Published: Mon 30 March 2015
While you can predict your own behavior fairly well, you are largely opaque to everyone else. Nobody can read your mind and nobody is observing you all that closely. Your good friends have enough history with you to have a general idea of your habits and preferences. That is, they can predict your behavior better than most.
To be social you'll often have to take action or share some information. Action generates information so it need not always be explicitly shared. Whether you simply act, state your intention, share your state of mind, or cast about for this information from others, the topic is future behavior.
In essence, you are communicating the fluctuating probabilities of the various options for your future behavior. It's a negotiation, sometimes with yourself. As a group gets larger, less connected, or the number of acceptable options becomes large, the negotiation becomes complex and more likely to result in discontent or deadlock.
Now, when you digitize your actions appropriately, even
simple algorithms can quantify the rough probabilities of your future behavior. Think of it as the probability that you'll say yes to something right now, putting a number on how open you are. With probabilities for a collection of behaviors, they can be sorted, managed, and presented in a visual way. All of this is what Benome does.
What does this mean, exactly?
First, you might find it useful to automatically share specific "openness" probabilities with specific friends. Or with the entire world. Twitter posts, subtext, graphic T-shirts, body language, and Facebook status updates are already used for this purpose. When digitized, there are other interesting possibilities.
Much more significantly is when a group of people have each digitized their actions related to a common interest, say mountain biking. Now you have a group of people each sharing their "openness" probability with a service (or amongst each other). This means that at any given time, there's a quantified measure of who will likely say yes to a suggestion to hit the trails.
You may recall that the
basic prediction algorithm can be viewed as producing an approximate smooth wave. What happens here is that those smooth waves are combined and scanned for high peaks. Every peak that passes the trigger threshold is a high-probability group activity. Now simply send out a suggestion to the most open members of the group and let them work out the details.
The result is automatically suggested group activities that works even with simple information and gets better from there. The problem isn't scheduling an exact date and time to meet; that's been long solved (although this can improve it). The real problem is getting the ball rolling on ad hoc social activity with minimal effort, cranking up the chance of that activity happening.
Automatic suggestion means you'll do more of what you enjoy: get out onto the trails more often, get out for coffee with friends, coordinate baby strolls or dog walks with your neighbours, get together for ball hockey. It also applies to everything else you do.