I've been mildly obsessing lately about applying predictive capabilities to content delivery through data, and what that might mean for media consumption and content producers of all kinds. So I decided to write a piece for Mediatel about it.
The announcement last week that Google is making its Prediction API generally available is potentially transformational, essentially meaning that it's now easier than ever for anyone to create apps with predictive capabilities supported by best-in-class technology. Ford have already partnered with Google to make cars that can use the API to learn from our behaviour and optimise for predicted scenarios. But for media owners and content producers, using such pattern-matching and machine learning capabilities to optimise user experience and power content recommendation seems like a no brainer.
My point in the piece though is that whilst there is good reason for content producers to get excited about the potential behind making content delivery services more intelligent (as Clay Shirky famously put it, it's not information overload, it's filter failure), we should not forget the power behind great curation, discovery and serendipity. There's nothing quite like just clicking around the internet and ending up at something that just blows you away. I guess that's what StumbleUpon (one of the most significant drivers of traffic on the web) is all about.
A few hours after submitting the piece I serendipitously (see what I did there) happened across Brainpicker's review of Eli Pariser's new book The Filter Bubble, a "compelling deep-dive into the invisible algorithmic editing on the web", describing the drawbacks and limitations of our personal universe of information ("unique and constructed just for you by the array of personalized filters that now power the web").
I'm surprised if I'm honest that, given it's potential impact on how we discover and consume content, this subject is not discussed more. Anyways, here's a link to the piece. I hope you like it.