UPD: Switz tried to make underground reactors and got what is now the most contaminated cave in the world. http://en.wikipedia.org/wiki/Lucens_reactor
also, the underground thing worked - no ? it contained the radioactivity - wikipedia says no pollution occurred... in the case of a ground level one, I'd guess the Vaud canton would be pretty f-up
Well you see there's design, like I designed a giant spacecraft when I was a kid, all on one piece of paper, and then I talked to people about it because I thought it was cool, and there's "really design" like handling mfgr issues, clearance/fit/wear issues, failure tolerance, etc. Thats how my spacecraft only took one side of one piece of paper but the Cessna I took lessons in had about six linear feet of shelf space for the A+P manuals for all the different parts and subsystems thats just enduser maintenance stuff, not actual design docs. I haven't watched the video, but my guess is the kid doesn't have a design of his own. If a design of a real reactor takes 1000 engineer-years from the best minds in the field, and the kid is not that old yet...
I've had kind of a nuclear interest for quite a bit longer than the kid's been alive. Its an interesting hobby, seeing how much you can learn. As a hobby its no weirder than seeing how much you can learn about the phone system, or the internet.
The fundamental problem we're discussing is balancing optimum nuclear characteristics (neutron activation, long life, capture cross section, etc), mechanical/chemical/safety/lifespan properties, and mfgr/supply issues, can never really optimize all three at the same time.
The Lucens reactor local optimization maxima for fuel cladding (the tube/shell the fuel sits in) ended up as Magnesium. Get it hot enough and it burns spectacular in pretty much anything, CO2, air, water... Also it corrodes pretty well underwater, or heck even in air. So burying a super flammable fuel underground and heating it up isn't so wise in retrospect although I'm sure as a local maxima it looked like a good idea. If we lived on an oxygen free planet then Magnesium would be quite a handy general purpose metal. Unfortunately on the earth...
Other popular fuel cladding over the years for gas reactors has been stuff like beryllium (WTF were they thinking?) and stainless steel.
I suppose the guys who selected Mg for cladding did pat themselves on their backs for not using a Be based cladding, which is probably the only way it could have turned out worse.
Needless to say, stainless steel in CO2 doesn't burn very well. Maybe that's the kids design, its certainly off the shelf and fairly conventional. Lifespan might be shorter, ruinous cost, and the worst part is stainless eats neutrons for dinner which aside from making a short life means huge waste of fuel and heating issues, its kind of like storing your fuel inside shielding crazy as that sounds. But as a local optimization its not bad for mfgr safety as stainless is pretty harmless (Your cafeteria is contaminated with stainless steel, oh noes!) and the flammability rating of stainless steel in CO2 atmosphere is as you'd guess pretty excellent.
Why all this fooling with cladding? Fundamentally nuke fuel seems to be awful structurally so you shove it into a pipe. Engineers are pretty cool with making heat exchangers outta pipes so a magic pipe that generates its own heat is pretty well understood. Its an object oriented encapsulation idea, sorta. If you made an indestructible fuel that didn't need cladding, then you'd either have wear issues on whatever it touches and/or processing/reprocessing would be terribly difficult dangerous and expensive and contamination issues so you'd just be moving the location of the accident, not reducing the rate of accidents.