The Charisma Machine: The Life, Death, and Legacy of One Laptop per Child (Infrastructures)

by Morgan G. Ames

A charismatic technology may promise to transform its users’ sociotechnical existence for the better, but this book will show that charisma is, at heart, ideologically conservative. Charismatic leaders confirm and amplify their audiences’ existing ideologies to cultivate their appeal, even as they may paint visions of a better world.50 A charismatic technology’s appeal likewise confirms the value of existing stereotypes, institutions, and power relations. This unchallenging familiarity is what makes a charismatic technology alluring: even as it promises certain benefits, it simultaneously confirms that the ideological worldview of its audience is already right—the charismatic technology will simply amplify it.

Technologists such as those involved with OLPC ignore the origins of charisma at their (and all of our) peril—at the risk of always being bewitched by the next best thing, with little concept of the larger cultural context that technology operates within and little realistic hope for long-term change.

Larry Cuban and David Tyack have shown that the same is true for educational technologies.56 This book provides a vivid example of what happens when a charismatic technology that was built around particular ideals comes up against the lived realities of its intended beneficiaries—beneficiaries who do not necessarily share those ideals.

Christo Sims’s description of “disruptive fixation,” where educational reformers become “fixated” on the promises of technology-driven disruption.

Through a detailed case study of the consequences of charisma—particularly the performativity of such projects, the treacherous allure of nostalgic design, and the catch-22 that short-term funding and performance expectations create—this book offers those interested in technology-heavy educational reform or development projects a way to recognize charisma, to either harness or resist it. Likewise, it offers technologists a means to critically assess the utopian promises that circulate not just in reform projects such as OLPC but across Silicon Valley more generally.

This book’s narrative begins historically. It first examines the intellectual and cultural history of OLPC, which reaches back over fifty years to MIT’s nascent hacker culture in the 1960s. It demonstrates the similarities between this hacker culture and the constructionism learning theory that forms a cornerstone of OLPC, showing how both rely on particular social imaginaries about childhood, learning, technology, and schools.

We examine the work that these imaginaries do for projects such as OLPC by evoking deeply held feelings about the value of (certain types of) play, creativity, and learning and the kind of child capable of them. In particular, OLPC implicitly invokes the social imaginary of the technically precocious boy, which developed as an alternative to dominant notions of masculinity in the United States.

they used the XO laptop’s design as a vehicle to evangelize this imaginary to their beneficiaries—specifically, the superiority of a childhood that is full of pedagogically playful experiences involving technology that would, in particular, be expected to engage clever, scientifically inclined, and often rebellious boys. Stories of finding escape in computer worlds and of teaching oneself to program in spite of clueless adults played an important role in the project, even as those stories discounted the social and infrastructural resources that enabled this exploration and ultimately perpetuated classed, racial, linguistic, and gendered tropes about what computing culture is and who belongs in it.

These chapters draw on data collected during seven months of ethnographic fieldwork in Paraguay: six months from mid-June to mid-December 2010 and an additional month of follow-up research in November 2013. This fieldwork was conducted in collaboration with Paraguay Educa, a nonprofit, nongovernmental organization that ran the local project. During this time, I observed children, their teachers, and their families in classrooms, their homes, and public spaces. I typically spent two to three days per week in schools, for a total of fifty-seven school days (or approximately 450 hours), observing classrooms and schoolyards. I conducted 144 interviews with children, parents, teachers, administrators, Paraguay Educa employees, and others involved in the project, nearly all in Spanish.60 I supplemented this ethnography with some quantitative data, which I use to triangulate results in my analysis.

Even though the laptops were built to be both rugged and repairable, in August 2010 about 15% of students had unusably broken laptops, and more had laptops with missing keys or broken pixels that made them difficult to use.

the social imaginaries that motivated the design and popularization of the XO across the technology world did not translate to most users: fully two-thirds of children hardly ever used their laptops. Some nonuse was due to breakage, which occurred along gendered and socioeconomic lines, complicating some of the benefits the project was supposed to provide.

what about those who were using their laptops? Chapter 4 also discusses how most of these children were not interested in learning to program, as OLPC’s learning theory had predicted. Instead, they used their machines as portals to music, games, and videos on the internet—even though the design of the laptop made these uses difficult.

Still, there was a very small subset of students—generously, about forty out of the four thousand who had laptops in 2010—who explored the constructionist and creative options of their laptops at least some of the time.

