As a general complaint, i feel like most of these "web/graphic design pattern libraries" miss the point of christopher alexander's original 'pattern language' idea, which specify cataloging as only one part of actually developing a larger, expressive, design language.
Still, the problem with these as actual 'design patterns' is that they're often wedded to specific implementations rather than the tasks they're trying to solve. I.e. "Endless Scrolling" (to pick an example) is not a pattern, it's a solution to a larger problem (e.g. user retention, communicating large datasets, immersive experience, &c).
The point being that if you're talking about design patterns in terms of the implementation (web forms, 404 page) you're restricting your pattern language to specific solutions rather than understanding what problems you are actually trying to solve (i.e. I WANT TO GIVE YOU MONEY, WHERE DID THIS GO?, respectively).
Although these sites often address these problems, they do so as a second level consideration rather than the primary atom of their pattern language, which diminishes their power. This is generally part of the problem with how design patterns are discussed and understood, which is a bummer, because it promotes the perception that they are 'reference catalogs' rather than 'languages for addressing problems.'
edit: fixed ambiguous language in last sentence
