Um, no. He didn't attack their methodology. Learn to read.
Silver's right. He's not challenging PPP's accuracy like those poll re-weighting morons or calling them discredited. He's just pointing out that yes disregarding a poll you conduct because you don't like the results is not scientific or correct. And worse of all this will just fill ammo for the critics, every PPP poll will get dismissed by right wing trolls in 2014 with this being the grounds (and perhaps ironically Silver's model as well due its inclusion of PPP.)
Cohn's criticism of the methodology is a bit more open to attack since yeah PPP is accurate despite all that. But I don't see anything wrong with criticizing PPP for this as it doesn't involve their accuracy.
I don't know how it's accurate to have Obama leading by 8-10 points in Colorado and Virginia during the summer of 2012 when nearly every other poll had it much closer.
Because Obama had a strong lead during the summer and all the other pollsters were wrong?