unit reflection

dddddd123

ConvergingTechnologies.pdf

Home >English homework help >unit reflection

CHAPTER 2

CONVERGING TECHNOLOGIES

Technology is the active human interface with the

material world.

— Ursula K. Le Guin

TECHNOLOGY AND SOCIETY

By starting with a chapter on technologies, I do not mean to imply anything about causality in the relationship of machines to society. There are those who, in the interest of explanation or simply as a shortcut, argue for technological determinism. The plow gave us feudalism; the steam engine brought a manufacturing society; the computer, an informa- tion age. Others go to the opposite extreme and believe that technologies are determined or shaped by society. In essence, machines are no more than congealed social relationships. Take apart a plow, a steam engine or a computer, they say, and we are left with the dominant social forms that existed at their creation. Accordingly, emerging feudal, industrial, and informational societies create the technologies appropriate to the social relations and structures of power of the time. Both

Mosco, V. (2017). Becoming digital : Toward a post-internet society. ProQuest Ebook Central <a onclick=window.open('http://ebookcentral.proquest.com','_blank') href='http://ebookcentral.proquest.com' target='_blank' style='cursor: pointer;'>http://ebookcentral.proquest.com</a> Created from purdue on 2021-03-27 07:52:29.

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

of these views are useful but I believe that it is more useful to take a position somewhere in between the two. Formally put, technologies, societies, and individuals mutually constitute one another. To put it less formally, they contribute to each others’ creation and development at every stage, setting important limits without any essential causal determination.

TO THE CLOUD

I begin with technologies because readers will be less familiar with the major elements that make up the Next Internet and because technological convergence represents a fundamental change from the network most people have used for the last three decades. Cloud Computing preceded data analytics and the Internet of Things and got its name from the network dia- grams that engineers used to describe a network. Typically, network nodes were drawn to look like clouds. As anyone who has visited a massive, server-filled, windowless data center can attest, there is nothing cloud-like about its look and feel. In its present form, the Cloud has been available to individual and institutional users for slightly over a decade. One reason why some see today’s Cloud as little more than a marketing ploy is because a form of what we now call Cloud Computing was prominent in the early years of computeriza- tion. Then it was referred to as time-sharing to identify the practice of computing on terminals linked to a central computer where the data and intelligence were stored. As a graduate student, I fed punch cards into a mainframe input device in order to have my research processed and analyzed. There were no personal computers back in the 1960s and early 1970s. We were all tethered to what today would be called a data center, albeit considerably more primitive than the vast information factories that fill acres of land today.

16 Becoming Digital: Toward a Post-Internet Society

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

This began to end in 1975 when the first personal computers

appeared and people slowly adapted to devices that stored

programs and data on their own computers and disks. In this

respect and on a far grander scale, Cloud Computing takes

us back to the future, where the electronic tether now con-

nects us to data centers. Cloud Computing does repackage an

old idea but one might also say that the package looks so

very different today that, marketing claims aside, it does

merit a new identity. This was signaled in a New Yorker

magazine cartoon that, like the one announcing the arrival of

the Internet, lifted the curtain on Cloud Computing in popu-

lar discourse. It featured a young boy looking sheepishly at

his teacher who expects him to deliver an assignment.

Instead, he delivers a new riff on an old excuse: “The Cloud

ate my homework.” Cloud Computing had arrived. The Cloud is one of the few information technologies that

actually comes with an official definition. Shortly after the

marketing departments of Cloud companies began to spread

publicity and pump up the hype, U.S. government technology

officials ordered agencies to adopt Cloud Computing.

Puzzled, many responded with, “What’s Cloud Computing?”

As a result, the government agency responsible for expertise

in science and technology was asked to produce a definition

and came up with this: “Cloud computing is a model for

enabling convenient, on-demand network access to a shared

pool of configurable computing resources (e.g., networks, ser-

vers, storage, applications, and services) that can be rapidly

provisioned and released with minimal management effort or

service provider interaction.”1 I cannot say with certainty

that this completely satisfied the demand for clarity. Other

analysts offered more pointed definitions, including “comput-

ing on someone else’s computer,” “the next step in outsour-

cing IT jobs,” and “a bullsh*t marketing term,” a definition

17Converging Technologies

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

I actually heard given by an IT expert at a cloud computing

promotional event. Each of these definitions is accurate. Cloud computing is a

system that moves data stored on individual computers and

in the IT departments of institutions to large, distant data

centers operated by companies that charge for storage and

use. Amazon is the largest of such firms. For over a decade its

Amazon Web Services division has benefited from the ability

to continuously drop prices through cross-subsidies from its

many other divisions, to take market share (now at about

one-third of the cloud market) and build a global network of

data centers. In addition to warehousing data for a fee, cloud

providers offer software and applications that enable users to

carry out operations without having to store and update soft-

ware on their own machines. In addition to offering storage,

Google, for example, provides its Gmail service through the

Cloud. Similarly, Microsoft, which is second largest in cloud

storage, now delivers its popular writing, spreadsheet, and

other productivity software through the Cloud on a monthly

subscription basis. Apple provides data storage through its

iCloud along with such additions as device syncing and

enhancements for those who store their mail, music, and

video files in the Cloud. To round out the familiar examples,

Facebook has created a virtual community in the Cloud for

its two billion users. Mastery of the Cloud is one of the pri-

mary reasons why Amazon, Google, Microsoft, Apple, and

Facebook are the most valuable companies in the world.

Some cloud businesses also offer services that make it possi-

ble for companies to outsource some or all of their sales, per-

sonnel, accounting, legal, and other functions. The company

Salesforce.com is a prime example of a cloud-based business

that draws from its network of data centers to manage mar-

keting and human relations for firms of every size.

18 Becoming Digital: Toward a Post-Internet Society

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

http://Salesforce.com

For the convenience of offloading data, software and

services, individual and institutional users pay a regular sub-

scription fee that can also vary with usage. This has proven

incentive enough to create a global construction boom in

data centers as large firms and some specialized medium-sized

companies jockey for a share of the growing market. In addi-

tion to subscriber fees, companies have access to vast stores

of valuable data, particularly when it can be grist for the big

data mill. For example, Google and Facebook use the infor-

mation that flows through their data centers on the email and

social media posts of users to sell customized advertising.

Moreover, data is used to develop algorithms that businesses

and governments are willing to pay for because it helps them

to predict the activities of markets and people. On the other

side, individual users benefit by lessening the need to store

data and software on their own computers. As the digital

world has become less safe for users, many would rather put

their trust in established commercial providers to protect their

data, or at least use more secure systems for backup, than to

rely solely on their own devices. The same holds, only on a

much larger scale, for businesses and other large institutional

users. These benefit by reducing the need to expand their

own IT facilities. Who needs an onsite data or server center

when another company will provide storage for your com-

pany or government agency? Why bother buying new soft-

ware that your IT staff has to regularly update and debug

when a cloud company like Microsoft will provide the

latest versions online? More radically, why employ a large

IT staff, as well as accountants, sales, personnel, legal, and

other costly staff professionals when a cloud company like

Salesforce will do the work for you? Cloud Computing

begins to make financial sense when it can be used to cut

professional staff and outsource to the Cloud.

