I started using news aggregators like Flipboard, News360, Zite, Summly, Rockmelt almost from the minute they came out. Initially, I found them amazingly useful–they helped me rediscover blogs or authors I had forgotten, they introduced me to new bloggers, and they widened my perspective on topics I was already reading about. I quickly realized that it was highly unproductive to try to follow even just a couple of them, so I picked Zite as my aggregator of choice.
For a while I was amazed by the productivity I could achieve and the quality of content the app allowed you to find—even when originally published on less known sources. Yes, every now and then the “algorithm” served up articles from content farms, but I could indicate that I didn’t want more of these, and they would go away.
After a few months, the discovery of new bloggers naturally declined. But more importantly, I noticed that the “algorithm” was increasingly incapable of sorting out the low quality content (things like top ten lists, if you know what I mean). In fact, it seemed to be less sophisticated than it initially appeared.
I gave some other services a second chance, but found that the situation was no different. So, in conjunction with the closing of Google Reader, I had to go back and select a new reader and, in the process, manually create a new list of feeds. Certainly I have Zite and co. to thank for many of the sources that I now follow, but in the end, I went back to a smaller, more qualified list of content sources that I refer to daily.
What this tells me is that the algorithm has still room for improvement. As providers of a deep linguistic/semantic platform, I can tell you that you can do much better from what the current apps are offering. However, this won’t be enough—human curation is still necessary.
The lesson here is this: You cannot completely avoid a certain level of complexity if you want to enjoy a certain level of quality.
The benefit of good technology is its ability to turn an unmanageable quantity of content into something manageable. In this case, a good technology is one that supports curation by making a rich set of attributes available in order to select what is important and filter information in a fast, rich and effective way. At this point, the information becomes something that human can manage.
And for now, at least, this is something that we are still better at.