Nature

You are currently browsing the archive for the Nature category.

fruit thought

If personal data is actually a commodity, can you buy some from another person, as if that person were a fruit stand? Would you want to?

Well, no.

Yet there is lately a widespread urge to claim personal data as personal property, and to create commodity markets for personal data, so people can start making money by selling or otherwise monetizing their own.

ProjectVRM, which I direct, is chartered to “foster development of tools and services that make customers both independent and better able to engage,” and is a big tent. That’s why on the VRM Developments Work page of its wiki a heading called Markets for Personal Data. Listed there are:

So: respect.

Yet, while I salute these efforts’ respect for individuals, and their righteous urges to right the wrongs of wanton and rude harvesting of personal data from approximately everybody, I also think there are problems with this approach. And, since I’ve been asked lately to spell out those problems, I shall. Here goes.

The first problem is that, economically speaking, data is a public good, meaning non-rivalrous and non-excludable. Here’s a table that may help (borrowed from this Linux Journal column):

Excludability Excludability
YES NO
Rivalness YES Private good: good: e.g., food, clothing, toys, cars, products subject to value-adds between first sources and final customers Common pool resource: e.g., sea, rivers, forests, their edible inhabitants and other useful contents
Rivalness NO Club good: e.g., bridges, cable TV, private golf courses, controlled access to copyrighted works public good: e.g., data, information, law enforcement, national defense, fire fighting, public roads, street lighting

 

The second problem is that nature of data as a public good also inconveniences claims that it ought to be property. Thomas Jefferson explained this in his 1813 letter to Isaac MacPherson:

If nature has made any one thing less susceptible than all others of exclusive property, it is the action of the thinking power called an idea, which an individual may exclusively possess as long as he keeps it to himself; but the moment it is divulged, it forces itself into the possession of every one, and the receiver cannot dispossess himself of it. Its peculiar character, too, is that no one possesses the less, because every other possesses the whole of it. He who receives an idea from me, receives instruction himself without lessening mine; as he who lights his taper at mine, receives light without darkening me. That ideas should freely spread from one to another over the globe, for the moral and mutual instruction of man, and improvement of his condition, seems to have been peculiarly and benevolently designed by nature, when she made them, like fire, expansible over all space, without lessening their density in any point, and like the air in which we breathe, move, and have our physical being, incapable of confinement or exclusive appropriation

Of course Jefferson never heard of data. But what he says about “the thinking power called an idea,” and how ideas are like fire, is essential in a very human way.

The third problem is that all of us as human beings are able to produce forms of value that far exceed that of our raw personal data.

Specifically, treating data as if it were a rivalrous and excludable commodity—such as corn, oil or fruit—not only takes Jefferson’s “thinking power” off the table, but misdirects attention, investment and development work away from supporting the human outputs that are fully combustible, and might be expansible over all space, without lessening density. Ideas can do that. Oil can’t, even though it’s combustible.

Put another way, why would you want to make almost nothing (the likely price) selling personal data on a commodity basis when you can make a lot more by selling your work where markets for work exist?

What makes us fully powerful as human beings is our ability to generate and share ideas and other combustible public goods, and not just to slough off data like so much dandruff. Or to be valued only for the labors we contribute as parts of industrial machines.

Important note: I’m not knocking labor here. Most of us have to work for wages as parts of industrial machines, or as independent actors. I do too. There is full honor in that. Yet our nature as distinctive and valuable human beings is to be more and other than a source of labor alone, and there are ways to make money from that fact too.

Many years ago JP Rangaswami (@jobsworth) and I made a distinction between making money with something and because of something. It’s a helpful one.

Example: I don’t make money with this blog. But I do make money because of it—and probably a lot more money than I would if this blog carried advertising or if I did it for a wage.

Which gets us to the idea behind declaring personal data as personal property, and creating marketplaces where people can sell their data.

The idea goes like this: there is a $trillion or more in business activity that trades or relies on personal data in many ways. Individual sources of that data should be able to get in on the action.

Alas, most of that $trillion is in what Shoshana Zuboff calls surveillance capitalism: a giant snake-ball of B2B activity wherein there is little interest in buying what can be had for free.