19Converging Technologies

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

Seeing opportunities to cut staff also inspired governments to hop onto the cloud bandwagon early and to spend lav- ishly. Facing budget cuts, many domestic agencies turned to the Cloud to save on IT costs. Intelligence and defense agen- cies, with far fewer budget worries, saw the Cloud as an opportunity to centralize operations for increased security and control. The National Security Agency, which became famous, or infamous, because of whistle blower Edward Snowden’s revelations about the extent of its legal and illegal spy operations, has built one of the world’s largest Cloud facilities in a secure Utah location. The Central Intelligence Agency also moved quickly to construct its own facilities and it also gave a contract to Amazon to purchase $600 million worth of cloud services. It is not unusual for government agencies, including those operating in a highly sensitive envi- ronment, to outsource at least some of its cloud requirements in order to build redundancy into its systems and create strong ties to key private sector organizations. The ability to call on its commercial friends comes in handy when the CIA faces intensely embarrassing situations, like the Wikileaks revelation that it routinely hacks ordinary consumer devices like Samsung’s “smart” televisions, recording the conversa- tions of households even when the sets are turned off, and analyzing them in the CIA Cloud.

These significant incentives are offset by technical chal- lenges. No minor investment, data centers are big projects that require careful planning if the interconnected system comprising tens of thousands of servers, switching equip- ment, and telecommunications facilities linking the facility to the world are to function properly. In addition to requiring real estate, cloud data centers need massive energy supplies to power servers and to keep them sufficiently cool to avoid overheating. Most people in the business are aware of the overheating incident in which, were it not for a nearby shop

20 Becoming Digital: Toward a Post-Internet Society

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

that sold electrical fans to panicked employees, Facebook’s first server center might have completely overheated and taken down the company. Finding low cost energy deals is not easy nor is locating inexpensive supplies of the water essential to keep servers cool. As a result, cloud facilities are often located in cold climates with abundant energy and water supplies, a reason why Canada’s province of Quebec, which has both, and especially Montreal, with its burgeoning tech industry, has become a location of choice for big compa- nies such as Google and Amazon. On the down side, how- ever, remote locations often mean high telecommunications costs to connect data centers to the global grid. Furthermore, given the sensitivity of both servers and data, cloud facilities cannot be located in places where earthquakes and weather disasters would spell catastrophe. In addition, customers demanding quality 24/7 service do not tolerate downtime. The only solution is the installation of expensive backup sys- tems that can take over when the electricity grid goes down. While a great deal of understandable concern, thought, and investment goes into the need for security against hackers and intruders, climate change may prove to be an equally for- midable challenge.

Users face significant challenges as well. Parting with data to save on storage costs might make economic sense, but it also means giving up the feeling of control that proximity often provides. Can you really trust a cloud provider to care for your data as much as you do? What happens to your data if a provider goes out of business or just decides to focus on something else, the next big thing that suddenly looks like a lucrative opportunity? It is in the interest of cloud providers to lock in customers with long-term contracts that contain expensive additions. What sounds good today often proves excessively costly tomorrow. Moreover, customers have to trust their providers to offer sufficient security. While they

21Converging Technologies

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

have many incentives to maintain a clean Cloud, those offering cloud services have to make delicate judgments that balance this requirement against the growing costs of adequate security.

It is also advantageous for providers to build data centers in numerous locations around the world because doing so creates new markets, advances global connectivity, and diminishes dependence on specific markets. But it also means subjecting data to different legal jurisdictions. For years, authorities in the United States sought access to files stored in a Microsoft data center located in Ireland which U.S. law- enforcement authorities maintained was central to a major criminal investigation. Microsoft and the government of Ireland resisted U.S government demands, citing EU jurisdic- tion and regulations protecting the data. The U.S. govern- ment lost that case but filed another one, this time against Google, to obtain emails stored on foreign servers. Apple, Amazon, Microsoft, and Cisco filed briefs to support Google and the outcome is uncertain. What is certain is that, as the Cloud becomes an essential feature of the Next Internet, situations like these will become more pervasive. Canadians worry about their data stored on U.S. servers, which subjects data to the Patriot Act and other legislation that justifies widespread surveillance of information practices. What will they and others think when their data is stored on servers in China or Russia?

The cloud computing industry has dealt with these issues in numerous ways. From the start it created tiers of services and levels of control that have provided some customization for users who demanded it. It also distinguished those who use and trust cloud services only for data storage from those who want, and trust the cloud company to provide a plat- form for functions like email (Gmail), and office software (Microsoft). Customers not keen to rely on Google for email

22 Becoming Digital: Toward a Post-Internet Society

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

can download their own software and store their emails on their own computers or in IT department servers. Again, this is no simple matter. Google might take some of the work out of email and for a price that only amounts to having ads appear alongside mail. But it also means storing mail in data centers whose location, and hence legal jurisdiction, is not clear. In addition, some cloud providers serve as robust alternatives to in-house functions replacing all or part of entire departments by offering traditional business services. Whether it is Salesforce.com or Uptime Legal Systems, this form of Cloud Computing, referred to as Software as a Service (SaaS), means giving up the greatest amount of con- trol and requires the most trust. It is one thing to store your data on someone else’s computer, quite another to outsource core functions like sales, legal, and personnel to another firm.

Alongside these service levels, cloud providers distinguish between public and private services. The Public Cloud is for those with the least need or concern for security and privacy. Although cloud companies insist that they provide the highest level of data protection and secure services for everyone, stor- ing data in the Public Cloud is like bringing it to a warehouse that is also used by tens of thousands of other customers. A public cloud data center takes on all customers and attracts them with the lowest rates in the industry. Amazon pioneered the Public Cloud and by taking advantage of economies of scale continues to lead the sector. Those who want and can pay for more security turn to the Private Cloud. These are the equivalent of gated communities in the Cloud where security is tighter and servers are sequestered for higher paying custo- mers. The distinction between public and private Clouds varies by company and there are different grades of service within each. Industry associations help develop standards with the financial services sector understandably taking a lead. As a result, standards have been adopted for “bank-grade” Cloud

23Converging Technologies

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

http://Salesforce.com

systems, more secure than those typically found in a private cloud service. There are also various hybrid cloud systems that bring together features of both public and private clouds for those who want more security but cannot afford a fully private cloud provider. These often also include some data storage and processing on the customer’s premises and access to some company data in the event of an outage or a hacking attack on one or more of the cloud provider’s facilities.

Before leaving this discussion of public and private cloud services, it is important to consider just what the term “public” means here, if only to specify what adding an “extra edge of consciousness” actually means. Unlike the way it is used in discussions of public space, citizenship, and democ- racy, the public in Public Cloud refers to a customer’s choice to buy services from a commercial provider, typically through a subscription or usage fee but also by accepting advertising. This is a far cry from the traditional view of a public as a set of citizens whose right to participate in a democratic society requires the equitable distribution of the tools for them to do so. A genuine Public Cloud would set aside access to secure cloud services as a fundamental right and a requirement for democracy. In this respect there is no Public Cloud. There are varieties of private Cloud Computing that come with a range of price tags and security levels. The erosion of the concept of the public has important consequences, particularly as socie- ties move into the next phase of the Internet. If planning for the future of one of democracy’s central tools is bur- dened with historical amnesia over a central principle, it becomes even more difficult to consider the future informa- tion space as anything other than private, commercial, and commodified.

Recent years have brought into the mix what is called Edge Computing, a new form of outsourced data processing so named because it pushes processing and intelligence out

24 Becoming Digital: Toward a Post-Internet Society

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

from the data center to the edges of networks. It warrants close scrutiny because it may challenge the dominance of Cloud Computing. Edge Computing emerged from a debate about whether cloud data centers are capable of managing the complex requirements of new devices that require near instantaneous data analysis and tight security. The automated car, what has been described as a data center on wheels, drones, industrial robots, embedded medical devices, all of these test the capacity of large, centralized data centers to operate with efficiency and timeliness. As a result, companies are exploring the potential to embed some of the intelligence contained in the traditional cloud inside the devices them- selves, thereby speeding up automated decision-making (e.g., sensors warning of a lane-hopping car up ahead) and permit- ting customized security. This would depart from, for exam- ple, how smart phones operate. When I go out for a walk and my data craving self (more about that in chapter four) checks to see if I have made my daily goal of 15,000 steps, I receive the results from a communication between my phone and a cloud data center which stores my iPhone infor- mation in the iCloud. Edge Computing contains the Cloud in the device.

