It’s deja vu all over again
One aspect of intentcasting that fascinates me is how recursive it is. VRM is the manifestation of a desire for intentcasting in an environment which lacks a framework for intentcasting. It needs to bootstrap itself in order to fully express itself. The whole project is so meta. This is why it is subtly amusing to talk seriously about the existence (or not) of demand that cannot possibly be measured in a vacuum. There’s a long list of things I anticipate will emerge after the VRM boot-up, all of which are predicated on the presence of something on which to build. These will in turn will give us the vantage point from which to imagine applications we can’t possibly predict or forecast demand for today.
I remember having an eerily parallel discussion with my manager at Equifax back in the early ’90s. You can’t do business on the Internet. Computers are just for playing games. No serious business would have a web presence. Consumers won’t want it.
Of course at the time you could have that discussion seriously due to the uncertainty of the platform. All those statements were true then and it wasn’t entirely clear that the Internet would be the globally disruptive force that it is today. At the time, I was the crazy one in that discussion and all I really wanted was online banking and electronic access to my own credit report. I didn’t know until much later that I wanted things like Wikipedia, Netflix streaming, and a WordPress blog that makes this post possible. In hindsight I don’t feel like the outlier. At the time I as the eccentric employee my manager was afraid to invite to meetings.
Now it’s 20 years later and I’m having the same discussions. You can’t do volume business if all your customers are peers. Privacy is dead and the Average Joe is OK with that. Average Joe has no interest in his transactional data and doesn’t know what he’d do with it if he got it.
There’s something I learned doing security that is broadly applicable and is generally overlooked in this argument. All systems are built on foundational assumptions. When you apply those architectures to new models, the foundational assumptions may break in non-obvious and potentially catastrophic ways.
Atoms to bits
Consider privacy in the digital world. Current attitudes about privacy inherit their legacy from a world of atoms when there was a significant cost associated with surveillance. This meant that the foundational surveillance technologies were all based on an assumption of targeted investigations. Moving from a world of atoms to a world of bits enabled surveillance at a resolution only dreamed of previously and at a per-person cost approaching zero. The baseline for surveillance technologies is no longer targeted investigation. Today’s strategy is simply to capture everything about everybody and mine the data as needed.
Our system of legal privacy protection is based on a world of atoms and embeds the implicit assumption that cost imposes a natural limit to the extent of surveillance. The move to bits breaks that foundational assumption in ways that are not obvious and potentially catastrophic. For example, in the 1983 United States v. Knotts case, the court found that “Nothing in the Fourth Amendment prohibited the police from augmenting their sensory faculties with such enhancement as science and technology afforded them in this case.” This ruling was later cited in October 2012 in the US v. Mendoza case as justification to allow warrantless installation of wireless video surveillance devices in exterior locations on private property.
The very same day in US v. Wahchumwah the 9th circuit court of appeals ruled that it was OK for undercover police who had been invited into a suspect’s home to secretly videotape during the visit without a warrant. Judge Smith wrote that “If the conduct and revelations of an agent operating without electronic equipment do not invade the defendant’s constitutionally justifiable expectations of privacy, then neither does a simultaneous recording of the same conversations made by the agent or by others from transmissions received from the agent to whom the defendant is talking and whose trustworthiness the defendant necessarily risks.”
But eyewitness testimony is notoriously unreliable whereas audio/video recordings are refutable as to integrity but not as to content. Unless you can prove the recording was altered, there is no defense. Whatever it shows is assumed to be reliable. In every detail. Forever. The move to bits enables reliability, granularity and durability of testimony that is orders of magnitude greater than eyewitness testimony, at an incremental cost approaching zero, and with no indicator that is detectable to the suspect. In dealing with this, new case law has a greater affinity for previous case law than it does for changes in the foundational assumptions on which the old case law was predicated. The result is the minting of unprecedented and broad new government power.
I used court cases as my example because the forces at work are transparent. The discussions I’m having lately are more about corporate behavior tracking which is anything but transparent. Unlike the government, corporations are under no obligation to afford you due process before surveilling you. Almost any web page you visit is packed with stealth trackers that follow you as you navigate the web site and even across multiple web sites. The technology enables extremely fine-grained tracking of your behavior, correlated across many aspects of your interests and over long periods of time. These trackers are designed to be undetectable to ordinary users. Tools like Ghostery reveal their presence but also trigger an arms race. Each new privacy enhancing tool spawns new versions of trackers designed to defeat it.
Bits to gigabits
Consider privacy in relation to IoT – the Internet of Things. Cisco says that 50 billion devices will be online by 2020. Others predict trillions of devices in the not-too-distant future. Where will all this instrumentation reside? Anywhere there’s electricity. Large appliances like washer/dryers and refrigerators are a given. Individual light bulbs, switches and outlets are on the market now. Forks are a current example of instrumentation soon to be embedded into much more mundane objects.
I like to use the example of electric motors. Steam engines were big and noisy and you needed large stockpiles of fuel to run them. The design architecture for the steam age was a single large motor and a system of belts and pulleys to distribute the motive power throughout the factory. The first electric motors were envisioned to replace steam motors within the same architecture: one big motor, lots of belts and pulleys. But what actually happened was that electric motors disappeared into the fabric of life. There’s one on my wrist as I write this. There are roughly 30 within arm’s reach of my chair. Electric motors are invisible. We don’t think of them as motors, we think of them as a watch, hard drive, CD/DVD player, printer, sprinkler valve, drill, toy, fan, vacuum cleaner, etc. At least with the electric screwdriver we make a passing allusion to the motor within but pretty much everything else it’s just assumed. A hand drill used to be just a drill. Now “drill” without the qualification is assumed to be powered.
