From what I'm reading, here are some issues impacting the quality of the new Panda algorithm:

1) keyword in domain title is given too much weight

2) Algorithm cannot detect the original piece of content amidst a bunch of copies

3) Scrapers outrank the original content, even with extremely shallow sites

4) *If* domain wide penalties have been imposed, they unfairly penalize good content along with bad

5) Algorithm is unable to account for content that is legitimately and necessarily copied, such as identical MLS or product listings

6) Some content farms still rank oddly high (e.g. eHow) compared to others