jgcopple

“Big data is the Enlightenment for the digital age.” So said Viktor Mayer-Schönberger, Professor of Internet Governance and Regulation at Oxford University and co-author, together with The Economist data editor Kenneth Cukier, of the popular 2013 book Big Data: A Revolution That Will Transform How We Live, Work, and Think. Just like the Enlightenment brought us reason and rationality, so will big data represent a paradigm shift that will help us unearth new insights as we gain a new perspective on reality.

But is big data ­- analyzing huge swaths of information to uncover insights and make predictions that were largely impossible in the past – truly poised to transform our lives, our jobs, and our science? In my view, the book is an important contribution to the public discussion on how the phenomenon of big data will shape many aspects of our human existence in the years to come. Yet, at the same time I had a sense that the authors at times succumbed to throwing ‘big data’ around as a vague and hyped-up buzzword and overlooked some important social implications of big data fetishization.

Take Google Flu Trends, for example. The authors tout the success of Google’s algorithms’ predictive power – pure statistical power that did not rely on solid theory – in using a plethora of search terms to predict flu outbreaks. However, only four years later Google Flu Trends failed spectacularly, as the theory-free, data-rich model overstated the figures of flu-like illness by 140 percent (see this WIRED article on the lessons of Google Flu Trends failure).

This story, based on the belief that algorithms coupled with a massive amounts of data points will save the day and ‘revolutionize’ our world, is just one example of what some critics have termed “big data hubris”. This is of course not to say that big data analyses are useless. However, one of the book’s themes, that we ought to dispense with reliance on causation in favor of correlation, carries some problematic implications. The authors write that for many things “knowing what not why is good enough”. This runs the risk of disregarding the reality that data does not, by itself, create meaning. We do.

Let us use the example of language. Google and many other companies have started mining books, magazines, blogs, Twitter, etc. for patterns on how humans use different languages. We can visualize the data using fancy graphs or word clouds, but we have to remember it is humans who create language, and algorithms alone are not sufficient to understand the complexities and nuances of how we choose to use, interpret and translate different languages. Google Translate is leaps better than the early translation software I remember from my childhood, but I agree with those commentators who remain doubtful whether machine translation will ever surpass the very human skills that professional translators offer.

Correlation does not equal causation. Big Data: A Revolution…spends a whole chapter showing how correlations within big data are now the hard and fast rule in retail, health, and government. I feel like this chapter was missing the rule of “correlation doesn’t always equal causation”, often pushed in the sciences. Big data is often pushed as the only way to gather information, and the book even has pushed so far as to say the scientific method and hypotheses are no longer strong forms of research. It bothers me that Mayer-Schonberger is making the case that predictability and probability will replace structured methods such as methodical experiments. Not everything needs to be rushed and “fast thinking.” I can understand Walmart’s use of correlations between hurricanes and selling more Pop-tarts, but once my health insurance starts rising because they think I watch Netflix for hours, then I have something to be worried about.

This week’s reading was a little bit of a circus. On the one hand, I’m not convinced that THE WEALTH OF NETWORKS was even written by a human author because the same words seemed to repeat infinitely. On the other hand, BLOCKCHAIN REVOLUTION felt like taking a tour of a chocolate factory hologram guided by two Willy Wonkas. Yochai Benkler writes unbelievably complicated sentences, while Alex Tapscott and Don Tapscott offer a vision of our future that is totally made out of candy. 

 Both books are fundamentally about building trust. Both argue that comparing the blockchain future to perfection isn’t the point; instead, we’re supposed to compare the blockchain future to our present-day broken “Franken-finance” system, which we can’t trust anyway. The Tapscotts reference an interesting Gallup poll that points out that there’s nothing we trust less than business and Congress, so I guess they’re trying to make us feel things can only get better.

