FTFY. Simply because he's correct on some things, doesn't mean he's infallible.

Yeah I know it exists :) I was just demonstrating that the technology already exists. The prospect of a true hypersonic missile would be to stay inside the atmosphere, relatively close to the ground, yet provide the capability to strike a target a short-to-intermediate range (<1000km, probably way less) with little to no over-the-horizon warning time. Something like the P-800 Oniks (which can reportedly do M4+ over a distance of ~100km), just much faster.

Yeah, I'm aware of the problems with corrosion of high-temp lead, though this can somewhat be kept in check by limiting the temperature (it hurts efficiency, but might be worthwhile - really depends on the design). It really comes down to materials science and like you say, research on lead cooling has been thin so far. I'd love to see a lot more open-access research in this area.
As for sodium cooling, it has its own problems, but AFAIK it doesn't react explosively in air. It can burn, but if I understand it correctly, the reaction isn't exothermic, so it'll quench itself quite rapidly. Now water is a different story. A sodium-cooled reactor in a potential flood area (ahem, Fukushima Nr.1) is about as dumb an idea as I can think of. Fortunately, the improved thermal efficiency gives you some more leeway to run pumps and pump the water up a hill (like reactors #5 and #6 at the Fukushima Nr.1 plant, which were largely unharmed by the tsunami which totally devastated their lower-sited siblings).

I'm not a fan of light water reactors either, but you need to understand that the public isn't aware of the details and intricacies of reactor design. To them terms which make an engineer cry happy like a little girl, don't mean anything. I mean FFS most of them still think nuclear reactors can explode like atom bombs. They saw Chernobyl and Fukushima, they saw "boom", it's a nuclear reactor, therefore "nuclear boom".
I also think and hope education can change that, but that's a long road ahead and TRU-burning BWRs could work in the interim to help jumpstart that (since we already have them).

You've got a mistake there, 2223K is the flame temp for CH4. Natural gas is 2233K or 10K higher (because it's not pure CH4). The difference, however, is largely inconsequential, it's a 1% increase in absolute flame temperature (exactly like I said), at best (actually 0.44%, but we'll let that go).

Now hang on, you can't just take the flame temperature and call that your working fluid temperature, that's not how it works in heat engines. Gas-based heat engines (CCGT - the most efficient ones, not talking about ICBs, those have very poor efficiency) use the hot flame to heat a working fluid (typically superheated steam), which is much cooler than the flame itself. I used an 18C coolant temperature (cold water) and 60% efficiency to back-calculate the minimum ideal working fluid temperature (~454C hot vs. 18C cold will give you ~60%). So a 1% increase in absolute flame temperature can at best give you a 1% increase in absolute working fluid temperature, which means your working fluid goes from 454C (727K) to 461C (734K). However, the gains here will be much more modest because you can't just step over a heat exchanger's maximum temperature willy-nilly, or bad things will happen to it, which is why I rounded to 460C (I'm a generous guy, I know :)). So 460C hot vs. 18C cold gives you 60.3%, or about a 0.3% gain.
Now as for "direct thermal usage (water/hot air/etc)", these applications are not heat engines, they don't convert heat energy into work (J -> J/s is a heat engine, J -> J is not), so for them an increase in flame temperature means nothing. For example, it takes the same amount of heat to raise 1kg of water by 1C, regardless if the absolute temperature of the heating source is 100C or 500C, it's still 4.18kJ/kg.C.

Understood, thanks.

Had to google the abstracts of the report and its conclusions are highly interesting. They claim to be able to breed at a ratio slightly above 1.0 in a BWR and even slowly consume TRUs by 10% per reprocessing step with unlimited reprocessing capability. Results of the report:

This has the potential to be a game-changer if true, as we could simply use existing reactor designs such as the ABWR (of which there are several operating already) to both burn waste and breed fuel indefinitely from U238 feedstock.

It's possible they plan to only burn stuff beyond Pu in there, as that can already be consumed in MOX (which however produces more of the higher TRUs for the reasons you noted). It's really hard to tell what they're trying to do here without more detailed data on the actual fuel composition.