Some believe that Edge Computing will replace the Cloud by reopening the potential for distributed networks of mil- lions, if not billions, of devices. Others foresee a division of labor with the distant cloud data center providing strategic decision-making, for example, advising and carrying out soft- ware and firmware updates, and leaving detailed operational decision-making to the edge-equipped device. Over the past 40 years I have observed hundreds of forecasts announcing the Next Big Thing that will rock the industry and society. Back in the 1960s, for example, color televisions were expected to soon give way to 3D. In the 1970s videophones were just around the corner. In the 1980s it was the

25Converging Technologies

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

interactive device called videotex. Few make it beyond the dream lists of prognosticators because the existing dominant technology adapts and absorbs the challenger, because the new device or system arrives too soon, or because the pro- blems it solves are not worth the investment. Nevertheless, Edge Computing bears watching if only to remind us that cloud data centers may not be up to the task of automation’s next wave. Such an outcome would have significant impacts on America’s dominant tech firms because they all have invested heavily in the Cloud.

Cloud Computing has been able to fend off challenges partly because it has been marketed extraordinarily well. After all, the principle of Cloud Computing is not at all new and there is nothing especially stunning about how it works. Moreover, as chapter five addresses, there are numerous social problems associated with the role of Cloud Computing in the Next Internet. Effective marketing was absolutely essential to sell the necessity of filling the world with giant data factories and convincing individuals and institutions to move to the Cloud. Beginning with the image of the ethereal, if not quite immaterial, puffy white cloud, that replaced the big, ugly data center in the minds of potential users, Cloud Computing benefitted from a global advertising campaign, including early, very expensive ads telecast during the U.S. Super Bowl, annually the highest rated broadcast in America. Nearly all of the major cloud companies produced ads but the prize for chutzpah went to Microsoft for a widely circu- lated video. It featured a mom about to take a photograph of her family. Like most families it was hard to get everyone to sit still, smile and appear, well, family-like and the photos reflected this. Undaunted, mom turns “to the Cloud” where she finds photos featuring faces of her family members in smiling repose and swaps them for the desultory images she has just taken. The result, “finally a photo I can share

26 Becoming Digital: Toward a Post-Internet Society

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

without ridicule.” Then the coup de grâce, “Windows gives me the family nature never could.”

Along with the big ad campaigns, Cloud Computing benefited from free publicity compliments of promotional blogs and posts by self-proclaimed industry experts who vouched for the revolutionary significance of the Cloud. Complementing these forms of advertising and pseudo- advertising, leading private think tanks, including Gartner, McKinsey, Deloitte, and Forrester, produced reports con- cluding the same thing, only in more scientific language. Even the World Economic Forum, the leading intellectual force behind neoliberal globalization, chimed in with its own report calling on the world to join the cloud revolution. Inevitably these reports would find their way onto blog posts to produce a supportive echo chamber. Finally, the cloud’s arrival was fortuitously timed because it appeared just when Silicon Valley was waking up to the need to build its lobbying presence in Washington on a number of issues ranging from immigration to surveillance. As a result, pro- ponents of Cloud Computing were able to create a strong promotional force just when legislators were learning about some of the social and environmental problems associated with the Internet.

All of this benefitted the cloud computing industry and especially the major U.S. participants. These were also fortu- nate enough to emerge at a time when Europe’s telecommuni- cations providers, including Siemens, Ericsson, Nokia and Canada’s Nortel, and Blackberry, which once dominated the industry, fell into decline. Failing to foresee and take advan- tage of the burgeoning Internet and unable to keep up with Apple’s near universally popular iPhone, these companies survived, some barely, as much lesser participants in a global information technology upheaval. That left Amazon, Google, Microsoft, Apple, Facebook, IBM, Salesforce.com,

27Converging Technologies

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

http://Salesforce.com

Rackspace, and a few others to control the market. The only genuine challenge to U.S. domination in the Cloud comes from China where the giant firm Alibaba is using its power in China’s domestic cloud marketplace to build an interna- tional cloud business sufficient to take on the major U.S. firms. The next chapter addresses the political ramifications of this development. Now it is time to turn to Big Data Analytics, the second leg of the Next Internet, in its own right a major force in the success of Cloud Computing.

THE NUMBERS WILL SPEAK FOR THEMSELVES

Converging with Cloud Computing, Big Data Analytics refers to a system for the quantitative analysis of large data sets. In spite of the proliferation of fancy new titles that fuel enthusi- asm, like data science professional, there is very little that a social scientist would find novel in the big data approach. It generally involves taking a large, often massive, and, almost always, quantitative batch of data, and examining the specific ways the data do or do not cohere or correlate. The aim is to draw conclusions about current behavior and attitudes and go on to make predictions. Specifically, big data analysts produce algorithms or a set of rules that specify conclusions to be drawn or actions to be taken under specific conditions. Facebook, for example, takes the data generated by its two billion or so users and relates the “likes” associated with posts about everything from celebrities, companies, and politicians to views about society, products, and, of course, cats. These enable the company to develop profiles on its sub- scribers, which Facebook sells to marketers who target users with customized ads sent to their Facebook pages. Google does the same for search topics and, until public pressure forced Google to announce that it would end the practice by

28 Becoming Digital: Toward a Post-Internet Society

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

the end of 2017, the company has scanned the content of Gmail. Amazon creates profiles of its users based on searches and purchases on its site.

What fuels the enthusiasm for Big Data is the massive growth in the amount of information readily available for analysis because it is now organized in an accessible form in cloud data centers. The term “data mining” is appropriate because it suggests that the payoff is in finding valuable nuggets in the relationships among seemingly unconnected behavior and attitudes, what is referred to as “productivising data.” The early development of the Cloud and Big Data, particularly the massive information gathered by search engines, powered predictions of an intellectual revolution that would sweep away traditional academic disciplines and research approaches. Leading the charge was Chris Anderson who in 2008 wrote in Wired magazine about “The End of Theory.” With enough data and applied mathematics we can finally declare: “Out with every theory of human behaviour, from linguistics to sociology. Forget taxonomy, ontology, and psychology. Who knows why people do what they do? The point is they do it, and we can track and measure it with unprecedented fidelity. With enough data, the numbers speak for themselves.”2 Anderson was not alone, as devotees from the sciences to sociology and even the humanities repeated this rallying cry, if with slightly toned down language. The prospect of taking the banality of everyday existence and turning it into a digital sublime of predictive power sparked talk of revolutionizing our ways of knowing and of creating new ways of being and doing.

Big Data promised to radically renew the philosophy of positivism that the nineteenth century social scientist August Comte took from the hard sciences and applied to under- standing society. For Comte, empirical observations and the data of experience alone qualified for a true scientific

29Converging Technologies

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

understanding of the world. There would be no room for the- ory or what was called, disparagingly, metaphysical specula- tion. Today’s versions of Comte aim for a digital positivism that realizes the French sociologist’s goal of making the facts of life speak for themselves. To carry out this goal, corpora- tions, government agencies, scholars, civil society organiza- tions, and social movements dig into the data to find connections and create predictive algorithms. A new occupa- tion, the data scientist, was created to lead the effort and universities developed new degree programs to train big data professionals.

