Help! An 8 year old asked me how to build a nuclear power plant I would really like to give an explanation similar to this one.
Here's my current recipe:
(i) Mine uranium, for example take a rock from here (picture of uranium mine in Kazakhstan).
(ii) Put the rock in water. Then the water gets hot.
(iii) [Efficient way to explain that now we are done with the question]
This seems wrong, or the uranium mine would explode whenever there is a rainfall. Does one need to modify the rock first? Do I need some neutron source other than the rock itself to get the reaction started?
As soon as I have a concrete and correct description of how one actually does I think I can fill in with details about chain reactions et.c. if the child would still be interested to know more.
 A: Everything is made of tiny things called atoms. All atoms have a tiny center part called the nucleus. Some atoms have an unusual type of nucleus that, every once in a long while, randomly explodes, sending tiny pieces in all directions. Normally those tiny pieces just bounce around until they join another atom. However, if you have a bunch of the right kind of exploding nuclei together, the exploding pieces of one nucleus can hit other exploding nuclei, and make them explode immediately, then those pieces hit even more exploding nuclei, and you get a chain reaction, sort of like dominoes.
To make a nuclear reactor, you dig up a bunch of rocks with the right kind of exploding atoms, and you carefully remove many of the other atoms so the exploding atoms are close enough together to make a chain reaction, then you put them in water*. All the exploding nuclei produce a lot of heat, which boils the water. The steam turns a fan, which spins a magnet, and creates electricity. You have to be very careful that you don't put too many of the pieces with exploding atoms together, or the atoms will explode too fast, and reactor will get too hot.
*If you want to get into more detail, you could explain that the exploding bits are going so fast, that they usually pass right through the other atoms, cartoon-style, unless you have other atoms, like those in water (a moderator), for them to bounce off of and slow down. You could also explain that reactors use "control rods", which are made of atoms that easily absorb the exploding bits, and therefore slow down the chain reaction. So, if they push the control rods further into the reactor, the chain reaction slows down more.
If you want to include more terminology:
Rocks = Uranium ore
Removing all the other atoms = enrichment
Nucleus exploding = nuclear fission
Exploding atoms = radioactive atoms (often Uranium)
Exploding pieces = neutrons (and some other particles)
Fan = turbine
A: RTG
The described approach mirrors https://en.wikipedia.org/wiki/Radioisotope_thermoelectric_generator devices used to generate power in, for example, space probes - simply take some radioactive material and extract energy from its decay, without needing to control a chain reaction or something like that.
A: 
This seems wrong, or the uranium mine would explode whenever there is a rainfall.

A natural nuclear "reactor" probably existed at Oklo, Gabon

The natural nuclear reactor formed when a uranium-rich mineral deposit became inundated with groundwater that acted as a neutron moderator, and a nuclear chain reaction took place. The heat generated from the nuclear fission caused the groundwater to boil away, which slowed or stopped the reaction.



Does one need to modify the rock first?

No, you just need enough of the right kind of rock in close enough proximity.
Nowadays, on our planet, most of the right kind of rock (containing lots of U235) has turned into the wrongish kind of rock (mostly U238 and U234) by the natural process of nuclear decay.
So you need to separate out the right kind of stuff (nuclear fuel) from your rock. This is done by a complicated process (gas centrifuge).


Do I need some neutron source other than the rock itself to get the reaction started?

The rock produces neutrons. You usually need a moderator to slow your neutrons down, water will do.
A: Well, if you have a really adventurous kid you can follow the recipe of the Radioactive Boy Scout

...and became fascinated with the idea of creating a breeder reactor
  in his home. Hahn diligently amassed this radioactive material by
  collecting small amounts from household products, such as americium
  from smoke detectors, thorium from camping lantern mantles, radium
  from clocks and tritium (a neutron moderator) from gunsights. His
  "reactor" was a bored-out block of lead, and he used lithium from
  $1,000 worth of purchased batteries to purify the thorium ash using a
  Bunsen burner.[2][3]
Hahn posed as an adult scientist or high school teacher to gain the
  trust of many professionals in letters, despite the presence of
  misspellings and obvious errors in his letters to them. Hahn
  ultimately hoped to create a breeder reactor, using low-level isotopes
  to transform samples of thorium and uranium into fissionable
  isotopes.[4]
Although his homemade reactor never came anywhere near reaching
  critical mass, it ended up emitting dangerous levels of radiation,
  likely well over 1,000 times normal background radiation.

However, that bit about Tritium seems suspect and is probably wrong
A: You may need to explain the concept of a chain reaction, at least enough to explain why the rocks don't explode.  Every fission event in Uranium generates neutrons.  This occurs naturally at a slow rate, or can also be triggered by a Uranium atom getting hit by a neutron.  The more densely packed the Uranium is, the more fission occurs (this is drastically simplified, skipping the issues of neutron speeds, but good enough to expand on later).  Rocks with a lower Uranium density do heat up water, but not to the extreme that concentrated Uranium does.  The larger the volume of Uranium, the more neutrons hit other Uranium atoms.  The more Uranium there is per volume, the more neutrons hit other Uranium atoms.  This is why we enrich Uranium for use in reactors.
Then you can show them this example of what chain reactions look like.
