But as far as discovering new content and helping me to eliminate content I don’t read or don’t want to read, Google Reader is not so great. Reader is terrible at suggesting new feeds to read to me; it constantly suggests Dilbert and Lifehacker, and has for the past 5 years or so. I have no interest in either of those. It has never analyzed my reading patterns to such a degree that it suggested something new that blew me away. I am looking for those kinds of interesting suggestions: not just for feeds that match what I already read, but things that are outside of my normal bubble but would be interesting to me.
Luckily, Google Reader has an API, and rather than just exporting my data and building a new service, I can start to build off of it. I get to keep a lot of the features I enjoy while extending it with my own code.
In Pariser’s talk, he talks about encoding algorithms of filtering and recommendation with a sense of civic duty. To some degree, this means having some journalistic integrity. Such an algorithm needs to present both sides of the story. One issue of many feed readers and other content online is that, well, it only shows one view. The view of the article you’re on.
But imagine a feed reader that was more like the front page of Google News. It would not only show you the blog post you’re reading, but all recent blog posts from other authors about similar topics. Maybe, if they’re responding to a news event or writing about a known fact, the reader could do the work to track down the original source. To take it even further, without even really needing to be aware of motive, politics, and other factors, a dumb feed reader with good suggestions could probably present both sides of the story. Both sides of an argument. Both liberal and conservative takes on the same bill.
Dreaming up features like this can be a deep rabbit hole. Start considering the consequences of pulling up all past articles you’ve read about similar keywords or tags. Or performing searches for academic papers on the topic. Pulling in data from Wikipedia. Looking up books you’ve read or are planning to read on Goodreads that are related to the blog post you’re reading. Or any other number of ways to slice and dice content. And it doesn’t have to stop with Google Reader.
Opportunities to make better tools and more intelligently consume information are all around us. At the same time, consumer-consuming corporations on the web try to trap us into filter bubbles. They try to provide us with what they think we want, but can they ever really know? In the end, it’s control-your-own-filters or be filtered1.
1 The name of this article was lifted from Daniel Rushkoff’s Program or be Programmed, a book which I didn’t really enjoy. It was not what I thought it would be about: why we should learn to program so that we can control the complex technical systems around us rather than be controlled by them, or how programming’s problem solving skills can be applied elsewhere. Instead, I found the book to be some technology-fear-mongering and a bunch of diatribes about how things online or in computers are “less real.” Suffice to say, I did not enjoy it. ↩