The results were mixed and the messianic zeal for Big Data has somewhat abated. In 2012, the leading research company Gartner predicted that by 2020 there would be a shortage of 200,000 data scientists. Five years later it forecast that by 2020 automation would kill 40 percent of existing data scientist jobs. Nevertheless, along with the growth in data sets, analytics research is expanding across a wide range of applications. The largest beneficiaries are the big data collectors, including Google, primarily through its search engine and email program, Facebook, the overwhelming leader in social media, and Amazon, which gathers data on its retail customers to be stored and analyzed in its own cloud division Amazon Web Services. In addition to using this data internally to match user profiles to what customers like and purchase, Amazon sells the data to advertisers who market products and services on their own sites and on those target customers are likely to visit. So when I searched Google to learn whether the Nordstrom Rack store that just opened in my neighborhood had the running shoes I like, ads for the store began popping up on my New York Times site and on most others I searched. That is because research determined that it was worthwhile for the wholesale advertising pur- chaser to make this information available on other sites that

30 Becoming Digital: Toward a Post-Internet Society

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

share an interest in my profile. Completing the circle, retailers contract with companies that have access to my actual pur- chases through credit card data and other sources. In that way, they can determine what advertising actually leads to a buy and construct an algorithm that applies to people with my profile in order to predict the advertising that is likely to work in the future. Retailers also use Big Data to speed up price changes and determine product placement.

Although advertising and retail help drive big data research, there are many other applications. Amazon itself sells access to the data that manufacturers use to improve the time it takes to get products to market, to improve supply planning, and to forecast product demand. Utilities compare customer profiles with data collected from intelligent home and business meters to improve forecasting of energy needs and respond more effectively to changes in demand. From the drones that carry out aerial photography to precision moni- toring of crop yields, water, and pesticide use, farmers are able to use data analytics to practice what is increasingly called “precision agriculture.” Researchers in the health care industry make extensive use of Big Data to forecast the spread of communicable disease and develop algorithms to predict with precision rates of well-being and mortality among different populations. These results are understand- ably of great interest to risk assessors like insurance compa- nies and credit bureaus, which use their own data to determine qualifications and set rates. With Big Data, risk assessment has become big business.

Governments have been making extensive use of Big Data at all levels. As the Edward Snowden revelations of 2013 and the WikiLeaks data dump of 2017 demonstrate, U.S. intelli- gence agencies have employed Big Data Analytics gathered in their own as well as in commercial cloud systems. Building profiles and carrying out sophisticated network analysis of

31Converging Technologies

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

potential terrorists, spy agencies provide the models for secu-

rity and policing in the United States and wherever it has

influence around the world. Moreover, law enforcement

increasingly, and not without controversy, makes use of

predictive analytics to forecast the neighborhoods and people

who are likely to commit crimes. Outside these areas, Big Data Analytics is an essential tool

to deliver government services, determine eligibility and track

users. Educational institutions have deployed analytics for

some time, particularly to manage enrolments and, at the uni-

versity level, to reach out to potential applicants and donors.

As one would expect, scholars in the hard sciences make

heavy use of big data tools, everything from high-energy

particle research to a project that organizes the 25,240

notebook entries that Charles Darwin produced leading up to

his Origin of Species. They are also increasingly deployed in

the social sciences and even in the humanities. In fact, as a

result of major U.S. government support, what is called the

Digital Humanities movement incorporates Big Data to carry

out with relative ease what were once daunting, if not impos-

sible, quantitative research on the contents of written texts.

While one might question the necessity of some of this

research, for example, we now know that James Joyce leads

major novelists in the use of exclamation points with 1,105

per 100,000 words, there is no doubt that the Digital

Humanities has changed how scholars carry out research in

centuries-old fields. Given the inexorable growth of urbanization worldwide,

big data researchers have made “smart cities” a special focus

of attention with high hopes of using the data gathered on

transportation, energy use, environmental pollution, waste

management, and the health of citizens to create more liveable

and sustainable urban areas. Cities like Oslo, Norway, Portland,

32 Becoming Digital: Toward a Post-Internet Society

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

Oregon, and the Madrid suburb of Rivas Vaciamadrid, home to the Ecopolis smart city project, are notable examples.

Notwithstanding these growing applications, Big Data Analytics has not lived up to Anderson’s forecast that it would radically transform research and bring about the end of theory. One reason why it has not sustained early expecta- tions for revolutionizing applied social science, policy forma- tion, and academic disciplines, is a series of highly criticized studies that demonstrated Big Data’s limitations. In 2009, Google researchers published a study examining search terms associated with the percentage of visits to physicians where influenza-like symptoms were diagnosed. This enabled, they claimed, the prediction of flu outbreaks with almost no time lag (one day) as compared to the best that the U.S. Centers for Disease Control could provide (two weeks). If verified, this might qualify as the kind of radical research finding that inspired Chris Anderson’s vision of sweeping away our intel- lectual past and replacing it all with numbers that speak for themselves. Optimism rose after the initial success, but the researchers’ algorithm failed miserably in the 2012�2013 flu season when the model forecast higher than average levels of flu that did not materialize. It turned out that people were searching a lot more for flu information because they were more aware of the damage flu can do and because media saturation spread a panic over flu that sent people to their search engines in record numbers. Google promised to improve its model. But the damage was done. Often described as a tool to find a needle in a haystack, Google’s researchers demonstrated that the needle could easily be lost again.

The Google flu fiasco gave way to stories about ethically challenged big data research at Facebook and the match- making dating service OKCupid. Facebook decided to engage in a big data-sized mass experiment to determine whether it

33Converging Technologies

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

was possible to use social media to influence emotional states. It did so by choosing 689,003 users whose news feeds were manipulated without their knowledge to present almost exclusively positive or negative stories. Using the standard obscure consent agreement that most people never read to justify the ethics of the project, Facebook was heavily criti- cized and Big Data got a big black eye. Following the furore that erupted over this 2014 story, OkCupid admitted that it too was engaging in user manipulation. The dating service was carrying out research to determine if love would blossom even if it recommended partners that its own algorithms had already determined were simply not compatible. The com- pany was not helped by its justification: everybody does it. Big Data took another big hit.

Then there was the doctoral student who undermined one of the research pillars of the neoliberal globalization strategy, which policy-makers at the International Monetary Fund, the European Union and other powerful institutions had been using to promote austerity measures. The study in question was a big data analysis of the relationship between national debt and economic growth. Specifically it drew from massive historical data sets to justify the claim that once a nation’s debt reaches 90 percent of its Gross Domestic Product, growth slows. The research was used to buttress painful eco- nomic measures, including cutbacks in government spending on social programs. These were justified as the necessary short-term pain that would lead to long-term social benefits. The authors became academic stars. All that ended when the student discovered major flaws in the research that signifi- cantly challenged the results. Many of the errors proved obvi- ous, including spreadsheet mistakes, but others involved intense scrutiny of how variables were weighted. Such scru- tiny is not easy in research involving large amounts of infor- mation and, while controversy continues to brew over

34 Becoming Digital: Toward a Post-Internet Society

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

the story, Big Data’s big promises were once again under- mined. Many concluded that Big Data can easily lead to big errors that are hard to track but which carry big human consequences.

As if the world needed another example of Big Data’s lim- itations, along came the election of President Donald J. Trump. Big Data was used throughout the two-year election cycle and then roundly criticized in the aftermath of the elec- tion. Building models based on historical polling of state and national races, almost all pollsters forecast a Clinton victory with a probability of anywhere from 70 to 99 percent with one major pollster agreeing, if Trump won, to eat a bug on live TV (he did). The reasons given for this failure seem end- less and some understandably defend Big Data because the popular vote margin predictions were generally upheld and because, in a story given very little attention, the Trump cam- paign successfully made use of analytics to slice and dice the American electorate to win just enough states to put him over the top. But there is no backing off from the conclusion that faith in Big Data played a significant role in the belief that Hilary Clinton would win, with relative ease, an Electoral College victory. There is also evidence that this faith may have influenced her campaign strategy. After all, advised by her analytics people that the important state of Wisconsin was a shoo-in, Ms. Clinton did not visit in the last six months of the campaign. Similar advice convinced her that Michigan was a lock and that Republican states like Arizona were in play. We do not know for sure but faith in big data-based forecasts may have caused some of her supporters to stay home from the polls. This is not an excuse for the campaign’s other failures and there is no gainsaying the ultimate bril- liance of the Trump campaign’s strategy. Like the Brexit campaign, it tapped into and exploited anger about the status quo that could not be easily measured by statistical

35Converging Technologies

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

techniques. Both events were major shocks and once again provided evidence that something is wrong with Big Data that needs to be fixed to avoid future failures that would threaten the legitimacy of the approach.

