Date

Today, ad networks like Google and Facebook use the following strategy to connect businesses with potential customers:

  1. Accumulate information
  2. Secure the information
  3. Predict behaviour by applying heuristics or algorithms
  4. Gain competitive advantage
  5. Assess and repeat

It's an effective strategy for them and clearly a worthy, lucrative problem to solve. Successful organizations and individuals have been following this pattern since forever and have applied it against any number of problems.

Wayne Gretzky skated to where the puck was going to be, not where it was. He wasn't predicting the puck, he was predicting the behaviour of the other players. Lionel Messi isn't thinking about the ball; he's in tune with what both his opponents and his teammates are going to do. The same pattern holds in sports, business, academics, politics, or anywhere else: whoever best predicts behaviour has the advantage. Though if the playing field is not level, or you lack a competitive drive, you'll lose out anyway.

The problem with online advertising is that step one - accumulating information - is achieved by tracking everyone's online activity. Whether through supercookies, tracking scripts, analytics, ads themselves, or web-app usage, most online activity is tracked and captured. It's a horrible mess that puts each individual in a weak and worsening position: those with all of the information and the biggest computers receive ever more power and advantage. Anyone without information, security, algorithms, or enough computer tends to lose out. Jaron Lanier speaks to this more eloquently than I can.

An ambitious minority uses this strategy on a small scale in concert with their personal computers, using organizers, information managers, and calendars, or even custom systems, but it's all sailing against the wind. The modern computing environment is inhospitable, to put it mildly: information about you is spreading everywhere, personal computers offer neither security nor the privacy it provides, and there's no easily accessible heuristics or algorithms to put to work. Certainly no coherent package to pull off the shelf. It's true that anyone can use the old brain, paper, and pen techniques (and many do), but this pales in comparison to what computers can do and are doing.

How Kill Online Advertising

To kill online advertising, flip the script by putting individuals first. Do exactly what they're doing but for yourself, without them. Don't invite them to your party. Take their general, time-tested strategy and instead center it on you:

  1. Accumulate information about yourself
  2. Secure your information (from anyone, including "the cloud")
  3. Predict your behaviour by applying heuristics or algorithms
  4. Gain competitive advantage for yourself
  5. Assess and repeat

As far as I can tell, there is no generally available computer or software designed specifically and exclusively for this purpose. There's no clear answer to the question of "what do I do?". Until such technology becomes available, our inhospitable computing environment will continue to degrade. The strong will get stronger, and the weak weaker.

Key Requirements

Fundamentally, this new software (we'll call it software) must offer a way to accumulate information, be highly secure, embed useful algorithms, and then result in competitive advantage. It must also represent a natural feedback cycle: useful predictions result in high quality information that is ideal for folding back in to the collection.

These key properties are necessary to achieve a legitimate information economy like Jaron Lanier talks about. For one, the information you have about yourself (generated out of self-interest) is of the highest quality, worthy of trust and capable of being intentionally packaged up and shared or sold. Since we know the mediocre stuff has value, the rarest and best stuff must have much more.

If the system is secure, there's much less risk of whatever's sold becoming freely available. First, security helps assure buyers they won't see the information they've paid for floating around on the grey market. Unlike with movies, music, or software, the value offered by information drops when copies are made. Second, fair prices incentivize the buyer to keep their purchased information secure and make good use of it, or else they've wasted their money. Third, sellers now have good reason to only share or sell with those they trust. But all that comes later.

Fine. What else?

There are many ways to build a computer system that technically meets those requirements but remains completely unusable. So we have to do better. Since the point is to offer value to the average non-technical person - those most vulnerable online - this software must also be broadly accessible, ridiculously easy to use, and still work while air-gapped.

Yes, air-gapped, or else there's no guarantee of control. All sorts of things happen beneath the surface of modern computers. Without control, there's no security. With neither, privacy is a dangerous illusion. Nothing (certainly not personal computers) is secure while connected to the Internet. No doubt this software will do more when networked, assuming the risks are understood, but the core value must remain available in isolation.

Why not throw in unicorns and rainbows while we're at it? This is getting pretty far out there. But then, who's to say how it should work or what it should look like? Nobody really knows! This thing doesn't exist yet. Perhaps it should.

My Approach

I believe the solution begins with the simplest possible personal computer. Hold onto your hat.

This computer begins with a single button. The user gets one thing to do: press the button. That's also what it looks like, one button beckoning "press me". Most likely on a touchscreen.

