For example, a user planning a trip may be very interested in news of other users who have traveled recently, in news of trips identified as events by other users, and in travel information, and then be much less interested in these relationships, events, objects, or categories or subcategories thereof upon his return. Thus, items of media content associated with another user who has traveled recently may receive a large weighting relative to other items of media, and the weighting will decay steeply so that the weighting is low by the time of the user’s return. (Zuckerberg et al. 2012: 5)
Afinity, weight, and time are as dynamic as the news feed itself. Constantly changed and fine-tuned, there is no clearcut way in which a certain state of the feed can easily be discerned. Like the prospect of moving furniture at the Facebook headquarters, the news feed ofers users a slightly diferently designed space to hang out in every time they log in. While it is unlikely that the sofas at the headquarters are really moved that often or to radically diferent areas of the building, the prospect of continuous change is what is important. The culture of not knowing how the ofice may look when entering the building each day or the uncertainties Facebook users face when logging in are not that diferent. The main diference is that users may never really know what they are exposed to, even after entering the digital building that is Facebook. A moving sofa is easily detectable; an A/B test is not. Users will never know exactly what they are witness to or how the world appearing on their screens came into existence. Living in an algorithmically mediated environment means learning to live in the moment and not to expect to step into the same river twice.16 It also means acknowledging that talking about EdgeRank, the news feed algorithm, or whatever one now wants to call it, means talking about a general underlying logic, not its exact mathematic formula.17
Life at the Top
81