Who Owns the Future? (16 page)

If you rooted a Bop, something strange happened. A gigantic firm called Booty arose based on building a proprietary database of what was going on in the bodies of people with rooted Bops. How did Booty get the data? Millions of people used pirate bopper sites that Booty could “scrape.” Millions more voluntarily accepted contracts they had never read on a social chemical networking site in order to get access to free formulas. In doing so, they opened their bodies to Booty’s competitor Bodybook.

Booty usually made money not by directly charging Boppers money, but instead by taking money from third parties in exchange for being able to influence what a Bopper would be exposed to online. For instance, a Bopper in excellent physical health might get an offer for a free fill-up at an army recruiting station. It turned out that this form of indirect manipulation worked well enough to earn Booty many billions of dollars.

Booty, Bodybook, and VitaBop coexisted awkwardly. Each collected a vast dossier on the metabolisms of everyone, but none could peer into the others’ data vaults. Booty hoarded the treasure of the
rooted, open world, while VitaBop did the same for the world of subscribers, and Bodybook for a world of “sharers.” Booty accused VitaBop of being closed and not supporting the public good of bio-openness. VitaBop accused Booty of violating people’s privacy and dignity. Pundits would say that VitaBop was a little like Apple, while Booty was a little more like Google or a hedge fund, and Bodybook was like, well, guess.

What they had in common was that each was shrinking the economy and the job prospects of everyone.

CHAPTER 9

From Above

Misusing Big Data to Become Ridiculous

Three Nerds Walk into a Bar . . .

Your always-amused author once served on a panel at UC Berkeley judging mock business plans submitted by engineering graduate students who had enrolled in an entrepreneurship program. Three students presented the following scheme:

Suppose you’re darting around San Francisco bars and hot spots on a Saturday night. You land in a bar and there are a bounteous number of seemingly accessible, lovely, and unattached young women hanging out looking for attention in this particular place. Well, you whip out your mobile phone and alert the network. “Here’s where the girls are!” All those other young men like you will know where to go. The service will make money with advertising, probably from bars and liquor concerns.

I looked at this geeky, sincere trio, and asked the obvious question: Will there ever, ever, ever be even the slightest chance that this service will provide even one bit of correct data? There was a tense pause. Was this yet another Asperger’s syndrome–like example of incredible technical intelligence coupled with appalling naïveté about people?

Their answer: No, of course not. There will never be good data. The whole scheme will run on hope.

I gave them the most favorable possible evaluation, not because I wanted to encourage them to apply their hard-won skills to such an unproductive plan, but because they demonstrated an understanding of how networked information really works, when it comes to people.
*

*
Silicon Valley comes through like clockwork. A year after this passage was written, a “where the babes are” app debuted at San Francisco bars. “SceneTap” used cameras and machine vision instead of volunteered information. No, I did not look into whether this startup was founded by the same students.

Your Lack of Privacy Is Someone Else’s Wealth

Occasionally the rich embrace a new token and drive up its value. The fine art market is a great example. Expensive art is essentially a private form of currency traded among the very rich. The better an artist is at making art that can function this way, the more valuable the art will become. Andy Warhol is often associated with this trick, though Pablo Picasso and others were certainly playing the same game earlier. The art has to be stylistically distinct and available in suitable small runs. It becomes a private form of money, as instantly recognizable as a hundred-dollar bill.

A related trend of our times is that troves of dossiers on the private lives and inner beings of ordinary people, collected over digital networks, are packaged into a new private form of elite money. The actual data in these troves need not be valid. In fact, it might be better that it is not valid, for actual knowledge brings liabilities.

But the pretense that we have a bundle of other people’s secrets is functioning like fine modern art. It is a new kind of security that the rich trade in, and the value is naturally driven up. It becomes a giant-scale levee inaccessible to ordinary people.

Few people realize the degree to which they are being tracked and spied upon in order that this new form of currency can be created. There is an extensive literature
¹
and a sphere of activism
²
already in place to address the situation, so I will present only the briefest exposition of the topic here, with plenty of footnotes to follow.

Even an innocent visit to a legitimate major newspaper site like the
New York Times
invokes a competitive swarm of more than a dozen tracking services, each of which is attempting to become a dominant compiler of spy data about you. One plug-in that attempts to block spying schemes, called Ghostery, is currently blocking more than a
thousand
such schemes,
³
though no one knows the true number.

There is no definitive map of network spying services. The allegiances and roles are multifarious and complex.
⁴
No one really knows the score, though a common opinion is that Google
⁵
has historically been at the top of the heap for collecting spy data about you on the open Internet,
⁶
while Facebook has mastered a way to corral people under an exclusive microscope.
⁷
That said, other companies you’ve probably never heard of, like Acxiom
⁸
and eBureau,
⁹
are also deeply determined to create dossiers on you.

Because spying on you is, for the moment, the official primary business of the information economy, any attempt to avoid being spied on, such as the use of Ghostery,
¹⁰
can seem like an assault on the very idea of the Internet.
¹¹

Big Data in Science

The seeming magic of using data over a network has been applied differently in the worlds of science and business. The operations of both worlds are increasingly enacted using almost indistinguishable big data tools, but they play by different rules. In science, verification and accuracy are paramount. In business and the culture at large, not so much.

