In lieu of an abstract, here is a brief excerpt of the content:

Reviewed by:
  • What Algorithms Want: Imagination in the Age of Computing by Ed Finn
  • David Nofre (bio)
What Algorithms Want: Imagination in the Age of Computing. By Ed Finn. Cambridge, MA: MIT Press, 2017. Pp. 272. Hardcover $29.95.

In recent years the term "algorithm" has become almost a synonym for the complex mixture of computational techniques and computer programs that support many of the online transactions that shape our lives. It is in this new guise, as Ed Finn reminds us in What Algorithms Want, that algorithms have acquired a quasi-mythical status, as some sort of sorcerers that [End Page 811] are able to reveal our most intimate selves, the inner workings of our societies, and even the deepest structures of the universe. But we should not be naive and think of algorithms as just numerical recipes for solving problems. For, as Finn argues in this book, algorithms are better understood as "cultural machines," that is "complex assemblages of abstractions, processes, and people" (p. 2) which inevitably involve all sorts of social, cultural, and economic assumptions.

Finn's What Algorithms Want is a valuable addition to the growing studies on the politics of algorithms. At the crossroads of the burgeoning fields of platform studies, critical code studies, and media studies, the book provides an insightful reading of some of the social and cultural inferences and filters embedded in today's most relevant social media platforms and web-based services, such as Facebook, Google, Netflix, and Uber. Chapters two through five form the core of the book. Each focuses (mostly) on one of these platforms and/or services. Chapter two deals with Apple's intelligent assistant Siri and Google. In this chapter, Finn shows how both services actually depend on a deep well of data that we provide to fulfill their Utopian promises: in the case of Siri, the appearance of intimacy and simulated humanity; in the case of Google, the promise to deliver absolute encyclopedic knowledge. Chapter three looks at Netflix, the film-on-demand web-service. Here Finn convincingly argues that Netflix, by offering content assembled especially for each one of us, embodies the ideal of personalization common to our neoliberal times. Chapter four shifts our attention to the process of gamification of social media platforms. In this chapter, Finn analyzes the aesthetics of Uber's interface and shows how its game-like features seek to make more palatable the experience of traveling with strangers. Finally, chapter five explores the libertarian ideology of mistrust that underlies the development of cryptocurrencies like Bitcoin.

These core chapters follow an historical overview in chapter one, which traces our current notion of algorithm back to the confluence in the 1940s and 1950s of the fields of formal logic, cybernetics, information theory, and computer programming. This chapter would have benefitted from a greater engagement with the work of historians of computing and historians of mathematics and logic. The historical discussion is at times too sketchy and includes some historically inaccurate conclusions: e.g. Finn states that McCulloch and Pitts proved that their model for biological neurons was computationally equivalent to a Turing machine (p. 28), a proof that they did not provide in their classic article of 1943.

More worrying in my view is the loose way in which Finn deals with the Church-Turing thesis and the notion of effective computability. These results from formal logic, which were formulated in the 1930s and 1940s, have come to play a fundamental role in the field of theoretical computer science. Yet their content simply does not account for the many abstractions [End Page 812] that go to make up our modern computational infrastructure, as Finn pretends (p. 25). The role of formal logic in shaping computer programming has a much more complex history that is not simply one of one-way influence.

These historical and conceptual inaccuracies are a significant shortcoming in this otherwise inspiring book. For What Algorithms Want offers us a penetrating account of the aesthetics of abstraction, simulated humanity, and personalization that characterize some of the most successful of today's social media platforms and web-based services. This aesthetics, as the book successfully shows, has been conceived and developed...

pdf

Share