Serendipity in the Digital Age

I'm currently attending a meeting in Chamonix, France, and one of the questions that was discussed tonight was, what is going to happen to serendipity in the digital age? Serendipity means fortunate happenstance, and some people seemed concerned that the digital age would reduce serendipity, because in their view, it requires real face to face human interaction. I think this argument is weak and perhaps reflective of a certain generation that feels nostalgic for the olden days. But I do agree that serendipity is a good that needs to be preserved in the digital age, and this is no trivial task.

In evolutionary biology, serendipity is captured by the concept of beneficial mutations. DNA, the underlying chemical basis of all life on earth, mutates all the time. Most of these mutations are bad in that they either render a cell inviable, or cause some type of malfuction. Less often, the mutation is neutral - the encoded protein might not change, or change in a way that has no effect. Rarely, but occasionally, the mutation is beneficial, and the new version of the gene is better in the evolutionary sense, i.e. it directly or indirectly contributes to getting itself more successfully copied into future generations.

I would be surprised if the distribution of outcomes in random human interactions were very different. Most interactions are simply not interesting enough to be even started. Imagine the hundreds of people you bump into every week - even if you started a conversation with each single one of these people, you would readily find that most of them are not your type, boring, or just not interesting enough to talk to, beyond polite chit chat. A few people may be interesting enough for you to keep the conversation going for a while, but the relationships will fizzle out after some time. Rarely, but occasionally, you meet a stranger with whom you connect strongly, and they become colleagues, friends, or even lovers. Many years in the future, we will think of the serendipitous moments when we met those people, while conveniently forgetting about all the other interactions the led nowhere.

Can anything more be said about this somewhat vague comparison between genetic mutations and random human interactions? I think so. The interesting question is, how can we optimize the rate of random interactions in a way that ensures that we get the most out of it? Simply more randomness cannot be the answer, and indeed, that's not what we find in nature. In fact, mutation rates are generally extremely low, as low as only 1 mutation per billions of base pairs. But do not think for a moment that the mutation rates in nature are fixed - like everything else, they are subject to natural selection, and they do vary widely both among and -even more so - between species. Viruses, for example, have some of the highest mutation rates among living things (if you grant them the status of being "alive"). The reason for this high mutation rate is that they are constantly attacked by their hosts' immune system, and they need a way to escape from that threat. If they mutate, then the chance of not being discovered can increase substantially. Not mutating is simply not an option, or else the immune system will bring the infection under control before it had a chance to be passed on - an evolutionary dead end. Indeed, the genetic makeup of a population of viruses within a person infected with HIV can change drastically in just a few days.

To bring this back to human interactions - how much serendipity we want often depends on the circumstances in our lives. In moments of great stability, we may not be all that interested in too many random interactions. When we are alone, or looking to expand into new social circles, we may actively connect with as many new people as possible. Professionally, it's probably best to always maintain a relatively high "mutation rate", especially if you work in a fast moving environments (and don't we all these days?).

In the digital age, many of these random interactions occur online. On one hand, since the online community is global, you can now sample from a vastly larger group of people, which may or may not lead to more serendipity. On the other hand, your sample is increasingly driven by algorithms, and in my experience, these algorithms try to predict what I like based on what I already like, which is a kind of targeted serendipity. But there is also the danger that I am getting too much of what I already like, and the true, random encounter with a new person or a new idea becomes increasingly rare. I would love for these algorithms to allow me to fine-tune the level of random interactions.

Ultimately, it is too early to say what the digital age will do to serendipity. My sense is that it will be a good thing, but I guess that's why my friends call me a tech optimist.