I think in the case of Google+, the boxes are necessary because of the dynamic capacity of the posts to rearrange themselves based on the width of the window. The boxes allow for an expected boundary between each post. This really helps in scenarios where the content is layed out in a staggered grid formation. In contrast to a list where each element such as the avatar, header and body are in a layed out in a predictable format when read linearly, the grid can lay each of these elements at varying heights (and maybe even widths) adjacent to each to other.
In short, the unpredictable nature of the layout of each item in the grid calls for some necessary boundary while the predictable nature of a list view doesn’t since the content itself serves as a cue to the start and finish of each post.