At a recent seminar (hosted jointly by the Royal School of Library and Information Science and KU’s Centre for Communication and Computing), I had the chance to listen to Lennart Björneborn‘s talk about serendipity. The history of science is full of examples of “happy accidents,” from the discovery of penicillin to that of post-its – and in general I wholeheartedly agree with the point that it’s important to “plan for the unplanned.”
The problem is, of course, that doing so is quite difficult. It is not easy to create work environments conducive to serendipitous discoveries, because what counts such seems to depend very much on individual factors. Remember that bit in The End of Eternity, where technicians from the future can consciously alter the rate of scientific development by e.g. rearranging items on a scientist’s bookshelf? For better or worse, we’re not quite there (yet?).
But the point immediately relevant to me is that, according to Björneborn, algorithmic recommendation systems – from Amazon to Google News and Spotify – are “great serendipity engines:” they seem to be very good at suggesting content that is interesting, inspiring, and that we would’ve missed otherwise.
If this is so, perhaps social news sites are the best serendipity engines – although sometimes I’m afraid that it’s also difficult to find a balance between procrastination (getting distracted) and productivity (getting inspired).
And of course there’s still that big if: what if all this serendipitous content that we get recommended is all fundamentally the same? Nod to Sunstein, Bohman and others: what we society needs is not just serendipity, but individuals’ exposure to markedly different contents and ideas- such that we could not be exposed to through recommendations which come from people just like ourselves.