epistemic insubordination
Aral BalkanServer farm (n): a place where human beings are farmed for their data by corporations.
See also: people farming.
Data (n): information about a thing.
With enough data about a thing, and the right algorithms, we can begin to create digital approximations of the thing.
If that thing happens to be a human being, the resulting approximation (proxy/simulation/profile) is considered highly valuable. In fact, the business model of mainstream technology (people farming) is to own these profiles to both exploit them internally and to rent out selective access to third parties.
jon r@aral What you mention reads to me as the end product of a process that encodes meaning in information, and through that turns it into data.
From a constructivist, and also from a deconstructivist perspective, we may also add to this, that the actual 'thing' in question is actually created by this selection of certain information and assignment of meaning to its relation in the world.
Why, to 2nd your argument, the production of subjects as ad-sensitive consumers is needed to extract value.
felix stalder (@[email protected])
How to define data? Today it’s important to differentiate between at least two types of data. The first covers data in the conventional sense, going back at least to the beginnings of statistics in the 18th century (think national population counts) or, more broadly, to the beginnings of writing (think cuneiform tablets). Here, data are alphanumeric representations of physical objects or states, organized into categories, created with specific methods, technologies, and interests. You can further differentiate between measurement data (e.g. a person's height) and metadata (e.g. the person's name), but the boundaries between these categories are fluid. Like all forms of representation, such data has an inherent paradox because it is supposed to be the same yet different from what it represents. How this paradox is resolved in practice, which aspects of a thing/state are represented in data, and into which categories these representations are organized are inevitably political processes. There are no objective, direct procedures here through which things can speak for themselves, but only certain methods and techniques whose use and direction are guided by changing interests. In the context of machine learning, however, data means something different. Here, data are sets of texts and/or numbers that are analyzed for statistical patterns. This is no longer about a representative relationship to an external world but about purely internal relationships. Accordingly, it is not the methods and techniques of collection that turn something into data, but rather the form of analysis, evaluation, and manipulation of statistical patterns. Everything that is digitized becomes data the moment it is treated as a collection of statistical regularities and peculiarities.