What Algorithms Want

What Algorithms Want: Imagination in the Age of Computing, MIT Press, Spring 2017.


The apotheosis of the algorithm is here. In the past several years we’ve hit a turning point, leaving endless debates about artificial intelligence behind in favor of tacitly accepting complex computational systems that tell us where to go, who to date and what to think about (to name just a few examples). The mythos of computation has become almost universal: with every click, every terms of service agreement, we buy into the idea that big data, ubiquitous sensors and various forms of machine learning can model and beneficially regulate all kinds of complex systems, from picking songs to predicting crime. Already these culture machines dominate the stock market, compose music, drive cars, write news articles, and author long mathematical proofs—and their powers of creative authorship are just beginning to take shape. This book proposes that we are missing the algorithmic sea change by focusing only on the crests of waves—we continue studying books, films and games when we should be paying much closer attention to search bars, mobile applications, text prediction systems and other rapidly evolving tools for thinking and authoring. Scholars and cultural critics assume algorithms are all about code. They’re actually about culture.

What Algorithms Want takes on the challenge by reading contemporary algorithms in the context of a long cultural history. The figure of the algorithm, which computer scientists use as convenient shorthand for “a method for solving a problem,” is a mythic concept much older than the invention of the computer, with deep roots in the Enlightenment and the philosophical tradition of rationalism. I excavate this historical narrative through a genealogy of the algorithm as a figure in contemporary culture, tracing its origins in cybernetics, symbolic logic and language philosophy. These foundations inform interpretive readings of a variety of algorithmically entangled cultural works: Apple’s Siri, Netflix’s House of Cards, Ian Bogost’s Cow Clicker and the cryptocurrency Bitcoin, among other objects of analysis. Though seemingly very different from each other, all of these works are algorithmic forms that have been authored by complex computational systems in collaboration with (often unwitting) humans. We work with and think through these culture machines, re-enforcing and reinventing the mythos of the algorithm as we go. I develop a method I call “algorithmic reading” to offer original interpretations of these new modes of hybrid authorship, which involve millions of computer processes and human beings thinking, creating and enacting culture together. Algorithmic reading is reading by the lights and shadows of machines: the brilliant illumination of computationally enhanced cognition and the obfuscations of black boxes. As all culture comes increasingly under the sway of the algorithm, I argue that algorithmic reading will be a vital method for the humanities in the 21st century.

Beyond helping us develop a new reading method, these cultural works teach us something important about the nature of algorithms themselves: namely, that algorithms can never be separated from the conditions of their implementation. Not only are algorithms cultural all the way down, they are systems for belief as much as they are rational tools—the latest incarnations of a tradition that encompasses Liebnitz’s quasi-spiritual mathesis universalis, medieval religious automatons and contemporary representations of god-like artificial intelligence. Coming to terms with this deep culture structure ultimately reveals the interpreter to be herself complexly enmeshed within algorithmic culture machines, from search engines and word processors to the social media platforms on which she shares her work.

The stakes of this conversation are high as algorithmic thinking reorders entire industries, cultures and creative traditions. Even the engineers behind some of the most successful and ubiquitous algorithmic systems in the world—executives at Google and Netflix, for example—admit that they only understand some of the behaviors these systems exhibit. But their rhetoric is transcendent and emancipatory, equating code, bandwidth and freedom. Our standard assumptions about algorithms are historically and critically shallow, and at best we comprehend them through layers of abstraction and analogy. To understand this sea change, we need to read and experiment with algorithms as they are: cultural machines of oceanic depth and complexity.