Bobby had just created something to be proud of, a real article of creativity. He had been working on his song for days, hunched over guitars, computers and microphones. A laborious process, but after he had his result, he was satisfied. It was an artifact worth the effort he had put into it.

The only thing left to do was to get it out to the world. In the music industry this would probably be more painful than creating the music itself, but luckily for Bobby he wasn’t in it for the money – so he does what any other modern musician would do nowadays. He posts it on social networks. Namely, Facebook.

After submitting it (and breathing a sigh of relief), he moves on to browse elsewhere for awhile. In the back of his mind, though, is the constant thought – are people listening? Are people enjoying my work? Creativity is largely a social act, and Bobby is in the final stage of it—validation.

He checks back every few minutes, and watches his update begin its inevitable trudging down the feed wall into oblivion. His work is slowly overtaken with gems such as:

This is a tragedy. And Facebook knows it.

The Problem of Worth

The problem here is simple: the gap between the quality of these two posts is very high. One of them is a genuine article of effort – someone truly creating something novel and sharing it with the world (whether your tastes align with the content or not). The other is half-hearted mumblings at best. Even the authors themselves wouldn’t deny the difference in quality.

Facebook has acknowledged this problem recently, with their introduction of the “News Feed” versus the “Live Feed”. The News Feed is meant to be a subset of the Live Feed – the things that are “important”, an algorithm primarily based on number of comments and likes. Both the implementation and the naming are a bit clumsy, but the effort is noted: Facebook understands the noise problem, and that it has only been exacerbated with the prevalence of third-party application notifications. They’ve tried to solve it, but haven’t really nailed a way to determine the quality of a post.

Quality is a User Problem

“Quality”, in this case, is a very nebulous concept. It’s not really measurable in machine terms – this makes it a particularly hard problem to solve by filtering. “Likes” are not exactly newsworthiness, for example. You could even make the argument that this enters the realm of Strong AI—machine intelligence that matches or exceeds a human’s—due to its requirement of very human characteristics like taste and emotion. And Strong AI is a very long way off from reality. This is simply not a problem made for machines to solve.

So if it’s not a machine problem, it’s a user problem. That means we have to rely on the authors, and their readers, to figure out the quality of their posts for us.

One way to implement this is below – a simple “update volume” slider, that asks the user “How loudly do you want to say this thing?”

Depending on the author’s choice (and a few other factors like frequency of posts), the update would be displayed more or less prominently on their readers’ update streams.

The authors themselves aren’t ignorant; by and large they know the worth of their own submissions. Putting the problem into their own hands to solve may seem a bit strange at first, but it provides us with a very valuable data point to begin from: Are your words worth anything?

This is the hard part; this is the user problem. From this bit of data, we can begin to solve the simpler problems programmatically: Is a user being arrogant and loud by posting high-volume content too often? Automatically scale their volume based on post frequency. Are application notifications annoying, but still potentially useful for some users? Set them as the lowest volume by default, but allow a reader to specify individual modifiers on specific Applications. These are all solvable problems once we’ve pushed the intractable one of Quality into the authors’ hands. This single data point provides us a lot of ground to stand on when making otherwise-difficult decisions about the worth of a post.

I’d be interested to see how this would play out in the Facebook ecosystem. It seems a particularly more noisy system than other social streams like twitter, due to third party applications posting junk. In the end, we’ll either need better quality control, or fewer friends.