Worse, surveillance capitalism’s business is making guesses about you so it can sell you shit. On a per-message basis, this works about 0% of the time, even though massive amounts of money flow through that B2B snakeball (visualized as abstract rectangles here and here). Many reasons for that. Here are a few:

  1. Most of the time, such as right here and now, you’re not buying a damn thing, and not in a mood to be bothered by someone telling you what to buy.
  2. Companies paying other companies to push shit at you do not have your interests at heart—not even if their messages to you are, as they like to put it, “relevant” or “interest based.” (Which they almost always are not.)
  3. The entrails of surveillance capitalism are fully infected with fraud and malware.
  4. Surveillance capitalism is also quite satisfied to soak up to 97% of an advertising spend before an ad’s publisher gets its 3% for pushing an ad at you.

Trying to get in on that business is just an awful proposition.

Yes, I know it isn’t just surveillance capitalists who hunger for personal data. The health care business, for example, can benefit enormously from it, and is less of a snakeball, on the whole. But what will it pay you? And why should it pay you?

Won’t large quantities of anonymized personal data from iOS and Android devices, handed over freely, be more valuable to medicine and pharma than the few bits of data individuals might sell? (Apple has already ventured in that direction, very carefully, also while not paying for any personal data.)

And isn’t there something kinda suspect about personal data for sale? Such as motivating the unscrupulous to alter some of their data so it’s worth more?

What fully matters for people in the digital world is agency, not data. Agency is the power to act with full effect in the world. It’s what you have when you put your pants on, when you walk, or drive, or tell somebody something useful while they listen respectfully. It’s what you get when you make a deal with an equal.

It’s not what any of us get when we’re just “users” on a platform. Or when we click “agree” to one-sided terms the other party can change and we can’t. Both of those are norms in Web 2.0 and desperately need to be killed.

It’s still early. Web 2.0 is an archaic stage in the formation of the digital world. surveillance capitalism has also been a bubble ready to pop for years. The matter is when, not if. It’s too absurd, corrupt, complex and annoying to keep living forever.

So let’s give people ways to increase their agency, at scale, in the digital world. There’s no scale in selling one’s personal data. But there’s plenty in putting our most human of powers to work.

The most basic form of agency in the digital world is control over how our personal data might be used by others. There are lots of developers at work on this already. Here’s one list at ProjectVRM.

Bonus links:

 

 

 

 

Just before it started, the geology meeting at the Santa Barbara Central Library on Thursday looked like this from the front of the room (where I also tweeted the same pano):

Geologist Ed Keller

Our speakers were geology professor Ed Keller of UCSB and Engineering Geologist Larry Gurrola, who also works and studies with Ed. That’s Ed in the shot below.

As a geology freak, I know how easily terms like “debris flow,” “fanglomerate” and “alluvial fan” can clear a room. But this gig was SRO. That’s because around 3:15 in the morning of January 9th, debris flowed out of canyons and deposited fresh fanglomerate across the alluvial fan that comprises most of Montecito, destroying (by my count on the map below) 178 buildings, damaging more than twice that many, and killing 23 people. Two of those—a 2 year old girl and a 17 year old boy—are still interred in the fresh fanglomerate and sought by cadaver dogs. The whole thing is beyond sad and awful.

The town was evacuated after the disaster so rescue and recovery work could proceed without interference, and infrastructure could be found and repaired: a job that required removing twenty thousand truckloads of mud and rocks. That work continues while evacuation orders are gradually lifted, allowing the town to repopulate itself to the very limited degree it can.

I talked today with a friend whose business is cleaning houses. Besides grieving the dead, some of whom were friends or customers, she reports that the cleaning work is some of the worst she has ever seen, even in homes that were spared the mud and rocks. Refrigerators and freezers, sitting closed and without electricity for weeks, reek of death and rot. Other customers won’t be back because their houses are gone.

