Sooner or later we'll need the waste to feed our breeder reactors. When you have pushed the uranium through a reactor and spent it, it still contains 59/60 of its original energy. Breeder retrieve the remaining 59 parts.

Peak oil is not an energy crisis. It is a liquid fuel crisis.

I'll briefly entertain the notion. On instincts, I would pick the Mariana Trench as one of the better candidates. Not only because of its depth, which gives the nuclear waste some headway, but also because the subduction of the oceanic plate has one of the steepest angles observed. As far as crust instability goes, the Pacific oceanic crust at the Mariana trench can hardly be beaten. (Crust instability has to do with crust isostacy and it is time dependent, related to the growth of the lithosphere. The older an oceanic crust gets, the more it wants to sink back into the mantle.)

Beside the point of retrievability Starvid highlights, there is the problem that dumping the nuclear rods does not guarantee immediate entry. Hence the containers will have to be absolute contamination proof. The option Plan9 posted is in that respect interesting: jettisoning the containesr into the sedimentary accumulated clays which work as a backup guarantee in case the containment barrels crack. I didn't know that option before, but it could solve helping the risks of relative short term contamination.

However... Subduction zones also are know to scrape off whole slivers of soft sediment of the subducting crust. It's a bit of a convoluted subject; they're hard to study. I'd say, at first thought, there's a real risk that the sediment with blobs of rod-containers just get piled onto the ocean floor. Do we want that to be our legacy? In the end (the really long end) this will pose no problem as all oceanic crust gets recycled. I'd like to hear Plan9's view on this, too. I'd think that even if the ocean sediments get sliced off, the radiation levels would have significantly dropped by that time. But still. Not exactly a clean-job in that case.