It's possible they're having trouble getting a dedicated TRU burner design approved and built (there might be little economic incentive and much public opposition to new nuclear plants, no matter the safety of the technology), hence why they might be motivated to try and design fuel that can consume TRUs in standard BWRs, of which Japan already has quite a few.

I agree that burning the crap off is a good thing, but why tack on an expensive piece of extra equipment when pretty much the same effect can be achieved by being smarter about core design? I'm just not seeing the big advantage here.

I never quite understood the allure of ADS. To my eyes it just looks like an exceedingly difficult way of achieving criticality. Given a good design, a reactor will self-regulate by its own negative temperature coefficient, so an external driver isn't strictly necessary and shutdown can be performed by passive systems that are equally dependable as cutting power to the accelerator, e.g. by suspended or spring-loaded SCRAM rods. There is the interesting proposition of not having to reprocess the fuel when running a thorium breeder cycle in order to extract the bred fissile and load it into the core, since one can boost the neutron budget externally, but that needs to be weighed against the pretty steep cost of a high-powered accelerator (in terms of current, not just particle energies) and accelerator reliability issues.

They don't, but the ratio of absorption to fission in the thermal spectrum for them is pretty bad, so that can mess up your neutron budget. Depends on the exact composition, though - each reactor produces a slightly different mix and that makes the TRU content in spent fuel fairly heterogeneous, which complicates reactor design and makes fabrication of reliable fuel fairly expensive (hence why MOX fuel only contains the Pu content, not all the other TRUs and even so it's much more expensive than fresh Uranium fuel).

I don't think they do so in the breeder cycle - their neutron loss margins are fairly thin, hence why most designs propose extracting at the Pu-238 step (unusable for weapons, but great for space batteries). The burner cycle might be better in this regard. Fast reactors are able to do it, they have plenty of neutrons to spare.

What is "neutron saturation transmutation"?
I'm also skeptical of their claims, as it appears to be a thermal-spectrum light water reactor and it's quite difficult to consume TRUs completely in the thermal spectrum, the neutron absorption cross sections are fairly large. Maybe they've got higher enrichment and so shitloads of excess reactivity, so they can afford to lose the neutrons, in which case I seriously hope they have a strong negative temp coefficient. Don't know, would be good to learn the details.
Not sure about the likelihood of meltdown being increased, though. I don't think the decay heat profile of MOX is significantly different from regular enriched Uranium fuel (decay heat melted Fukushima fuel, not fission heat).

What if your point *is* to cause fallout badness.

I'm not convinced of the first-strike capability of hypersonic vehicles. Even at fairly highly hypersonic speeds (M10), the vehicle still takes considerable amounts of time to travel a substantial distance (1000km takes about 5 minutes at M10) - by that time satellite-based detection systems can react and a ground-based counter strike can be initiated (modern ground-based ICBMs can launch in less than 30s, and SLBMs are also an option). At certain distances a good exo-atmospheric missile on a depressed trajectory can strike faster than that. Assuming a 90s boost phase with the final ~10s being used to depress the trajectory arc downwards + a few minutes to travel the 1000-2000km towards the target at easily 4-5km/s. Military solid-fueled missiles have very high thrust-to-weight ratios to shorten the boost phase as much as possible. I think the more problematic aspect is that defending against low-altitude (well, relatively, I mean we're still talking 10-20km in altitude, otherwise the air resistance and shock heating just kills it) hypersonic vehicles in a local theater war scenario can be very difficult - at 10km altitude with the over-the-horizon flight time you only get maybe 30-60s of warning (by my rough calculation at 10km horizon is ~250km away) - depends on radar position and capability, of course.
Your analysis on usability by crazy/unstable countries, I think you're spot on. The big boys have bigger and perhaps more capable toys. It's those crazy wackos who might be tempted (Iran to Israel is only about 1000km, as is NK to Tokyo, so 5 minute strike capability would sound like a sweet deal there).