This may be the case. But having carried out research on both large and small data sets, I cannot conclude that there is anything inherently wrong in Big Data Analytics, assuming that it is carried out carefully and also provided that the methodology is viewed, and used, as one among several research approaches. Contrary to the dreams of enthusiasts, Big Data Analytics is not the answer to every question. It is especially not useful to apply Big Data to solve problems that may not actually exist or are less than significant. The ten- dency to address topics not because of their importance as social problems but rather because they can be easily quanti- fied using one or another technological tool, what Evgeny Morozov aptly calls “solutionism,” wastes resources, ignores genuinely significant issues, and does little for the reputation of scholarly research.

Big Data analysis requires a degree of care that is not suffi- ciently recognized. As the research on national debt and economic growth demonstrated, small errors like a few inaccurate spreadsheet cells, can have major negative conse- quences. Errors buried under a mountain of data can easily be left uncorrected and lead to inaccurate results. Moreover, even mountains of quantitative data are often no substitute for qualitative research whose findings result from in-depth interviews and actual observations of human behavior that reveal the fullness of human subjectivity. It is not easy, if at all possible, to understand how people make meaning out of their lives and their social situations with quantitative data alone. Quantitative research often tries to do so by assigning numerical values to subjective states. However, these are, at best, rough approximations that suggest more than they can

36 Becoming Digital: Toward a Post-Internet Society

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

actually warrant. Assigning a numerical value to “agree” and another to “strongly agree” might help create a large data- base, but actually say little about the qualitative feelings of respondents. Even if it were ethical for Facebook to carry out research on emotional states by manipulating newsfeeds, which most now agree it was not, one should doubt that its researchers could really determine the change in emotional states simply by assigning quantitative values to positive and negative user comments. This would at best be a very rough approximation, a precursor to methods, such as in-depth interviews, that are much more reliable for understanding changes in subjective states.

It is tempting and more than a little exciting to think that Big Data Analytics can sweep aside traditional approaches. But those who refuse to see it as just one among a set of approaches also miss the importance of theory and of histori- cal context. Avoiding what have been the twin pillars of social research makes it easier to find causality in the correla- tions that big data analysts discover. But correlations are merely associations between variables, suggesting at best the strength of a relationship between, for example, search terms and the incidence of flu. Before Big Data euphoria promised an intellectual revolution, elementary social science explained that correlation is not causality. It still isn’t. Apparent corre- lations are often spurious and therefore misleading because they can neglect another factor, like media attention to flu, that might lead people to increase their search for flu answers without any apparent symptoms.

A rich understanding of theory and context is the best way to avoid the problem of misinterpreting the results of correla- tional analysis. Theory provides the narrative that helps make sense from a mass of data. Best when it is carefully grounded in both data and past research, theory requires an appreciation of causality. Not necessarily the search for the

37Converging Technologies

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

sole determining force, causality can also mean setting limits and describing a process of mutual constitution. Data do not speak for themselves. We give data voice with theory.

Because Big Data tends to examine human behavior as a set of data points comprising discrete, individual events, it tends to resist the need to examine context and history. This is fundamentally misguided because behavior is not made up of distinct events and people are not data points. What we do, our behavior, and who we are, our character, are all deeply embedded in richly connected sequences. The popular- ity of stories is evidence that we understand this. The digital positivism of Big Data tends to neglect, if not reject, this view, seeing in our attraction to stories little more than evi- dence of a more primitive state that needs to be replaced by examining statistical relationships among a mass of discrete data points. To use the language of big data enthusiasts, unless used carefully, the approach that claims to find nee- dles in haystacks often produces only bigger haystacks. Understanding haystack formation is the job of history, which makes it an important tool, one certainly not to be discarded in the race to embrace the new.

To summarize, Big Data would be far better an instrument for understanding the world if researchers, especially data scientists, paid explicit attention to the lived, qualitative expe- rience of subjects over simply focusing on what can be repre- sented in data. Big Data Analytics would be a far better approach if it paid attention to constructing causal argu- ments, rather than assuming that correlation is causation, which it clearly is not. Nor is causation a hammer with which to beat reality into submission. Causality sets limits, suggests relationships, and maps the processes of mutual constitution and multiple determination. In addition, Big Data would benefit from more attention to theory and not burden data with the need to explain itself. Finally, it would be a sharper

38 Becoming Digital: Toward a Post-Internet Society

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

tool if it rebalanced the tendency to unabashed presentism with a willingness to entertain the importance of history and context.

Problems aside, there remains significant support for Big Data. Some of this inevitably comes from the power of the method as a revenue-generating machine, particularly among the leaders of Big Tech such as Facebook and Google who use analytics to dominate the digital advertising industry. But support also comes from those who are sharply critical of what these companies are doing. For example, the noted criti- cal journalist George Monbiot has made the case for deploy- ing Big Data as an instrument to advance democracy. He has argued that analytics can be used to make political decision- making more transparent. Social movements can use the “wisdom of crowds” to expand democratic policy-making, to propose ideas that do not occur to professional politicians, and to find problems and loopholes in proposed legislation. In Monbiot’s view, people will either control technology, or it will control them. I will have more to say about this in chapter six when the book turns to proposals for addressing the major social and political problems associated with the Next Internet. It is time to turn to what is arguably the most significant reason support for Big Data continues to grow, the third leg of the Next Internet: the Internet of Things.

BRINGING THINGS TO LIFE

The Internet of Things is a system for measuring, monitoring, and controlling the activity of objects and living organisms through sensors that gather, process, and report data over networks, including the Internet. From watches that measure blood pressure to refrigerators that tell us when we need more milk; from assembly lines worked by robots, to drones

39Converging Technologies

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

that deliver what robots produce, the Internet of Things pro- mises a profound impact on individuals and society. We refer to the admittedly awkward term the Internet of Things because unlike the Internet we know, the Internet of Things electronically connects things in digital networks. The sensors in a refrigerator form a network in constant communication with a cloud data center. They might monitor food use and report to us, perhaps through a panel on the door, when we need to stock up again. The sensors that surround the body of a Tesla automobile form a network that scans the car’s internal functioning and external environment in order to permit various degrees of autonomous driving. The Internet of Things is made possible by advances in the ability to mini- aturize scanning devices and provide them with sufficient pro- cessing power to enable the ability to monitor activity, analyze data on their use, and deliver the results over elec- tronic networks.

Most of what has been written about the Internet of Things is technical, promotional or both. Like most early analyses of new technological systems, it is satisfied to describe the technology, to encourage its use, and to situate the Internet of Things in utopian discourse. It is now hard to imagine an object that cannot literally contain a computer or enough computing power to warrant branding as an Internet of Things device. Those who have worked with digital scan- ning technologies are correct to point out that the Internet of Things is not an entirely new development in the information technology world. For the past 30 years, since the develop- ment of microprocessors and network-based instruments, companies in the oil and gas, chemical, pharmaceutical, manufacturing, and mining industries, have explored how to make use of sensors to make their business processes more efficient and safe. Unanticipated technological complexity leading to less than flawless operation and serious security

40 Becoming Digital: Toward a Post-Internet Society

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

issues (the Internet of Hackable Things is already an overused phrase), lead experts to advise that the Internet of Things is a slowly developing technology rather than the “killer app” that will instantly disrupt the world. Nevertheless, today’s Internet of Things is qualitatively different because scanner costs continue to decline and because the limit on packing more computer power into devices has not been reached. The Internet of Things also benefits from the global digital net- works that are now in place, and which continue to expand. Finally, the Internet of Things is significantly supported because of the opportunities for convergence with Cloud Computing and Big Data Analytics.

