The Atlantic:

We’ve been trying to figure out the moment Twitter turned, retracing tweets to see whether there was something specific that soured the platform.

Something is wrong on Twitter. And people are noticing.

Or, at least, the kind of people we hang around with on Twitter are noticing. And it’s maybe not a very important demographic, this very weird and specific kind of user: audience-obsessed, curious, newsy. Twitter’s earnings last quarter, after all, were an improvement on the period before, and it added 14 million new users for a total of 255 million. The thing is: Its users are less active than they once were. Twitter says these changes reflect a more streamlined experience, but we have a different theory: Twitter is entering its twilight.

The publishing platform that carried us into the mobile Internet age is receding. Its influence on publishing will remain, but the platform’s place in Internet culture is changing in a way that feels irreversible and echoes the tradition of AIM and pre-2005 blogging. A lot of this argument comes down to what we feel. Communities can’t be fully measured by how many people are in them. So as we suss out cultural changes, relying on first-hand experience is a first step.

Twitter used to be a sort of surrogate newsroom/barroom where you could organize around ideas with people whose opinions you wanted to assess. Maybe you wouldn’t agree with everybody, but that was part of the fun. But at some point Twitter narratives started to look the same. The crowd became predictable, and not in a good way. Too much of Twitter was cruel and petty and fake. Everything we know from experience about social publishing platforms—about any publishing platforms—is that they change. And it can be hard to track the interplay between design changes and behavioral ones. In other words, did Twitter change Twitter, or did we?