Highway 101, one of just two freeways connecting Northern and Southern California, runs through town near the coast and more than two miles from the mountain front. Three debris flows converged on the highway and used it as a catch basin, filling its deep parts to the height of at least one bridge before spilling over its far side and continuing to the edge of the sea. It took two weeks of constant excavation and repair work before traffic could move again. Most exits remain closed. Coast Village Road, Montecito’s Main Street, is open for employees of stores there, but little is open for customers yet, since infrastructural graces such as water are not fully restored. (I saw the Honor Bar operating with its own water tank, and a water truck nearby.) Opening Upper Village will take longer. Some landmark institutions, such as San Ysidro Ranch and La Casa Santa Maria, will take years to restore. (From what I gather, San Ysidro Ranch, arguably the nicest hotel in the world, was nearly destroyed. Its website thank firefighters for salvation from the Thomas Fire. But nothing, I gather, could have save it from the huge debris flow wiped out nearly everything on the flanks of San Ysidro Creek. (All the top red dots along San Ysidro Creek in the map below mark lost buildings at the Ranch.)

Here is a map with final damage assessments. I’ve augmented it with labels for the canyons and creeks (with one exception: a parallel creek west of Toro Canyon Creek):

Click on the map for a closer view, or click here to view the original. On that one you can click on every dot and read details about it.

I should pause to note that Montecito is no ordinary town. Demographically, it’s Beverly Hills draped over a prettier landscape and attractive to people who would rather not live in Beverly Hills. (In fact the number of notable persons Wikipedia lists for Montecito outnumbers those it lists for Beverly Hills by a score of 77 to 71.) Culturally, it’s a village. Last Monday in The New Yorker, one of those notable villagers, T.Coraghessan Boyle, unpacked some other differences:

I moved here twenty-five years ago, attracted by the natural beauty and semirural ambience, the short walk to the beach and the Lower Village, and the enveloping views of the Santa Ynez Mountains, which rise abruptly from the coastal plain to hold the community in a stony embrace. We have no sidewalks here, if you except the business districts of the Upper and Lower Villages—if we want sidewalks, we can take the five-minute drive into Santa Barbara or, more ambitiously, fight traffic all the way down the coast to Los Angeles. But we don’t want sidewalks. We want nature, we want dirt, trees, flowers, the chaparral that did its best to green the slopes and declivities of the mountains until last month, when the biggest wildfire in California history reduced it all to ash.

Fire is a prerequisite for debris flows, our geologists explained. So is unusually heavy rain in a steep mountain watershed. There are five named canyons, each its own watershed, above Montecito, as we see on the map above. There are more to the east, above Summerland and Carpinteria, the next two towns down the coast. Those towns also took some damage, though less than Montecito.

Ed Keller put up this slide to explain conditions that trigger debris flows, and how they work:

Ed and Larry were emphatic about this: debris flows are not landslides, nor do many start that way (though one did in Rattlesnake Canyon 1100 years ago). They are also not mudslides, so we should stop calling them that. (Though we won’t.)

Debris flows require sloped soils left bare and hydrophobic—resistant to water—after a recent wildfire has burned off the chaparral that normally (as geologists say) “hairs over” the landscape. For a good look at what soil surfaces look like, and are likely to respond to rain, look at the smooth slopes on the uphill side of 101 east of La Conchita. Notice how the surface is not only a smooth brown or gray, but has a crust on it. In a way, the soil surface has turned to glass. That’s why water runs off of it so rapidly.

Wildfires are common, and chaparral is adapted to them, becoming fuel for the next fire as it regenerates and matures. But rainfalls as intense as this one are not common. In just five minutes alone, more than half an inch of rain fell in the steep and funnel-like watersheds above Montecito. This happens about once every few hundred years, or about as often as a tsunami.

It’s hard to generalize about the combination of factors required, but Ed has worked hard to do that, and this slide of his is one way of illustrating how debris flows happen eventually in places like Montecito and Santa Barbara:

From bottom to top, here’s what it says:

  1. Fires happen almost regularly, spreading most widely where chaparral has matured to become abundant fuel, as the firefighters like to call it.
  2. Flood events are more random, given the relative rarity of rain and even more rare rains of “biblical” volume. But they do happen.
  3. Stream beds in the floors of canyons accumulate rocks and boulders that roll down the gradually eroding slopes over time. The depth of these is expressed as basin instablity. Debris flows clear out the rocks and boulders when a big flood event comes right after a fire and basin becomes stable (relatively rock-free) again.
  4. The sediment yield in a flood (F) is maximum when a debris flow (DF) occurs.
  5. Debris flows tend to happen once every few hundred years. And you’re not going to get the big ones if you don’t have the canyon stream bed full of rocks and boulders.