Imagine the same thing with instrumentation. In the near future a “smart switch” will just be a switch. A “smart” anything will just become that thing and the old version will become a “dumb thing.” The instrumentation will no longer be a novelty but will recede invisibly into the fabric of life. When steam engines were replaced by electric motors, it was hard to imagine a time when motors would fit on your wrist. It’s just as difficult today to imagine why we’d want sensors and actuators in all our devices and objects but let’s table that and stipulate that it happens.
What is the impact of all that data?
The move from atoms to bits reduced cost and improved granularity of surveillance by orders of magnitude. The move from coarse-grained “smart objects” to pervasive embedded instrumentation will advance us several additional orders of magnitude. We dream of living like the Jetson’s or the crew of the Enterprise, but struggle with a legal system on a trajectory that more resembles Gattaca or Minority Report. When it becomes possible for sensors in your car to analyze alcohol on your breath will it be acceptable that the car will refuse to start? Will we allow consumers to opt out of such controls and if so how much of a surcharge will insurers be allowed to tack on? After all, if you have nothing to hide, why hide your data?
By the way, to those who claim we can refuse to participate, have you tried lately to buy any kind of content player that does not include an iDock? Tried opting out of a smart power meter in jurisdictions where they are protected by law? These are trivial examples compared to the question of whether you’ll be able to opt out of using sinks and toilets with built-in chemical analysis capabilities. Currently employers who conduct drug tests do so occasionally at random or specifically in connection with an investigation because cost limits the practice. But when incremental cost drops to near zero the baseline will shift to ubiquitous surveillance. Anything with a liquid drain will analyze contents. The presence detection and recognitions we have today are sufficient to correlate your identity to the chemical analysis.
In the very near future your casual behavior and activities will be trackable with the precision and detail only possible today in the confines of a lab. Every device, object or surface will potentially be a sensor. The physical constraints assumed by the current legal framework and that balanced the power of individuals against corporate and government interest are disappearing. The digital representation of you that was once a rough tile mosaic is coming into focus for vendors and government as a hi-def, crystal image.
VRM, or Vendor Relationship Management, is a new approach to conducting business in which the missing physical constraints have been replaced by technological and policy constraints that restore the balance of power between individuals and their vendors, and perhaps to some extent also their governments.
One of the issues is asymmetry in the cost of data collection. Vendors spread the capital cost of data collection over a large population of customers. Given enough time, the cost of data collection drops to near-zero or in some cases actually generates returns. Consumers on the other hand have no such infrastructure. You are co-owner of your transactional data but your grocer records each line item of your purchase in real time and you get a cryptic paper receipt which you have the option to transcribe into a database. If you had a database. And knew how to program.
VRM proposes to provide that platform so that individuals will have the means to capture more of their own data at a cost that is competitive with their vendors. Indeed, the vision is that the vendors who already have that data will some day participate in the VRM ecosystem by sharing it with their customers, in real time and full resolution. Instead of just a crappy paper receipt with unreadable abbreviated names, you’ll get the actual line items with UPC codes, prices and for some products possibly even the cradle-to-grave history and status. You’d get your smart meter readings in real time so that you could program home automation behaviors based on load, utility rate, occupancy and so forth. When you purchase online, the terms of the contract, price and all other metadata about the purchase would either be captured by you or delivered to you in real time by the vendor.
But VRM is about a lot more than just replacing today’s functionality. Just as electric motors transcended the function occupied by stem engines, VRM enables entirely new capabilities. Many are yet to be discovered but a key new capability is intentcasting. This is a direct signal from the individual to the market about preferences, requirements and purchase intent. For example, you would have the option to set a preference that your vendor must deliver your transaction data electronically. You may even be able to set a bounty on such a service: “I’ll pay a 5% premium on digital content if you deliver full purchase metata electronically at close of the transaction.” At the moment such a setting would be nothing more than a preference that few, if any, vendors would honor. However, if people set that preference in sufficient numbers (or with sufficient bounty attached) to motivate vendors, that capability would naturally emerge.
So VRM is about restoring the balance of power between vendors and their customers, and it is about crating new forms of interaction that only become possible when the consumer has technology comparable to the vendor.
Haven’t we been here before?
In the ’90s this conversation was about embracing the Internet. I took the position that everyone would be on the Internet and that business would be conducted there. Lots of business. My manager thought I was a crackpot. Two years after I left that job, Bill Gates published The Road Ahead, a book largely dismissive of the Internet, then almost immediately revised it when he realized the Internet really was the next big thing.
Today my position is that IoT is coming so embrace it. It is inevitable and it is closer than you think. If you start with 50 billion instrumented things (or trillions if you are ambitious) and work backward, what do we need to build to pave the road between here and there? I propose that privacy and parity of power in the commercial relationship must be protected primarily by code rather than law and that these must be restored before IoT can emerge and blossom. There may be laws behind the code, but you can’t police 50 billion things other than through code. Systems based on the VRM philosophy are where such code would live. In that respect, I don’t arrive at VRM by starting in the current day and working forward, but rather I see it as a hard prerequisite for any possibility of managing a world where everything is instrumented.
This is all speculation and there’s enough uncertainty that reasonable, serious people take the opposing view and I am once again the crackpot. I’m OK with that. I’ve been here before. Reasonable, serious people will continue to debate the viability of VRM even as it rises around them, built by an army of people willing to back their vision up with code, caffeine and pizza. I tried to contact Bill Gates to discuss all this but he was unavailable for comment. Presumably, he’s busy reading my blog and revising the draft of his next book.
(With a shout out to Drummond Reed who inspired the post with one of his own.)