded and there were no spares nor access to them in prompt fashion. Without switchgear you can't power the pumps and without pumps you're dead in a conventional nuclear plant. Losing grid-tie was part of it, having their generators flooded out (with the exception of two) added to it, and then the destroyed switchgear was the cherry on the cake. All of this was avoidable but humans make mistakes and the "rule of 3s" got them -- the first error rarely kills you, the second is bad news, and the third is normally the one that takes you over the cliff.
LFTRs are not a perfect technology but they solve two problems at once -- fuel supply and waste reduction. They also produce much-higher quality process heat. One of the very serious problems with traditional nuclear power is the quite-low quality heat they produce; this is caused by the use of water as a moderator and thus you are limited by the critical temperature (~374C.) While it's theoretically possible to go beyond that there are serious control and moderation problems that arise when you reach that point, so existing designs all stay away from it on purpose. A PWR, for example, runs at an exit temperature of about 315C, leaving a roughly 50C margin between operating temperature and critical temperature.
The problem is that this low-quality heat sucks. It leaves you with Rankine-cycle turbines for generation, which is why these plants need ridiculous amounts of cooling water and are thus all sited near huge fresh or seawater heat sinks (big rivers, lakes or oceans.) That of course exposes them to serious "natural event" risks (e.g. tsunamis.) It also is why the thermal efficiency is typically in the 30% range; the rest of the energy is literally thrown away. Thermodynamics, in short, sets the limits beyond which one cannot operate, just as is true for vehicle engine efficiency.
Incidentally, the Canadian CANDU design is a derivative of the standard PWR that Canada developed because it can run on unprocessed natural uranium and is capable of being refueled while in operation. Canada went this way due to not having access to enrichment technology at a reasonable cost when they were developing their nuclear program. The price of doing what they did is that CANDU uses heavy water as a moderator which is god-awful expensive.
The are also gas-cooled and liquid-metal-cooled designs out there but they have problems as well; Fermi I was a liquid sodium commercial power-producing plant but the others have been experimental, fuel-producing or research in their operational goals. There are a number of reasons why these designs have not turned into commercial successes with some of the most-serious being positive-temperature-coefficients in some of them, the potential to assemble a prompt critical mass by accident (theoretically possible in all breeder designs and most fast-neutron designs) and thermal conduction issues (common in gas-cooled designs) that limits the rational output that one can obtain via scaling up for commercial use.
The LFTR was originally designed with the intent of attempting to power an AIRCRAFT with a nuclear reactor. That didn't work out so well but the technology is wildly different and has a number of very attractive features compared to the traditional uranium/plutonium fuel cycle. You must reprocess online with an LFTR to remove neutron poisons, which is one of the "greenie weenie" objections; the NIMBY folks target reprocessing because without it all nuclear power systems will eventually choke on their own waste and have to be shut down. The LFTR, however, produces only about 5% as much high-level waste as does a traditional uranium-cycle plant as it inherently avoids producing much in the way of transuranics and what it does produce it burns up quite efficiently, extracting the energy instead of wasting it in a used fuel pool or burying it in a mountain somewhere. It also runs at a process temperature of about 650C which is of dramatically higher quality than any water-moderated reactor, and this makes for
much higher thermal efficiency. Combined thermal efficiencies approaching 50% are very much within the realm of reason for such a plant, where you simply can't get out of the 30s with a PWR or BWR design. This in turn means that the plant's thermal output is about half of what an equivalent PWR or BWR requires, which in turn means it burns less fuel over time. It too is fueled online and is passively safe; since the moderator is in the reactor vessel and the fuel dispersed in the working fluid in an emergency simply draining the working fluid into containers of sufficient linear space allows passive cooling of fission byproduct heat and an unattended, no-machinery shutdown. Because the fluid does not boil there is no need for high-pressure-rated piping and containment systems, simplifying the design. If there is a pipe break the affinity of the fluoride salt for other elements will tend to bind the reaction products rather than release them into the environment and in addition once the fluid cools it solidifies and is thus much easier to clean up and contain than a water release. One material negative is that due to the reactivity of fluorine Hastelloy is required for the reactor vessel and piping, and it's expensive, but the material itself is a known quantity and is regularly used in chemical processing plants today for its corrosion resistance (in other words its expensive rather than being an engineering problem to solve.)
There's a lot more but I'm probably boring people here...
(Yes, I'm somewhat of a physics and chemistry wonk in addition to other things... my next-door neighbor, during my childhood, was one of the plant physicists at Fermi I and I developed quite an interest in the technology behind nuclear power in my youth and have maintained that interest and investigation over time...)