It’s hard to imagine how blockchain infrastructures will suddenly cause companies to become transparent and personal data to become private, as if we’re being handed an immensely powerful pair of glasses that correct everyone’s vision at once. I love the reference to the Economist’s reporting, which describes blockchain technology as “the trust machine.” Benkler and both Tapscotts desperately want to convince us that blockchain technology produces unimaginably high levels of trust among global strangers because, put simply, “the ledger itself is the foundation of trust.” (?) More than anything, my takeaway is that these books might have been written about any emerging technology in any time, urging us to consider electricity, ride an airplane, or try chemotherapy. They sound insane, like everyone always does. They’re describing an economy based on simplified payment verification (SPV), which seems to require nothing in order to participate. I guess I feel impressed by everything that blockchain might make possible but also disappointed that these authors sound like Dr. Oz trying to sell you on magic vitamins. Why is it so hard to sound honest? 

I love this. It’s exactly what these pro-blockchainers want us to think. Not that i’m a part of some conspiracy theory movement against blockchain, but it’s easier to say that something is going to fix all our problems and create an easier process for something. Chances are whether blockchain is accepted in the financial industry or not, it’s never going to directly affect me. And you’re right, this is inherently a very trust-driven innovation. If something happens, even though it sounds like it won’t by the explanation of the technology behind it, the entire system goes down (like the left-pad debacle earlier this year). 

Bitcoin is a technology with possible game changing applications for the developing world, but is the developing world ready for it? The idea of developing countries in leapfrogging technologies is so “hot” these days, but is what is the evidence to back it up? There is some truth to this idea: most Africans never had a landline and went right to cell phones, but will developing countries Africa really leapfrog over the world and adopt Bitcoin first? To try and answer this question, I’m going to discuss the money environment in Africa and talk about existing financial systems.

Proponents of the leapfrog theory of development often say the developing world’s lack of old technological infrastructure leaves the market wide open for new technologies. This appears to be true for the financial industry. In the continent as a whole, there is a huge pent-up demand for banking services with only 66% of the Africans having access to financial services. Banks have a hard time reaching people in rural areas and people without a lot of capital. One can see the potential for digital financial services to take off in Africa and reach these customers by looking at the success of mobile money on the continent.  

Mobile money is a technology that has taken off in Africa and but not in the developed world. Mobile money is a technology that allows users to deposit or receive money in kiosks and use their phone account as a bank account. It has done well in Africa because there was a demand and all the new technology (mobile phones) were in place. This method lowers startup costs for banks to reach rural people and takes advantage of telecoms existing infrastructure to reach out to new customers. Mobile money is also used for people to have a safe place to store their money and for sending remittances. Several of these reasons for using mobile money apply to Bitcoin so why wouldn’t Africa just leapfrog again and move from mobile money to bitcoin?  

Some of the reasons for adopting mobile money are similar to those for adopting bitcoin but there are several major disadvantages to using bitcoin: lack of political capital, high fluctuation in price, and general trust issues. It took years of political negotiation in countries to get create the mobile money policy structure to allow mobile money businesses to flourish. Regulators created new policies around Know Your Customer (KYC) and Anti Money Laundering (AML) laws in order for the technology to take off. This kind of steady policy environment was what businesses needed to feel secure to make a large investment in countries. Considering the amount of energy it took to secure this political capital, it seems unlikely that Bitcoin will be able to do so in a few years, especially with the entrenched interests in mobile money and banking putting their weight against keeping them out of the market. Also many of the anonymity features of Bitcoin don’t lend themselves well to KYC and AML regulations.  

Secondly, companies that have gained the trust of poor and rural customers had to educate them on how mobile money technology works and the barriers to teaching people about bitcoin technology are much higher. I have a hard time understanding how it works and can imagine how hard it would be to explain to an uneducated rural farmer. In markets that rely on trust and can be reticent to change their saving strategy, Bitcoin can seem like a risky bet. Poor people are unlikely to make risky financial decisions. The fluctuation of Bitcoin prices makes the decision an even riskier bet for poor, rural people. There is a theory that if more people use Bitcoin, then there will be more people in the market and so the fluctuations will be smoothed out. There are more people buying  into the currency every day and so this theory might just turn out to be true.