The term Internet of Things is not precisely accurate because the devices that comprise it also connect people. For example, the Apple Watch reports on its user’s activity and sends to the user (and perhaps to other parties such as health authorities, insurance companies, and credit bureaus) infor- mation on the body’s functioning and activities (steps walked in a day, heart rate, blood pressure, etc.) Some of these devices are actually embedded under the skin to improve health monitoring but implants are also intended to deepen surveillance as well. For example, a company in Sweden pro- vides subcutaneous chips to its employees, removing the need for passwords to enter secure areas and to login to company sites. It also enables detailed performance monitoring. A firm in the United States markets a smart condom, which not only measures sexual activity, but it also claims to test for sexually transmitted disease. Another company is selling an app- enabled vibrator that can regulate and adjust sex toys from a phone or tablet (although it cost the company $5 million to settle a privacy lawsuit when a customer learned that the company was selling information on users without their informed consent). It is also the case that the Internet is not the only network for the Internet of Things. Companies and

41Converging Technologies

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

governments often make use of private networks with limited access and expanded security. Nevertheless, because the phrase Internet of Things is already in widespread use, it is best to go with it rather than add to the already long list of terms that fill descriptions of the online world.

The Internet of Things is also about the mundane, but powerful, economic and political demands of a global system. Whether global corporations are in the business of manufacturing things or providing services, their leaders see the Internet of Things as absolutely vital for their ability to produce more efficiently by expanding opportunities to add artificial intelligence to production facilities through so-called smart assembly lines, and robotics. They also imagine greater opportunities to distribute goods more efficiently and effec- tively at a lower cost. Not everyone will mimic Amazon’s army of delivery drones, which the company expects will move goods to customers within 30 minutes of an order, but the principle of automated delivery, whether of clothes or university courses, provides a popular model for organiza- tions in the twenty-first century. The opportunity to make more productive use of technology with many fewer workers makes the Internet of Things an essential consideration for organizations of every type, private and public, small and large, whether producing goods or delivering services.

The mere prospect of realizing this potential has greatly expanded the number of companies drawn to the opportunity to develop Internet of Things devices. The prospects for accel- erating demand appear irresistible. Factories filled with robotic devices, scanners embedded all along assembly lines, and sensors monitoring industrial equipment like pumps, gen- erators, motors, gauges, and switches comprise the industrial Internet of Things. These automated factories will increas- ingly produce the objects that make up the consumer Internet of Things whose driverless cars, intelligent homes, and

42 Becoming Digital: Toward a Post-Internet Society

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

biometric scanners, provide enormous opportunities to expand the sheer number of products on the market. There will be more things to want and more motivation to replace the things we like because smarter versions will constantly appear. The Internet of Things takes what used to be called planned obsolescence to entirely new levels. Create a small wand that can be passed over bar codes to ease the process of reordering and the two forces that often slow the consump- tion process, having to leave the home and facing a price label, practically disappear. In fact, in some cases consumers will not even need to “push now” because smart kitchens and laundry rooms will be programmed to automatically reorder. Wear a chip under your skin and you no longer need keys or passwords. The potential for self-monitoring is practi- cally unlimited. In addition to the value of expanded con- sumer choice in goods and services, the Internet of Things massively increases the amount of valuable data gathered on consumers that can be sold to advertisers and other inter- ested parties including, for example, insurance firms that provide medical and health coverage, credit bureaus that determine mortgage, auto, and other loan qualifications, and law-enforcement authorities monitoring citizen behavior (Table 1).

The Internet of Things emerges in a political as well and an economic context. Since 9/11 that means governments committed to vastly tightened security and strengthened intel- ligence gathering, all in the interest of fighting terrorism. So it is no surprise that governments everywhere are giving special attention to the massive new surveillance opportunities that the Internet of Things presents. Billions of interconnected devices monitoring the most minute activities of people and things can provide the National Security Agency (NSA), the Central Intelligence Agency (CIA), and their counterparts in other nations with what amounts to billions of new eyes

43Converging Technologies

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

trained on people and institutions all over the world. Any human activity that makes use of a connected device and any process or practice of the businesses, government agencies, schools, and voluntary organizations that use the Internet of Things is trackable. So in addition to the efficiencies that the Internet of Things can bring to public agencies at every level, there is the specter of Big Brother and the seemingly insatia- ble appetite to gather, process, analyze, and use electronic surveillance.

For companies that provide Internet of Things services, mastery of the new technology of connected devices is abso- lutely vital for their growth, if not for their survival. These include the firms that are known around the world as leaders in the online world, such as Amazon, Microsoft, Google, Apple, and Facebook who, in their own different ways, are developing a strong Internet of Things presence within, and independent from, their social media strengths. The Internet of Things is especially critical for Amazon because, more than the others, the world’s largest bookseller is now in the

Table 1. Forecast Growth in the Number of Internet of

Things Devices (in Billions). 2018: 23.1

2019: 26.7

2020: 30.7

2021: 35.8

2022: 42.6

2023: 51.1

2024: 62.1

2025: 74.4

Source: Forbes, November 22, 2016, http://bit.ly/2oLMkwz.

44 Becoming Digital: Toward a Post-Internet Society

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

http://bit.ly/2oLMkwz

business of delivering a wide variety of goods. That is why Amazon is racing to develop consumer friendly digital assis- tants, robotic processing of warehouse orders, drone delivery, and new forms of packaging that automate reordering. The Internet of Things is also critical for older or legacy firms like General Electric and IBM that view the new technical system as an opportunity to retool their industrial base by expanding robotic production, taking advantage of enhanced scanning and monitoring capabilities, and making extensive use of Big Data Analytics.

With the growth of the Cloud and Big Data Analytics, the Internet of Things has a strong foundation upon which to build. The IT giant Cisco projects that 50 billion devices will be connected by 2020 and 82 billion by 2025, with the tech- nology and services market reaching $14.4 trillion by 2022. Even if we discount the hyperbole that typically accompanies industry forecasts, these are, by any measure, massive growth expectations. Intel, using a broader definition of a device envisions 200 billion connected objects equipped with sensors and connectivity by the year 2020. Nevertheless, as Cisco also reports, there is a great deal of attrition in the Internet of Things market with three out of every four projects ending in failure.

It is easy to get lost in the mountain of terms and statistics that accompany promising new technologies like the Internet of Things. Before proceeding further down that road, it is useful to consider concrete examples of how successful com- panies are envisioning the future in this brave new world. Amazon leads the consumer Internet of Things and General Electric is arguably the major corporate power in industrial applications. Although its shareholders complain that the company is too eager to opt for new lines of business over expanding earnings, Amazon, by most measures, is one of the world’s more successful companies. Most people are aware

45Converging Technologies

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

that it is a leader in mail order retailing but fewer know about its success in the post-Internet world of Cloud Computing where Amazon is widely recognized as the domi- nant player. Its cloud subsidiary Amazon Web Services con- trols over one-third of the market in the Public Cloud, the primary form of Cloud Computing, well ahead of its nearest competitor Microsoft with less than 20 percent market share. Amazon entered the cloud computing business because it stored customer information in data centers. The company decided to take advantage of its technological expertise sell- ing cloud storage and extended that by offering value-added data processing. By combining this with its experience in retail discounting, Amazon was able to undercut the competi- tion on price and rapidly rise to a dominant position.