I found that each student had a constellation of resources that encouraged them down this path: families that steered them toward creative and critical thinking, a focus on the importance of education, and in many cases another computer in the home. Although children’s own interests were of course at play as well, their motivation for doing these activities appeared to be less a product of interactions between the universalized child and laptop than a negotiation between many factors, especially parents and peers.

Finally, Seymour Papert, who in 2005 was professor emeritus at the MIT Media Lab, had originated the very idea of “one laptop per child” and had discussed how this idea was meant to transform the world.

Negroponte readily admitted throughout his marathon of talks in 2005 and 2006 that the very idea for the project was actually Papert’s.

Papert, by contrast, was seventy-six years old and retired when OLPC was first announced in 2005, though he did do some early interviews for the project.4 This participation was cut short on December 7, 2007—around the time that OLPC’s first laptops were coming off Quanta’s assembly lines—when Papert was hit by an auto-rickshaw in Vietnam and suffered a brain injury from which he never fully recovered.

The OLPC community knew that Papert had written a best-selling book about the idea of one computer per child some thirty-five years before OLPC was announced. Many of them, especially Media Lab affiliates, had been captivated by this book and by previous charismatic projects built by Papert and his students.

Many of Papert’s projects—spanning some forty-five years of work at MIT, from the 1960s to the late 2000s—express aspects of this learning framework. Though he first began fleshing out details of constructionism in the 1970s in a National Science Foundation grant and a series of working memos, his book Mindstorms: Children, Computers, and Powerful Ideas (1980) brought constructionism to a wider audience and remains the core reference on the subject, with a smattering of other books, articles, working memos, and talk transcripts fleshing out additional details.

Papert’s constructionism borrows heavily from Jean Piaget’s theory of constructivism, as reflected in the confusing similarity between the two names. Before joining MIT in 1964, Papert spent five years at Piaget’s International Centre for Genetic Epistemology in Geneva, around the time that Piaget’s constructivist theories of child cognitive development—first articulated in the 1920s—started to gain popularity worldwide.7 Papert adopted from Piaget a focus on children’s learning as an active process of constructing knowledge about the world. Both stressed that children (and adults) learn by relating new concepts to what they already know and sometimes adjusting their theories of the world to accept new concepts. Piaget called this “assimilation” and “accommodation,” respectively, though Papert called both “Piagetian learning.”

Piaget is often still read alongside John Dewey, Lev Vygotsky, and others in schools of education, and the educational research community has amassed some evidence of the value in Piaget’s constructivist theories, though with some modifications and critiques along the way. Papert’s constructionism, on the other hand, is less commonly taught in education schools, though it remains popular among technology designers.

Papert asserts that when given unlimited access to a computer, children can also be empowered to think “like a computer,” learning the “language” of the machine just as living in another country when young would enable them to learn the language spoken there with a proficiency that adult learners can only envy. Under these conditions, Papert says, computers could finally scale up the ideals of “progressive education.”30

Thus, one of the central tenets of Papert’s constructionism—that children should have unrestricted access to computers—originated not with Piaget but with MIT’s hacker culture.

Another part of constructionism that differs from Piaget’s constructivism—and bears a clear relationship to hacker practices—is its focus on debugging. In constructionism, this refers to the process of embracing “wrong” ideas and then iteratively revising them by testing.36 In computer programming, it refers to the process of finding and correcting the mistakes (or bugs) in computer code.37 Papert argues that an openness to debugging is among the most important aspects of constructionism: it encourages children not to internalize feelings of failure when they have a “wrong” answer but to see this as an integral part of the learning process.38

childhood that is characterized by seemingly innate creativity, fearlessness (whether with physical feats or technology), innocent mischief, and “rugged individualism”—one that is in close communion with the natural world but utterly out of sync with the adult world.39 This imaginary seems so natural to many, especially in the United States, that it is often assumed to be timeless and universal; it has been a constant refrain in English and American media for more than 150 years, evinced by the perennial resonance and many variations of characters such as Huckleberry Finn and Peter Pan.

in particular of the worldview of the Romantic era, a set of nineteenth-century ideological shifts in the United States and western Europe that coincided with the Industrial Revolution.

The imaginary of the naturally creative child de-emphasizes the extent to which creativity and play are socially motivated and socially learned. In particular, it ignores the important role that parents and other adults play in structuring a child’s world—by providing toys, environments, activities, role models, and more—to elicit certain kinds of play and creativity.

This shared ethos between constructionism and hacker culture at MIT and beyond—a belief in the universal appeal of computers, in the natural creativity of children, in the detrimental effects of school, and in the value of certain kinds of rebellion—are all aspects of a particular social imaginary of childhood that is widely recognized across American culture: that of the technically precocious boy.