This computer is unusual in that the simplicity isn't hiding anything; it's not a facade for apps, settings, or gestures to hide behind. Until the user presses the button, it's not possible for this computer to offer any value. There's simply nothing for it to do. Any value it offers depends entirely on the button-pressing pattern, now and in the future.

Each time the user presses the button, the computer saves the time. After two or three presses the next press can be roughly predicted by extending the time of the most recent press by the average of the last few intervals. This heuristic is rough but a fine place to start. Accuracy would be counter-productive anyway. When a prediction is available the user is visually informed of when the predicted press is expected.

So when is the button pressed? Like with regular computers, it's a garbage-in garbage-out situation. When just messing about it's all meaningless; the user won't see any value at all. If the user is earnestly pressing the button but for the wrong reason, they'll see very little value in it. However, there is one specific reason to press the button that opens up many interesting possibilities.

It's this: the user presses the button only when they take action in their life. Any action, no matter how major, minor, or how long it took, or whether it was just now, an hour ago, or yesterday. One press per action, no detail needed. For each button, the actions should probably be related. But it's all good as long as each button-press maps to a real action. While there's only one button at the beginning, there's no upper limit; the user decides what each button means, how many they want, and how they're organized.

That's the simplest personal computer.

Unlike how newbies first experience regular computers, you're already an expert on how to use it. I can say with great confidence that you're a button-pressing fiend and you're also the leading authority on what you get up to.

It's much more personal than usual because it's actually all about you. The only information to work with comes from your button presses about your actions; there's nothing else. Any structure that arises is of your own making, so you will understand it. And the overall simplicity means it works just fine without ever touching the Internet. That's wonderful for privacy.

From the time series, rough predictions are made about your future. This is because you are mapping presses to actions. The best predictions come from past actions, and that's exactly what the computer has to work with. You might be surprised at how predictable you are when your various activities are isolated into their own frequency patterns.

That's Not All

I've implemented all of these ideas in Benome. Although it's newly released and early in its development, you can already find in it many of the above properties, with more to come. It's not quite as simple as it could be, but close; I've pulled it back for now. It'll be open sourced soon enough.

Simple beginning

Drag from the orange circle to the middle to create a new activity. Drag that activity back to the orange circle to record an action. That's all you need at the beginning and it's simple enough for almost anyone.

Simple beginning

Past the early basics, it's tap to navigate and drag and drop to reorganize. Then more capabilities whenever you're ready. Each person progresses at their own pace to develop a well-structured pool of high quality information that mirrors their own life and goals.

Because it's a web app, it's immediately accessible to a wide range of people - anyone with a remotely modern computer or somewhat recent smartphone (I use a Nexus S from 2010).

Most significantly, it will be capable of standalone operation (no Internet required) like classic software. That's the only way to get complete control over privacy and security.

After a few months

This second screenshot is the top-level view of my personal benome. I track hundreds of activities with minimal effort. The circles normally have labels but they've been removed for the sake of privacy.

When navigating by tapping circles (as shown in the next screenshot), they become the new central focus and expose further activities. All of the original basic interactions still apply.

Visual middle

In case you hadn't noticed, it's visually very different than the usual piece of software: smoother, more fluid, and more organic. No boxes, borders, corners, lists, layering, stacking, or even necessarily text. Those attributes tend to evoke much more structure and form than is appropriate for an intensely personal experience. They are useful, but best suited for broad distribution, impersonal scenarios, and specialized needs.

Deeper focus

This third screenshot shows the social activities I'm involved in (minus labels). It may look a bit busy but there's a couple of things to note:

  • Your structure is completely about you
  • Because you develop your own structure, you know exactly where things are and what they mean
  • This view shows everything but it's easy to reduce the level of detail to what's important right now, relative to the current focus.
  • You can see how you got here; there's a visual return trail indicated by the white glow
  • It's very quick to navigate both nearby and vastly separated activities

Focus

Interactive demos

Video

A birds-eye time lapse. Context

Really

Online advertising feeds on personal information and will die without it. It dies when you make it irrelevant, when you choose to behave in a way that values your personal information. Take care of yourself first.

My hope is to share these ideas to spark discussion and kick-start exploration of your own vision. If you think my premises are flawed, I'd love to hear about it. If you think I'm onto something, get in touch. And if your vision differs wildly from mine, that's quite alright. You have enough to get started.