The USPTO is overwhelmed and understaffed, so I can understand why they have poor QC on patents, particularly when they require technical knowledge to determine triviality.
What I don't get though is why it seems to be such a challenge to prove the invalidity of these crap patents in court. The google loss to bedrock (the linux kernel troll) is a high profile example of this. Maybe the standards that need to be looked at are around litigation and if enough crap is thrown out, all the incentive is gone and all of the sudden the USPTO application volume goes way down.