Many readers will remember the "sodium hydride as an oxidizing reagent" story from last year. (For the non-chemists in the audience, the problem here is that sodium hydride is most certainly not what you'd think of as an oxidizing reagent, quite the opposite, in fact. Seeing the paper's title was, for an organic chemist, a bit like reading about a new way to sweeten drinks with vinegar).
This was famously live-blogged over at Totally Synthetic and picked up on around the chemical blog world. The current thinking, though, is that adventitious oxygen is really doing the work here. If you run the reaction under strict inert atmosphere conditions, you get no more oxidation. (And it still doesn't appear that any note has been added to the original JACS paper). Update - make that no note added to the abstract page. The paper itself is still accessible, although it does have notes that it was withdrawn.
Well, now we have another one. This paper in press in Tetrahedron Letters claims oxidation of benzoins to benzils with good ol' sodium hydride. In this case, anyway, the authors (from Korea) did try running the reaction under inert atmosphere, and saw their yield go down. Their proposed mechanism involves molecular oxygen, in fact, and is quite plausible. (I've seen anion-oxygen chemistry myself - if you deprotonate Strecker amines of benzaldehydes, you'll convert them into amides via oxygen in your solvent, that is, if you don't saturate things with inert gas first). Still, I'd rather that they titled this paper differently, since it's not sodium hydride that's doing the oxidation here. You could probably get this to happen with NaHMDS, t-butoxide, or the base of your choice.
And, weirdly, the authors (as far as I can see by going over the PDF) manage not to cite the original JACS NaH oxidation paper at all. You'd never think that anyone had tried this before, especially not in one of the most high-profile chemical journals in the world, just last year, with plenty of added press coverage. What does it take a get a paper cited? Update: given the withdrawn-but-still-available status of the original, this becomes a trickier question. The earlier paper seems to have clearly gone through the same sort of chemistry, but the mechanism - and thus the point of the whole paper - was misassigned. Do you cite it, or not?