The company has done the same with Big Data Analytics. From its early days, Amazon made use of analytics to assess and use the mountains of customer data it receives every day. Now Amazon is offering low-price data science, what it calls Amazon Machine Learning (AML), to customers, like small businesses, that might benefit from Big Data Analytics but cannot afford a full-service data science operation like IBM. AML is built on the same technology that Amazon uses to create algorithms that send book recommendations to email inboxes every day.

Amazon’s vision of the consumer future begins with Alexa, a voice-operated digital assistant that responds to commands and carries out tasks such as turning on home lights, ordering music, and carrying out searches. This is com- plemented by the Dash button, a device about the size of a flash drive that appears on products and carries the logo of a product and a button. It is integrated into home Wi-Fi sys- tems and, at the push of the Dash button, can reorder the product from Amazon for next day delivery to Amazon Prime customers. The company which pioneered one-click

46 Becoming Digital: Toward a Post-Internet Society

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

shopping on its website is also working with manufacturers to embed the Dash button in product packaging. In addition, Amazon is producing a device that can read bar codes for instant reordering of any product and with others, like Whirlpool, to automate the reordering process through the appliance itself. Think of an intelligent washing machine that orders its own detergent. The goal is to automate consump- tion. In the process, Amazon also collects massive amounts of information on its customers’ everyday lives that the com- pany uses in its own business and which it markets to compa- nies wanting to know about customer tastes and shopping habits. Competitors recognize the threat and are responding. In May 2017, Walmart filed a patent for a sensor that the company would embed in all of its products to monitor con- sumer behavior, tighten inventory control and permit auto- matic reordering.

Once Alexa, Dash or their equivalent orders a product, attention shifts to Amazon’s automated warehouse where robots, more like driverless vehicles, motor down “streets” to find, fetch, and deliver to a human “pick worker” the ordered product. The laborers who used to roam warehouses in search of products are no longer needed, at least not in nearly the same numbers. Their job now is mainly limited to lifting merchandise orders from a pod and placing them in a mailing box. Amazon expects that the process will soon be completed by automated delivery through an army of drones. Traveling at 50 MPH through airspace between 200 and 500 feet and weighing about 55 pounds, these flying delivery devices will carry packages weighing up to 5 pounds through travel corri- dors up to 10 miles long. Amazon’s goal is to have drones complete delivery within 30 minutes of an order.

Automated ordering, automated order fulfillment, and automated delivery. This is Amazon’s vision for the near future with the Internet of Things. Alexa is one of the

47Converging Technologies

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

company’s most heavily marketed products (“Alexa,” goes the commercial, as Dad reads to his child, “what sound does a whale make?”). Driverless vehicles are already operating in the warehouses; you can order a free (if you are an Amazon Prime customer) Dash device now (one click!); and the drones are in the test stages of development. What justifies such a massive investment is the enormous savings in labor costs, the likelihood that the easier it is to make a purchase, the more people will buy, and the more finely detailed data to be gathered on people’s purchasing habits and used internally or sold to other companies. There is no guarantee that any or all of these systems will work well on their own or together to produce a seamless purchase�delivery process. Much can go wrong along the way to disturb the process, from hackers making digital mischief to the many errors that disrupt physi- cal systems, not the least of which is the potential for drone failures. Yes, they test well in the wilderness of British Columbia but this is no guarantee they will work on the streets of Chicago. Nevertheless, the Amazon plan empha- sizes just how close the world is coming to the deployment of Internet of Things technologies and, as subsequent chapters detail, just how essential it is to develop policies to deal with them.

When we turn to the industrial Internet of Things, it is use- ful to provide illustrations from General Electric, by today’s standards an ancient company, but one that has remade itself many times over. On April 10, 2015 the then 123-year-old firm announced yet another makeover, this time finding its future in its past. As the banner headline in the Financial Times read: “GE to shed financial services arm in radical return to industrial roots.” The move earned the attention because it marked a major retreat from the financial services sector, which, in the years leading up to the 2009 industry meltdown, dominated global capitalism. The company still

48 Becoming Digital: Toward a Post-Internet Society

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

brought in half of its earnings from its GE Capital unit, so selling off the unit and returning to its once core business of making jet engines, oil drilling equipment, and medical devices was no minor decision. However, returning to its industrial roots has nothing to do with factories filled with skilled assembly workers earning a union wage. Rather it means a manufacturing space that deploys robots, 3D prin- ters, ubiquitous sensors, and massive connectivity with far fewer human beings in the mix than would populate most industrial plants in the twentieth century.

GE’s iconic slogan used to be “We bring good things to life.” Today, that phrase holds a deeper meaning because the company not only plans to once again bring things to life, it also plans to do so by bringing to life the very factories that build them. The company calls it the “brilliant factory,” a manufacturing system whose machine parts communicate con- stantly with operators who can schedule maintenance before breakdowns occur and otherwise adjust the manufacturing pro- cess to changing conditions and fluctuating product demand. Factories have always been able to make adjustments, more in small batch than in larger production runs. But there is a quali- tative difference in the agility of production when an assembly process contains 10,000 sensors that oversee and communi- cate instantly about the production of a battery in GE’s Schenectady, New York plant. GE believes that the Industrial Internet of Things could add between $10 trillion to $15 tril- lion to global GDP by 2035 through more efficient and agile production, maintenance schedules that are fine-tuned to the specific materials that make up the manufacturing process, and lower cost repairs to industrial equipment. That would be the equivalent of adding another China or United States to the global economy.

Like the Cloud and Big Data Analytics, the Internet of Things faces major technical challenges. These include

49Converging Technologies

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

significant security issues. The Internet of what some call Hackable Things is vulnerable to breaches because producers, especially of Internet-enabled consumer devices, have been so eager to get their toys, watches, and thermostats to this highly competitive market that they have paid insufficient attention to software and data protection. Experts on the subject agree that there is almost no incentive for manufac- turers to make their products secure because it is cheaper to sell them with minimal protection and then market new ver- sions. That is why some industry people, to the surprise of observers, are calling for government regulation to prevent massive market failure. In the meantime, hackers have attacked individual devices and, more importantly, have mobilized large numbers of devices and ordered them to attack key network nodes. These “denial of service” attacks have crippled the networks of major providers by overload- ing their systems with demands from easily hackable thermo- stats, toys, and other Internet-enabled objects. In one such attack, hackers leveraged 2.5 million unsecured Internet of Things devices to target servers operated by a U.S. company providing domain name servers to several large and popular sites, including Netflix and Reddit. As a result, large portions of the Internet simply disappeared from view.

Moreover, consider that the billions of new online devices will require frequent software updates to remain secure and also to improve functionality. Many of these operations will be the responsibility of the consumer who is unlikely to have much interest in downloading an update to a teddy bear that can deliver music stored online. Companies will likely pro- mote new teddy bears, thermostats, or watches because they contain the latest in secure software and added capability. Given how often people now buy new smart phones and other gadgets equipped with “the next new thing,” however banal, firms are confident that customers will open their

50 Becoming Digital: Toward a Post-Internet Society

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

wallets for the devices equipped with 2.0 capabilities. However, this strategy is not without the risk of consumer fatigue and frustration that might tempt users to stash their Internet-enabled devices in the basements and attics that already contain numerous generations of what used to be their favorites, as they choose to return to a more analog life. The early signs of worry for the Internet of Things industry were evident by the first half of 2017 when it was reported that 75 percent of IOT projects had already ended in failure.