I hate to be a grumpy, skeptical old man, but since I turned 30 recently, I’m feeling more and more comfortable telling kids to get out of my yards and bursting people’s idealistic tech bubbles. I’ll start here with Steven Johnson’s book Emergence: The Connected Lives of Ants, Brains, Cities, and Software. Let me say first, that I liked this book. Who doesn’t like learning that ants are leaderless societies that have complex organization and even create their own trash dumps? And did you know that slime molds can be taught to move through a maze and don’t have a brain or central decision-making cell? However, Johnson’s obsession of networks doesn’t take into account how human bias makes network systems a poor solution for many of humanities problems.  I still love these ideas and think they could be applicable to certain parts of human society, but I’m not sure how helpful these lessons are for a diverse society that needs to complete tasks far more complicated than moving through mazes.

The book is about how small decisions at a lower level of society create higher level order and efficiency.  While Johnson’s examples are interesting and could hold lessons for planners, networks that select for one thing inevitably leave something out. Any system still has to choose what it will select for and often does this in a Darwinian manner. But when we want to create societies that are just, fair, and equal, the Darwinian method and the networking method begins to fall apart. A great counter-example can be seen in segregation.

Thomas Schelling’s 1972 paper Dynamic Models of Segregation uses game theory to prove that people’s small preferences to live with people of a similar race, lead to large segregation effects. These are small decisions made by many people that create a higher level order. This sounds awfully familiar, right? It is almost exactly the emergence network effect. But, it leads to an outcome that sorts people into homogenous groups which is exactly what we don’t want in our cities.

This idea can be modeled and visualized easily using game theory. Don’t worry, the website uses adorable triangles and squares so the exercise isn’t as depressing as it might sound. The model allows the user to move the controls up and down for preferences and baseline segregation to see how people would triangles and squares sort themselves. It’s an incredibly effective learning tool to see the side effects of emerging networks.

This website and model teaches us that we can have diverse neighborhoods, but we will probably need to be more conscious about what emerge from our networks. Secondly, we need to be careful about calling people racist, because everyone has some kind of preference to be with a similar people in a similar group. If we are to build a society that we want we need to be conscious of these and try to create systems and neighborhoods that go against some of our inherent bias. And this will not happen based on a network choosing what it wants.

Johnson mentions how networks can be biased towards majority viewpoints in an example he gives about a news aggregator website called Slash Dot. You may have heard about similar problems with Facebook’s newsfeed that favors cats and things you will ‘like’ over more difficult news. These websites tend to focus on majority viewpoints because they get more likes and positive feedback. These kinds of networks again, are great for one group, the majority, but fail to take into account all of the other things we want in our lives, like a diverse set of viewpoints.

Johnson’s insight into the generative capacity of emergent behavior was a foreshadow to the work of Yochai Benkler and other Internet scholars who would spend the rest of the 2000s investigating the social, legal and economic implications of the internet. Benkler’s seminal work, The Wealth of Networks, led to the idea of the networked information economy, in which commons-based peer-production enabled anyone with a computer and the right software to, of their own volition, contribute to a project or participate in a web-based society. The motivations behind participating in these new modalities of interaction are the subject of much inquiry in Emergence. And it is found that social norms often play a much greater role in shaping street level collective action than any centralized authority. The inherent architecture of the Internet facilitates heterarchical programs and systems, and Johnson’s “myth of the ant queen: the assumption that collective behavior implied some kind of centralized authority,” demonstrates the potential for artificial emergence occur with little to no authoritative direction (pg. 168).

Wikipedia, for example, is an ant-hill colony of emergent behavior where individuals act in accordance with low-level rules that shape macro-level behavior (pg. 181). The creator of Wikipedia, Jimmy Wales, exerts his authority as the “leader” of Wikipedia in the form of shaping the values the community will uphold. Much like the ant queen supplies the gene pool from which all other ants develop, creators of platforms which utilize emergent behavior really only set the ground rules for that behavior, and those mores are enforced by street-level users who believe in them. Other examples of these types of leaders include Tim Berners-Lee, the inventor of the World Wide Web, and Peter Sunde, the creator of the Pirate Bay. Both leaders have specific agendas entrenched into the very foundations of their respective enterprises, and individuals who participate in these systems are profoundly affected by the virtues and practices embodied in them.