http://arxiv.org/abs/1305.6289
Very readable and nice.

Also claims a result I had conjectured here before (I had

outlined a proof using Zhang, but never checked that the proof plan could

really be carried through)

is indeed true: there are infinitely many "de Polignac numbers."

One of the most impressive claims in Pintz's paper is this. Let a "near twin"

prime p be a prime p such that p,q are two primes with p<q<p+70000000.

Then: the near-twin primes contain arbitrarily long arithmetic progressions.