Technology tracks our every move. Much of this happens automatically, as when Google records our web searches, our interests, and our whereabouts. Facebook tracks our friendships, our relationships, our interests, and our location. Many other entities, including most tech companies, track us.
This is pretty much what we wanted, right? We’ve dreamed of personal assistant AIs for a long time, and more generally of technological systems that know everything about us. In Star Trek, the Enterprise knows everything about each of its crewmembers, which is usually quite helpful. Most visions of the future assume that technology will know us a lot better.
The dream is that technology will remind us when we’ve forgotten something, notify us when our diet is deficient, let us know when our commute will be longer than usual, give us context for places we’re visiting or activities we’re doing, congratulate us when we’re doing well, and reprimand us when we’re not meeting our goals.
Between Google, Facebook, Fitbit, and other services, a significant amount of this happens, albeit in limited scope. It seems like it’s just a matter of time before tomorrow becomes today.
I’m starting a project to make it easier and more appealing to log information about ourself. My technical thoughts are in another post, but here I answer the question, “Why am I not satisfied with the current state of affairs?”
Most visions of the future didn’t include cloud computing. It turns out to be really convenient and efficient to store everybody’s data in the same place and use it for them, giving them the results. As we use this for more and more of our data, two issues become more and more acute.
First, corporations get access to huge amounts of your personal information. Most people don’t have to be convinced why this is a bad thing, but if you want some light reading on privacy concerns regarding Google, I recommend the Wikipedia article by the same name. I pick on Google because they’re big, but most others are no better.
It’s worth noting that data, once collected by a corporation, doesn’t tend to stay there. Data leaks happen. Companies sell data, or they sell themselves. Government agencies request data, often in very broad terms, and usually receive it.
Besides the instinctual aversion most have to large, faceless entities knowing everything about them, I recommend Googling “nothing to hide argument” to understand more about the issue. You’ll find reference to Edward Snowden’s statement that, “Arguing that you don’t care about the right to privacy because you have nothing to hide is no different than saying you don’t care about free speech because you have nothing to say.” I have no more to say on the subject than has already been said.
Secondly, when corporations store your data, you rarely have access to it in a useful manner. Even if Facebook knows what my and all my friends’ faces look like, I can’t use this information to, for example, organize photos on my smartphone or desktop. I can’t even use it to organize photos on my Dropbox because that’s a different information silo.
Most other services keep our data similarly difficult to obtain, and even in cases where it can be exported, the tools often don’t exist to use it. This is in part because there isn’t a large number of people who want to leave these services but keep their data, so there’s little demand for such tools.
Thus, an effort to improve our ability to log about ourselves needs to be decentralized in nature. Each person should have control of their own data, and other entities shouldn’t have access to it unless they’re specifically given it.
Doesn’t track the interesting stuff
Everybody wants to track different stuff for themselves. Some care a lot about tracking their diet and fitness, which is why tools and companies have sprung up to do so. Others use tools to track their money. Still others track more non-standard things, like crying episodes, number of people they interacted with each day, or pretty much everything.
The general principle is that if enough other people are interested in tracking what you want to track, then tools exist for you to track it for yourself. If a corporation has a financial interest in tracking something about you (for example, interests, so they can show you targeted ads), they’ll do that, but you’ll generally not have access to this data. If you want to track something that few others wish to track, you’ll have to do it manually, and build any tooling you need.
Since a fairly narrow set of things are commonly-tracked enough to have tooling, most interesting logging requires you to build your own tools and processes. There’s a strong temptation to try to get technology to passively collect information about you, and there’s some merit to that approach. You can collect a lot of data that way, but it’s rarely the actually interesting data. But to flesh this out we need a clearer idea of what data is interesting.
In general, “interesting” stuff to log is what Tarn Adams (Dwarf Fortress) calls the “narratively interesting” parts of the world. These are those things that might be included in a story, if you were to write one about yourself. It can be helpful to think, “what would be in my journal if I were to just write what happened in my day?” What would be in your autobiography, and what wouldn’t?
Things like “number, severity, and causes of crying episodes” or “number of poeple I interact with each day” are interesting things to track. In general, you probably need to track with greater precision than would actually be included in a story because it’s the conclusions that are interesting. In other words, it’s less interesting that you interacted with five people a day for these few months than that you went this many years interacting with the same people each day, then two years interacting with nobody, then a bunch of people, and so on. This is especially interesting if you can correlate it with other data about yourself, like general wellbeing, weight, etc.
Any serious effort to make it easier to log personal data needs to recognize that choosing the narratively interesting data is much more important than just gathering any data that’s available.
Personal logging can have a couple of purposes, but the primary one is to analyze the data afterwards and look for insights, like “you’re generally happier when you eat three meals a day” or “you tend to get stressed on weekends” or “half your expenses are from eating out”.
As mentioned before, we don’t often have the ability to analyse our own data, unless we manually log it and build our own tools. Analysis can never have a one-click solution because it’s all about drawing the kinds of insights that we want from our data. Still, many tools can be built to help.
If you’re lookihg for a kind of insight that nobody’s looked for before, then you’ll have to write your own tooling. Otherwise, you hopefully won’t have to. Of course, if you don’t have access to your data, or if that data isn’t interesting, then those tools won’t exist, and you can’t build them.
Logging without analysis has no purpose.
So, then, a personal logging solution must be decentralized (i.e. owned and operated by the user), it must track only whatever is narratively interesting, and it must be possible, as an individual, to analyze the data and find useful insights.
I haven’t come across a project that checks off all three of those boxes, but I sure hope there is in the future. The best way I know to make that future happen is to start building it and see what happens.
I would appreciate pointers to similar existing projects (current or past, successful or failed), and if anyone has ideas on the subject and is interested in comparing notes or working with me, get in touch.
Addendum on world modeling
One of my secondary goals with this project is to experiment with modeling worlds. This stems from a childhood filled with made-up worlds like those of Brian Jacques, J.R.R. Tolkein, C.S. Lewis, and George MacDonald. Using computers to model worlds is particularly challenging. Good examples I see include Dwarf Fortress (which creates the most compelling and diverse stories of any computer game I know), the Civilization series of games (which takes a different tack on which things to model, but each round has fairly unique storyline), and Chris Crawford’s Interactive Storytelling ideas.
Basically, I’d like to separate the problem of building a world from the problem of modeling it, so I may as well use the real world as my first world to model. Perhaps the concepts can be applied to generating made-up worlds.