Further, every child has something to contribute; we need a free and open framework that supports and encourages the very basic human need to express. Give me a free and open environment and I will learn and teach with joy.”

Three key themes present in these principles—play, connectivity, and freedom—constitute the key pillars of OLPC’s charisma, and all were built into its laptop.

Indeed, the laptop’s hardware was otherwise puzzling. The first-generation laptop’s 433-megahertz processor was nearly ten times slower than state-of-the-art machines in the mid-2000s; it was typical of entry-level machines in 1999.11 It had a one-gigabyte storage capacity on solid-state Flash memory, which employees said would be faster and lacked the jammable moving parts of the then-common hard disk drive. However, this storage capacity was miniscule for the burgeoning multimedia-rich computer landscape: whereas one gigabyte was standard for a computer from the mid-1990s, computers in the mid-2000s typically came with fifty- to one-hundred-gigabyte hard drives. Likewise, its 128 megabytes of RAM (random-access memory, a computer’s working memory) was one-quarter of the 512 megabytes typical of entry-level machines in 2005. Slow as they were, each of these specifications was reportedly at least double what the team had first prototyped.12

This chapter explores how this justification for these hardware choices is an example of nostalgic design: it targeted the social imaginaries that resonated best with OLPC contributors’ own identities, especially the social imaginary of the technically precocious boy who found joy in understanding the machine deeply.

However, nostalgic design violates some of the core principles of user-centered design: to design for one’s users, not for oneself, and to account for the messy realities of use.

We will see that play is reflected in the toy-like appearance of the laptop as well as in the proliferation of games, game engines, and game-oriented constructionist software on the device—a proliferation that reflects the importance of video games in the childhoods of many of OLPC’s developers. Connectivity was built into the XO via its “mesh network,” which was designed to enable easy connections to other laptops and the internet and taken as a prerequisite for its Bitfrost security system—and which reflected the kinds of computer-mediated connections that were most important to many on the OLPC team. Encoding freedom, the XO laptop’s software was all free and open source, and the XO keyboard had a View Source button to emphasize this.

OLPC seemed to implicitly expect that children using the XO laptop would adopt their definition of play as one that involved video games and competition and one that ultimately led to programming—a starting point to learning deeply about computers and adopting the values of the programming community. In this way, they projected their own values onto the children they hoped to reach. They assumed that these children would find the same kinds of play captivating, and that this play would lead to the same positive outcomes in these children’s lives as it appeared to have in their own.

one of the project’s most hotly anticipated features (perhaps after its nonexistent hand crank): its mesh network. Though the mesh network did not work in practice and was in fact dropped in an update to OLPC’s Sugar software during my fieldwork in 2010, it generated much excitement in the programming community and technical press when it was announced. It would let XO laptops connect to one another without the need for an access point to moderate the connection, allowing them to circumvent any surveillance, censorship, or other attempts at control. It would let a group of laptops share one internet connection that might be on one edge of town (never mind how abysmally slow that connection might become). Even when it was off, a laptop could act as a relay in the network (never mind the battery-draining implications of this).

This focus on connection was also a result of nostalgic design, reflecting developers’ own memories of positive experiences online in their youth. Online spaces such as bulletin board systems (BBSs), chat rooms, Internet Relay Chat (IRC) channels, blogs, multiplayer online games, and social media are not places of harassment and trolls, they argued; they are places where computer-loving iconoclasts, technically precocious boys like many of them had been, could really be themselves, find others like them, and broaden their horizons.

Like the old hardware specification described above, this illustrates another clear instance of nostalgic design: it assumed a world with easy-to-obtain peripherals that were compatible with one another, such as the self-designed and easily repairable desktop “tower” computers that were especially prevalent in the 1990s.

these narratives of being self-taught invariably smoothed over the social support and scaffolding that, though unacknowledged, had generally helped these hackers. These might have included a stable home environment and a resource-rich, infrastructurally stable community that supported technically precocious children such as themselves, and also often included a parent who was a computer programmer or engineer. OLPC’s idea of the self-taught learner who disdains school for computers thus discounted the critical role that various institutions—peers, families, schools, communities, and more—play in shaping a child’s educational motivation, technological practices, and identities.

In his descriptions of “yearners” and “schoolers,” Papert entirely left out many potential users—from teachers to children with diverse interests or those lacking the requisite technical expertise—and a range of potential uses. Instead, he—and OLPC—designed the XO to support technical interests and ascribed those interests as innate.