A couple of weeks ago I got an email from Google saying that my Location History data would be erased soon since I haven’t used their tracking service for a while. In fact, I now think that the only reason why I stopped using that feature – which probably I never turned on myself but was a default one – was that I switched from an Android phone to an iPhone a couple of years ago, more or less at the time the tracking appears to have stopped. Google also pointed out that before “my data” was erased I could download a copy of it, which – pushed more by curiosity than by a sentiment of privacy – I in fact did. The so-called Takeout I got from them came as a zip file including a quite annoying and in my opinion somewhat useless “archive_browser.html” that is supposed to help one to navigate throughout the data. After one minute I abandoned it and I went directly to the Location History directory. There, a Semantic Location History folder caught my attention.
Organized by year and month, a number of json (JavaScript Object Notation) files offered a semanticized data landscape of my life covering almost a decade. A period of my life where, additionally, lots of things happened. Consequently, beyond curiosity, I have to admit that I suddenly found these records strange and fascinating. I dived into them and checked portions of the data that, in a way, made me revisit moments that were important for me. It was as if the notion of history – and of story for that matter – had been recast in variables, coordinates, and probabilities that, nonetheless, could take my memory to an exact day and place. The image above for example shows the moment when I arrived in Germany in February 2018 – I’ve been here since then.
But then, some elements in the json files appeared to me as a much more interesting matter. The data showed how the tracking system – Google’s machine – weighted the probabilities of where I was and what sort of transport I was using – see in the image, more or less in the middle, how the “locationConfidence” variable is statistically weighted with the 98.299385 value. Using the GPS in my phone and the constantly changing coordinates that device sent, measuring the motion speed between one coordinate state and the following, and contrasting that with the places where I was – the names of stores and institutions voluntarily provided by other humans to their system was key for that – this machine could established if I was walking, riding a bike, driving a car, or flying with “HIGH_CONFIDENCE.” All this became, then, particularly relevant and, say, revealing to me because it was totally reminiscent of questions of statistical forecasting I have been studying in the context of my research on cybernetics. In other words, navigating these all too contemporary objects of notation – Aufschreibesysteme with Kittler – this machine emerged as connected to that old epoch of systems of prediction and anticipation. Now with Wiener:
Behind all statistical work lies the theory of probabilities. The events which actually happen in a single instance are always referred to a collection of events which might have happened; and to different subcollections of such events, weights or probabilities are assigned, varying from zero or complete improbability (rather than certainty of not occurring), to unity or complete probability (rather than certainty of occurring). The strictly mathematical theory corresponding to this theory of probability is the theory of measure, particularly in the form given by Lebesgue. A statistical method, as for example a method of extrapolating a times series into the future, is judged by the probability with which it will yield an answer correct within certain bounds, or by the mean (taken with respect to probability) of some positive function or norm of the error contained in its answer.
Norbert Wiener, Extrapolation, Interpolation, and Smoothing of Stationary Time Series: With Engineering Applications (Cambridge: The MIT Press, 1964[1949]), 3-4.
Perhaps more importantly, navigating through my Semantic Location History folder I had the impression, a somehow vivid impression – what a beautiful and important word impression is, at least in media-theoretical terms – that we are in fact part of the machines that shape our lives; not only in relation to our present, and certainly our future, but regarding our past, our history, too. It was as if my phone, all those roads and cell antennas, airports and stores named and tagged by people, had then become part of that machine with me, and now – unknown to me as they continue being – are part of my history, along with all those data centers, servers, and data flows inscribed, as machinic impressions, in the json files of my life.
A couple of months ago Francis Hunger published a working paper arguing that Artificial Intelligence had to be unhyped, and that in order to do that we should instead – aiming to avoid the biases anthropological tropes bring with them – use terms such as “automated pattern recognition” or “weighted networks.” I agree. But I also agree with Simondon when he argued in favor of a collectivism formed by technical objects, natural non-organic beings, and organic (human) individuals. From the statistical forecasting and pattern recognition systems of the cybernetic epoch to their current inheritors – AI included – it seems to me that the present-future of our kind, just as our near past, has been, is and will be hybrid. Not entirely artificial, nor fully organic. Both and beyond. And thus – hopefully without and beyond big Techs – above all machinic.
The existence of the collective is necessary for information to be significative […] To receive an information is in fact for the subject to carry out within itself an individuation that creates the collective rapport with the being from which the signal arises. To discover the signification of the message that comes from one being or several beings is to form a collective with them and individuate through the group individuation with them.
Gilbert Simondon, Individuation in Light of Notions of Form and Information, Volume I (Minneapolis: University of Minnesota Press, 2020[1958]), 344.
To close, this brief note on cybernetics, daily telecommunications and being is certainly an informal Sunday reflection made from the perspective of German media theories and, in a way, media philosophy too. From there it is possible to state, I’m quite convinced about it, that despite the evidence that shows the increasingly insurmountable oppressive character of media technologies and machines, they still conceal a tremendous potential for emancipation. In that sense I dare to recall a recent tweet by Shane Denson. He referred there to Friedrich Kittler’s influential and somewhat polemic opening of Gramophone, Film, Typewriter. Denson wrote that “while ‘Media determine our situation’ was a great way to stir up controversy, Kittler’s opening could actually be translated: ‘Media orient us in situations that call for a briefing.’” I replied to that with the original German version: “Medien bestimmen unsere Lage, die (trotzdem oder deshalb) eine Beschreibung verdient.” Then Denson answered back in a challenging, yet revealing and eye-opening way: “Yes, the hidden compounds are the key. Lagebestimmung = ‘orienteering,’ and Lagebeschreibung = ‘briefing’ (with all the original military overtones).”
Maybe, once we dive into contemporary objects of notation with cybernetic trained eyes, we could understand the machines of our times as the “Media [that can] orient us in situations that call for a briefing.”