About this set of debris flows in particular:

  1. Destruction down Oak Creek wasn’t as bad as on Montecito, San Ysidro, Buena Vista and Romero Creeks because the canyon feeding it is smaller.
  2. When debris flows hit an obstruction, such as a bridge, they seek out a new bed to flow on. This is one of the actions that creates an alluvial fan. From the map it appears something like that happened—
    1. Where the flow widened when it hit Olive Mill Road, fanning east of Olive Mill to destroy all three blocks between Olive Mill and Santa Elena Lane before taking the Olive Mill bridge across 101 and down to the Biltmore while also helping other flows fill 101 as well. (See Mac’s comment below, and his link to a top map.)
    2. In the area between Buena Vista Creek and its East Fork, which come off different watersheds
    3. Where a debris flow forked south of Mountain Drive after destroying San Ysidro Ranch, continuing down both Randall and El Bosque Roads.

For those who caught (or are about to catch) Ellen’s Facetime with Oprah visiting neighbors, that happened among the red dots at the bottom end of the upper destruction area along San Ysidro Creek, just south of East Valley Road. Oprah’s own place is in the green area beside it on the left, looking a bit like Versailles. (Credit where due, though: Oprah’s was a good and compassionate report.)

Big question: did these debris flows clear out the canyon floors? We (meaning our geologists, sedimentologists, hydrologists and other specialists) won’t know until they trek back into the canyons to see how it all looks. Meanwhile, we do have clues. For example, here are after-and-before photos of Montecito, shot from space. And here is my close-up of the latter, shot one day after the event, when everything was still bare streambeds in the mountains and fresh muck in town:

See the white lines fanning back into the mountains through the canyons (Cold Spring, San Ysidro, Romero, Toro) above Montecito? Ed explained that these appear to be the washed out beds of creeks feeding into those canyons. Here is his slide showing Cold Spring Creek before and after the event:

Looking back at Ed’s basin threshold graphic above, one might say that there isn’t much sediment left for stream beds to yield, and that those in the floors of the canyons have returned to stability, meaning there’s little debris left to flow.

But that photo was of just one spot. There are many miles of creek beds to examine back in those canyons.

Still, one might hope that Montecito has now had its required 200-year event, and a couple more centuries will pass before we have another one.

Ed and Larry caution against such conclusions, emphasizing that most of Montecito’s and Santa Barbara’s inhabited parts gain their existence, beauty or both by grace of debris flows. If your property features boulders, Ed said, a debris flow put them there, and did that not long ago in geologic time.

For an example of boulders as landscape features, here are some we quarried out of our yard more than a decade ago, when we were building a house dug into a hillside:

This is deep in the heart of Santa Barbara.

The matrix mud we now call soil here is likely a mix of Juncal and Cozy Dell shale, Ed explained. Both are poorly lithified silt and erode easily. The boulders are a mix of Matilija and Coldwater sandstone, which comprise the hardest and most vertical parts of the Santa Ynez mountains. The two are so similar that only a trained eye can tell them apart.

All four of those geological formations were established long after dinosaurs vanished. All also accumulated originally as sediments, mostly on ocean floors, probably not far from the equator.

To illustrate one chapter in the story of how those rocks and sediments got here, UCSB has a terrific animation of how the transverse (east-west) Santa Ynez Mountains came to be where they are. Here are three frames in that movie:

What it shows is how, when the Pacific Plate was grinding its way northwest about eighteen million years ago, a hunk of that plate about a hundred miles long and the shape of a bread loaf broke off. At the top end was the future Malibu hills and at the bottom end was the future Point Conception, then situated south of what’s now Tijuana. The future Santa Barbara was west of the future Newport Beach. Then, when the Malibu end of this loaf got jammed at the future Los Angeles, the bottom end of the loaf swept out, clockwise and intact. At the start it was pointing at 5 o’clock and at the end (which isn’t), it pointed at 9:00. This was, and remains, a sideshow off the main event: the continuing crash of the Pacific Plate and the North American one.

