References in published papers point to previous works. So it has always been possible to follow the graph backward, reference to reference, scooping up more and more of the history of a topic. A consequence is that it has been advantageous to cite later papers for review purposes rather than earlier ones, since they potentially link to more literature.
With Google Scholar and such tools of the internet, this bibliographic arrow of time has lately been eliminated. Now, given a paper X, it is easy to find the subsequent papers that cite X. Links can be followed in both directions. In choosing which paper to cite for review, there is less of a case for preferring the recent one.
All this is in a world where each publication at least has a fixed timestamp. Though it makes me uncomfortable, this too may be changing.
[17 September 2021]