The Loss of Serendipity (my take)

Let’s get on the same page here. Serendipity means to find something you didn’t expect, a.k.a. finding something interesting you weren’t looking for in the first place.

Many services try to make sure that you only see things you are really interested in, this is accomplished by using all kinds of metrics to figure out what might be of interest to you, and you alone. Did you spot the error? An algorithm decides what are you interested in. I don’t know about you, but I don’t think that this is remotely possible for the wide spread of interests which tickle my fancy at any given time. I would venture it is close to impossible; let’s have a look at some of the topics I’m currently interested:

  • Motion graphics
  • Chinese writing
  • Quantum mechanics
  • Video compression
  • Philosophy
  • IOS development

The real list is longer, by the way. I just need some of the topics to explain the dilemma.


The results are filtered. This is a good thing, isn’t it? Not necessarily, do you remember the times when Google used to be a time-sink because of all the interesting things you found while searching for something completely different? Dilemma-alert! Finding the exact things you are looking for is great, but some of the topics above became interesting to me because I found something I didn’t expect to find in the search results.

Should an algorithm succeed in getting one the aforementioned topics right, then all is fine and dandy as far as our brave algorithm is concerned, but–when I’m looking for an explanation concerning odd or even line priority when encoding an interlaced video for example–am I not interested in search results regarding the other topics? Right, that depends on my search mode! Should I be “doodling” with search, then I want to be surprised; should I be time pressed to find a solution, then I’m not. Getting this right as far as the algorithm is concerned is next to impossible.


You only get the news you are probably interested in. I define news in a very broad sense here, this means that I’m talking about actual news from the wires and blog posts at the same time. How could this work? Let’s assume I had a bad day editing video for whatever reason. Do I want to read news related to that topic when looking at news? Maybe, maybe not. Worst case: I decided to not get bothered until next morning, how should an algorithm be aware of that?

Music, moving pictures, and books

“I didn’t like it at first, but it grew on me.” Do I need to say more? My taste regarding entertainment is the worst nightmare as far as marketing is concerned. Just because I happen to like True Blood at times does not mean that I’m interested in vampires during their puberty. I’m a great fan of Supernatural, but I also like Downton Abbey and was really angry about the person who decided to can Caprica or Dirty Sexy Money. Did I mention that I love almost everything Woody Allen did? Feel free to extend the list to books and music. I’m sure you won’t be able to figure out what I like. Simply because it’s a matter of the mood I’m in.

Should an algorithm be able to deduce the mood I’m in, then it will fail nevertheless. Some books or movies are of interest to me because they were written in a certain way, shot using a certain camera, or were edited using a certain software. Regarding the latter, not necessarily the software I’m using at the moment.


Even the best filtering algorithm in existence won’t be able to cater to my needs. I don’t think that I’m the only one coming to this conclusion. And sometimes I don’t want an algorithm trying to figure out what I might be interested in.

The straightforward way of achieving this would be a checkbox titled unfiltered honored by all search algorithms. Like this is ever going to happen. Search is a business, and we have to live with the consequences.

I’ll update this post should I find a way around this dilemma.

Speak Your Mind