THE POWER AND PERIL OF CONVERGENCE

Cloud Computing, Big Data Analytics, and the Internet of Things each contain great technical power and potential. This power is multiplied when they are brought together or converge in the networks that comprise the Next Internet. The Cloud provides essential storage and processing; Big Data offers new opportunities for adding value to this stored information; and the Internet of Things collects mountains of data for analysis. Technological innovations, including work- place robotics, autonomous vehicles, and weaponized drones all require the tight integration of these three systems. Through their sensors, production-line robots are in constant communication with cloud data centers, which modify robot activities on the line based on performance indicators that undergo constant analytics review and whatever changes managers want to make in output specifications. No auto- mobile could drive without a human behind the wheel and no battlefield drone would be able to produce “bugsplat,” the gruesome term pilots like to call their human prey, with- out a high degree of technological convergence.

Convergence is far easier to talk about than to achieve in practice. Technical systems that power the Cloud and the

51Converging Technologies

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

Internet of Things require complex designs or architectures to enable their functioning, even as they protect themselves from external attacks. It is all the more difficult when these already complex systems try to achieve seamless interconnection. Having installed a Wi-Fi thermostat in my home, I learned just how difficult convergence can be when it stopped com- municating with the data center servers that were to enable me to make changes in my home environment from an app on my phone. Having experienced a furnace shutdown that can lead to disaster in a Canadian winter, I installed the system to permit me to keep an eye on it when, like many other aging Canadians, I flee to the warm South in winter. Whether it was a data center outage, overloaded servers, a bad router, or faulty networks, I will never know. But for more than one frustrating month, the system was down more than it was up and running.

As the earlier discussion of Edge computing demonstrated, figuring out the proper relationship between devices and data centers is a big debate topic among those designing and build- ing the Next Internet. Convergence is not just about intercon- necting systems and enabling them to communicate. It also crucially concerns how to do this most effectively. In the rela- tionship between cloud data systems and the Internet of Things, there are questions about whether the former is equipped to bear the demand that the proliferation of network-enabled devices will continue to generate. Companies continue to build data centers as the devices proliferate. Will there be sufficient data center capacity and will these informa- tion factories be able to respond quickly and efficiently enough to manage the host of complex sensor systems that have little tolerance for error? Some argue that more of the intelligence and management functions need to be shifted to the devices themselves with architectures designed to create local networks that rely less, if at all, on the Cloud. At the

52 Becoming Digital: Toward a Post-Internet Society

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

extreme, there are those who envision a significantly reduced role for cloud systems in Next Internet networks. Others maintain that the Cloud has worked well enough so far and will likely be able to take on generations of new devices. There is a great deal at stake here, not the least of which is the safety of autonomous vehicles that demand near instanta- neous sensor response. Or the safety of the nuclear weapons arsenal.

There are, as yet, no clear answers and there may never be. Undoubtedly, adjustments will be made in the relation- ship between centralized servers and local devices for quite some time. However, what the debate does reveal is that con- vergence is more than a technical matter. It is not just about system architecture and design; it is also about the power relationships among companies that stand to gain or lose if one or another design path is chosen. Big investors in cloud computing facilities and services are counting on a massive boost in business as the Internet of Things advances. They do not want to see this expectation diminished or nullified by new architectures that take functionality from data centers and build it into the devices that foreshadow what some are already calling a “serverless future.” In essence, technology, convergence, and the design of networks are not just a matter of applied science. As the next chapter emphatically con- cludes, politics plays a significant role in each.

More than the Cloud, Big Data Analytics, and the Internet of Things make up the Next Internet. Telecommunications systems connect these systems and their wired and wireless links are vital. Big providers like Verizon see the value in expanding beyond serving as distributors of valuable data and are getting into the action. This helps to explain the com- pany’s purchase of Yahoo, a once successful Internet com- pany that still enjoys a following even though it has dipped in value well below the Big Five tech firms. Developments like

53Converging Technologies

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

this warrant close scrutiny, but so far telecommunications firms have not been particularly good at turning their net- work success into profitable Next Internet firms with a full set of Cloud Computing, Big Data Analytics, and Internet of Things services.

Telecommunications networks are important for conver- gence to succeed. However, even more important is a key element of convergence that tends to be ignored: the relationship between technical systems and human beings. Human�computer interaction has been an interest of scho- lars since the advent of the big, clunky word processors that initiated the personal computer era. In spite of rapid changes in systems, there was always something familiar about the relationship. After all, people had worked on typewriters for generations and so keyboards and even touch systems were not far from the human experience. Today’s systems present a new set of challenges. The rise of intelligent devices, robot- ics, and artificial intelligence systems now means a different relationship to devices. Previous systems, however complex, tended to give users a large degree of control. That is no longer the case. Increasingly, people work alongside or as assistants to robots and other intelligent devices that are more critical to the production process. Increasingly, people give up most, if not all, control to autonomous vehicles and to the algorithms that are now trusted to make decisions in business, government, and social life. Human convergence, for more and more people, means subservience to the tech- nologies and the algorithmic decision-making processes that make up the Next Internet. So much attention has been understandably paid to the concern that intelligent devices will lead to automation and the replacement of human work- ers that we tend to ignore the fact that most people, both at and away from work, will establish relationships, including strong emotional ones, with intelligent devices. What we are

54 Becoming Digital: Toward a Post-Internet Society

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

beginning to see with robotics and smart phones will only accelerate with the Next Internet. How this element of con- vergence does or does not work will go a long way to deter- mining the success of the Next Internet and most likely, at the risk of sounding dramatic, very much of our future.

NOWHERE AND EVERYWHERE

Complicating matters is an ontological shift, a genuine change in the nature of those things we somewhat anachro- nistically call computers. What is a computer? Is it the device sitting on your desktop? Is it your laptop? Your tablet? Your smartphone? Your watch? Your thermostat? Your car, which now operates like a data center on wheels? It is increasingly difficult to determine what is and what is not a computer because computers appear to be everywhere. In addition to the power of the three systems described in this chapter, genuine ubiquity is a vital characteristic distinguishing the Internet as we now know it from its successor.

Another way to think about this is that digital is disappearing even as it becomes more powerful. The analogy to electricity is useful. When the technology that would power so much of the twentieth century appeared, there was no missing it because electricity required the presence of big, bulky generators on site. Combined with the appearance of transmission cables seemingly everywhere, it was impossible to ignore the presence of electricity. Over time, things changed. The generators were centralized, the cables were buried (at least in most urban areas) underground, and the wires disappeared into the wood- work. Electricity lost visibility, even as it became more power- ful. It is not precisely the same for digital because digital is not quite intelligent electricity. However, the analogy is worth con- sidering because it appears that, as digital is becoming as

55Converging Technologies

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed

ubiquitous as electricity, it too is disappearing into the wood-

work of everyday life. We once viewed electricity as a thing

outside ourselves, embodied in very material technologies

obviously external to the people who used them. Similarly we

once viewed computers as contained within material machines

external to their users. Whether it was a mainframe, a work-

station, a personal computer, a tablet, or a smartphone, we

knew a computer when we saw it. It was external to and

clearly distinguishable from us. That is no longer the case. Not quite ubiquitous, computers are nevertheless increas-

ingly embedded in so many devices and bodies, including our

own, that they appear to be everywhere, yet nowhere in

particular. This represents at least the beginning of an onto-

logical shift in the nature of these devices. Less a thing out

there and more a characteristic of most everything we con-

front, including ourselves, the arrival of what was once called

ubiquitous computing is also a form of convergence. In this

case, computers converge with the world as we know it,

including objects, people, and nature itself. Computational

power is increasingly derived from its presence nearly every-

where, marking another significant difference between the

Next Internet and its predecessor. Another significant differ-

ence is the political economic context that shapes this power,

a topic taken up in the next chapter.

56 Becoming Digital: Toward a Post-Internet Society

C op

yr ig

ht ©

2 01

7. E

m er

al d

P ub

lis hi

ng L

im ite

d. A

ll rig

ht s

re se

rv ed