Radioisotope Thermoelectric Generators
Using radioactive materials for passive heat production (without the need for a reactor) is the simplest method, but power is inversely proportional to half-life, so after thousands of years, there won't be much useful material left. With a half-life of 432 years, Americium-241 is probably the best choice here. It is currently being considered for space probes due to the shortage of Plutonium-238. Even after thousands of years, the remainder should be seperable without too much problems, as the decay product is Neptunium-237, which has a half-life of 2 million years (and is useful in its own right).
Current fission processes all start with uranium. Natural uranium consists of 99.3% U-238 and 0.7% U-235. Uranium-235 is fissile, but Uranium-238 is not. It will eat neutrons though, especially at higher energies, which means you can't get a chain reaction in natural uranium without a very clever reactor encasing it. For most purposes, this means the uranium will have to be enriched, increasing the percentage of U-235.
Except for bombs, practical enrichments still contain mostly U-238, which means you will produce Plutonium-239 as it absorbs neutrons. Pu-239 is itself a good fissile material and has a lower critical mass than U-235, though if it stays in a reactor for too long, it will grow fractions of Pu-240, Pu-241 and Pu-242, which will make it unsuitable for gun-type bombs.
Fissile energy production
All fissile isotopes, upon neutron-induced fission, will release roughly 0.1% of their mass as energy, so big gains can't really be made here. Fusion will give you about 0.5%, but there are other answers that describe that.
An alternative fuel cycle starts with Thorium-232, the only primordial isotope of thorium. Like U-238, it's not fissile, but when it eats a neutron, it will turn into Uranium-233, which is fissile.
This fuel cycle is considered (by some) to be better because it doesn't grow transuranium elements (such as plutonium) by design, and as fissile isotopes (U-233 in this case) that fail to fission (and instead eat the neutron) will have extra chances to do so (U-235, Np-237 (fast neutrons only), Pu-239), very few transuranics will be produced in practice.
Up to this point, we've discussed things that are available on earth right now, in sufficient quantities. Some transuranics have half-lives sufficient to stash them away for a long time, while still having useful properties.
Neptunium-237, mentioned earlier, has a half-life of 2 megayears, can be used in bombs and fast reactors, and can be used to breed Plutonium-238, which is, for most purposes, the best RTG material.
Curium-247 is an isotope of Curium that has a half-life of 15 megayears. It is fissile and has a critical mass much smaller than the currently used fissiles. This enables the creation of small bombs (suitcase nukes) and small reactors. If you leave the stuff lying around long enough, it will grow fractions of U-235 (half-life 703 megayears), Pu-239 (24 kiloyears) and Am-243 (7 kiloyears), with U-235 being the largest fraction by far. Amusingly, this produces chemically separable, pure U-235. This can also be done simply by storing Pu-239 for long enough.
Other curium isotopes may yield even better results, but only Curium-245, with a half-life of 8 kiloyears, can make even a claim at being storeable.