Here is an image that helps, from that same link:

Find more geology, with lots of links, in Making sense of what happened to Montecito. I put that post up on the 15th and have been updating it since then. It’s the most popular post in the history of this blog, which I started in 2007. There are also 58 comments, so far.

I’ll be adding more to this post after I visit as much as I can of Montecito (exclusion zones permitting). Meanwhile, I hope this proves useful. Again, corrections and improvements are invited.

30 January

 

The term “fake news” was a casual phrase until it became clear to news media that a flood of it had been deployed during last year’s presidential election in the U.S. Starting in November 2016, fake news was the subject of strong and well-researched coverage by NPR (here and here), Buzzfeed, CBS (here and here), Wired, the BBC, Snopes, CNN (here and here), Rolling Stone and others. It thus became a thing…

… until Donald Trump started using it as an epithet for news media he didn’t like. He did that first during a press conference on February 16, and then the next day on Twitter:

And he hasn’t stopped. To Trump, any stick he can whup non-Fox mainstream media with is a good stick, and FAKE NEWS is the best.

So that pretty much took “fake news,” as a literal thing, off the table for everyone other than Trump and his amen chorus.

So, since we need a substitute, I suggest decoy news. Because that’s what we’re talking about: fabricated news meant to look like the real thing.

But the problem is bigger than news alone, because advertising-funded media have been in the decoy business since forever. The difference in today’s digital world is that it’s a lot easier to fabricate a decoy story than to research and produce a real one—and it pays just as well, or even better, because overhead in the decoy business rounds to nothing. Why hire a person to do an algorithm’s job?

In the content business the commercial Web has become, algorithms are now used to target both stories and the advertising that pays for them.

This is why, on what we used to call the editorial side of publishing (interesting trend here), journalism as a purpose has been displaced by content production.

We can see one tragic result in a New York Times story titled In New Jersey, Only a Few Media Watchdogs Are Left, by David Chen (@davidwchen). In it he reports that “The Star-Ledger, which almost halved its newsroom eight years ago, has mutated into a digital media company requiring most reporters to reach an ever-increasing quota of page views as part of their compensation.”

This calls to mind how Saturday Night Live in 1977 introduced the Blues Brothers in a skit where Paul Shaffer, playing rock impresario Don Kirshner, proudly said the Brothers were “no longer an authentic blues act, but have managed to become a viable commercial product.”

To be viably commercial today, all media need to be in the content production business, paid for by adtech, which is entirely driven by algorithms informed by surveillance-gathered personal data. The result looks like this:

To fully grok how we got here, it is essential to understand the difference between advertising and direct marketing, and how nearly all of online advertising is now the latter. I describe the shift from former to latter in Separating Advertising’s Wheat and Chaff:

Advertising used to be simple. You knew what it was, and where it came from.

Whether it was an ad you heard on the radio, saw in a magazine or spotted on a billboard, you knew it came straight from the advertiser through that medium. The only intermediary was an advertising agency, if the advertiser bothered with one.

Advertising also wasn’t personal. Two reasons for that.

First, it couldn’t be. A billboard was for everybody who drove past it. A TV ad was for everybody watching the show. Yes, there was targeting, but it was always to populations, not to individuals.

Second, the whole idea behind advertising was to send one message to lots of people, whether or not the people seeing or hearing the ad would ever use the product. The fact that lots of sports-watchers don’t drink beer or drive trucks was beside the point, which was making brands sponsoring a game familiar to everybody watching it.

In their landmark study, “The Waste in Advertising is the Part that Works” (Journal of Advertising Research, December, 2004, pp. 375–390), Tim Ambler and E. Ann Hollier say brand advertising does more than signal a product message; it also gives evidence that the parent company has worth and substance, because it can afford to spend the money. Thus branding is about sending a strong economic signal along with a strong creative one.