Scientists are using new technologies to observe previously murky layers of nature in detail for the first time, but there are so many details that it would be useless to even try without big computers and networks. Genomics is as much a branch of computer science as it is of biology, for instance. The same is true for the frontiers of materials science and energy.

In the sciences, the arrival of a fresh source of big data means a lot of hard work for researchers, no matter how much technology
is made available.
*
It is routine for new big data in medicine to transform our previous best guess about how to treat disease. And yet, new cures take years to arrive. In science, big data is magic, but
difficult
magic. We struggle with it, and expect to be fooled at first. The means to be rigorous with big data are still evolving.

*
For a while it looked like there was a statistical effect hiding in a giant sea of numbers showing neutrinos traveling faster than light. The compelling illusion survived a number of challenges until it was finally shot down months later.

No one in science thinks of big data as an automatic silver bullet. There is no shortage of common reference points to corroborate that assessment. Medicine provides the most consequential example. It is improving and yet improvement is tragically slow. Weather forecasting is better than it used to be, and is getting better. Satellites feed data we didn’t used to have into computer models that can handle the vast data volume, and the result is better guesses about next week’s weather, even next year’s overall weather. And yet, the weather still surprises. Big data gradually improves our abilities as we work with it, but it doesn’t instantly grant omniscience. Chasing a dynamic, ever-better-but-never-perfect statistic result is the very heart of modern cloud computing. Big data must be mastered in order to be valuable. It is not an automatic cornucopia, or a substitute for insight.

The spread of a flu outbreak can be tracked online faster than it can be tracked through the traditional medical system.
¹²
A research project at Google found that flu outbreaks could be tracked well by noting relevant searches in geographical zones. If there’s a sudden lift in concern about flu symptoms in a particular place, for instance, there is probably flu there. The signal is observable even before doctors receive the first wave of complaints.

Tracking the flu online is science. That means it isn’t automatic. Scientists must scrutinize the analysis. Maybe a rise in flu-related queries is actually in response to a popular movie in which the lead character has a bad flu. Without scrutiny, data isn’t trusted.

However, even in the world of big scientific data,
magical-seeming
results can come before the understanding. Big data can occasionally reverse the sequence and confuse the incentives that have driven science and commerce since the beginnings of each.

A spectacular recent example is the dawn of mind reading. In the first decade of the century there was a sequence of increasingly impressive examples of “brain reading.” This might involve a person learning to control a robotic arm through direct brain measurement. But would it be possible to measure what a person was seeing or imagining from reading the brain? That would be more properly described as “mind reading.”

Results started to appear early in the second decade of our century. Psychologist Jack Gallant and other researchers at UC Berkeley showed they could approximately determine what a person was watching simply by analyzing brain activity. It was as if computers became psychic, though a better way to understand the work is as an example of the challenges of scientific big data.

In Gallant’s experiment, a movie was computed of what someone was seeing, based on nothing but fMRI
*
scans of the activity of the person’s brain. The images looked blurry and otherworldly, but did conform to what was actually seen.

*
fMRI, or functional MRI, is a higher-power version of the familiar MRI scanner. fMRI is usually used to detect blood flow in the brain, which reveals which parts of the brain are most activated moment to moment.

The way it worked was approximately this: Each subject was shown a batch of movie clips. Their brain activation patterns were recorded each time. Then, when the person watched a new, previously unseen clip, activation patterns were once again recorded. Then the original clips were mixed into a new clip proportionally, according to how similar the activation pattern for the new clip was to each original clip. With enough previously seen clips mixed together, a fuzzy new clip emerges that does look like what the subject is watching.

This was a remarkable result, of great importance,
but
it was only the first step of scientific inquiry. It didn’t reveal how the brain codes visual memories. It did achieve something very important, which was that researchers had found a way to measure the brain that was relevant to specific visual cognition. Furthermore, similar techniques turn out to work for sound, speech, and other domains of experience and action. The age of high-tech mind reading has begun.

Jack Gallant is the first to point out that as spectacular as it is, the achievement is a beginning, not an end. The full cycle of scientific understanding will hopefully include additional attainments of insight and theory.

A Method in Waiting

You never know how long it will take for scientific conclusions about big data to form. Science gives up the best punch lines ever, but delivers them with the most inconsistent timing.

Big business data happens fast, as fast as people can take it in, or usually faster. Faster feedback loops make big business data ever more influential. We have become used to treating big business data as legitimate, even though it might really only seem so because of its special position in a network. Such data is valid by dint of tautology to an unknowable degree.

Science demands a different approach to big data, but we don’t know as much about that approach as we will soon. Scientific method for big data is not yet entirely codified. Once practices are established for big data science, there will be uncontroversial answers to questions like:

• What standard would have to be met to allow for the publication of replication of a result? To what degree must replication require the gathering of different, but similar big data, and not just the reuse of the same data with different algorithms?

• What is publication? Is it just a description of the code used? The code itself? The code in some standardized form or framework that makes it reusable and tweakable?

• Must analysis be performed in a way that anticipates standard practices of meta-analysis?