Plain old brand advertising also paid for the media we enjoyed. Still does, in fact. And much more. Without brand advertising, pro sports stars wouldn’t be getting eight and nine figure contracts.

But advertising today is also digital. That fact makes advertising much more data-driven, tracking-based and personal. Nearly all the buzz and science in advertising today flies around the data-driven, tracking-based stuff generally called adtech. This form of digital advertising has turned into a massive industry, driven by an assumption that the best advertising is also the most targeted, the most real-time, the most data-driven, the most personal — and that old-fashioned brand advertising is hopelessly retro.

In terms of actual value to the marketplace, however, the old-fashioned stuff is wheat and the new-fashioned stuff is chaff. In fact, the chaff was only grafted on recently.

See, adtech did not spring from the loins of Madison Avenue. Instead its direct ancestor is what’s called direct response marketing. Before that, it was called direct mail, or junk mail. In metrics, methods and manners, it is little different from its closest relative, spam.

Direct response marketing has always wanted to get personal, has always been data-driven, has never attracted the creative talent for which Madison Avenue has been rightly famous. Look up best ads of all time and you’ll find nothing but wheat. No direct response or adtech postings, mailings or ad placements on phones or websites.

Yes, brand advertising has always been data-driven too, but the data that mattered was how many people were exposed to an ad, not how many clicked on one — or whether you, personally, did anything.

And yes, a lot of brand advertising is annoying. But at least we know it pays for the TV programs we watch and the publications we read. Wheat-producing advertisers are called “sponsors” for a reason.

So how did direct response marketing get to be called advertising ? By looking the same. Online it’s hard to tell the difference between a wheat ad and a chaff one.

Remember the movie “Invasion of the Body Snatchers?” (Or the remake by the same name?) Same thing here. Madison Avenue fell asleep, direct response marketing ate its brain, and it woke up as an alien replica of itself.

It is now an article of faith within today’s brain-snatched advertising business that the best ad is the most targeted and personalized ad. Worse, almost all the journalists covering the advertising business assume the same thing. And why wouldn’t they, given that this is how advertising is now done online, especially by the Facebook-Google duopoly.

And here is why those two platforms can’t fix it: both have AI machines built to give millions of advertising customers ways to target the well-studied eyeballs of billions of people, using countless characterizations of those eyeballs.In fact, the only (and highly ironic) way they can police bad acting on their platforms is by hiring people who do nothing but look for that bad acting.

One fix is regulation. We now have that, hugely, with the General Data Protection Regulation (GDPR). It’s an EU law, but it protects the privacy of EU citizens everywhere—with potentially massive fines. In spirit, if not also in letter (which the platforms are struggling mightily to weasel around), the GDPR outlaws tracking people like tagged animals online. I’ve called the GDPR an extinction event for adtech, and the main reason brands (including the media kind) need to fire it.

The other main fixes begin on the personal side. Don Marti (@dmarti) tweets, “Build technologies to implement people’s norms on sharing their personal data, and you’ll get technologies to help high-reputation sites build ad-supported business models ABSOLUTELY FREE!” Those models are all advertising wheat, not adtech chaff.

Now here’s the key: what we need most are single and simple ways for each of us to manage all our dealings with other entities online. Having separate means, each provided by the dozens or hundreds of sites and services we each deal with, all with different UIs, login/password gauntlets, forms to fill out, meaningless privacy policies and for-lawyers-only terms of service, cannot work. All that shit may give those companies scale across many consumers, but every one of them only adds to those consumers’ relationship overhead. I explain how this will work in Giving Customers Scale, plus many other posts, columns and essays compiled in my People vs. Adtech series, which is on its way to becoming a book. I’d say more about all of it, but need to catch a plane. See you on the other coast.

Meanwhile, the least we can do is start talking about decoy news and the business model that pays for it.

[Later…] I’m on the other coast now, but preoccupied by the #ThomasFire threatening our home in Santa Barbara. Since this blog appears to be mostly down, I’m writing about it at doc.blog.

 

 

 

Tags: , ,