What We Wish We Were Investing in Now: Serendipity Systems

 Photo by  Sapan Patel  on  Unsplash

Photo by Sapan Patel on Unsplash

Humans love serendipity. Unlike so many other species, we crave new joys; we delight in dancing past the edge of our experience into the unknown, and then returning safely to the fold.

I have often said that serendipity is the key to media. Editors-in-chief, during my era among them in the 1980s, got that job generally because they could—either through incredible gut instinct or by a ferocious study of audience analytics—deliver serendipity to their readers. The primary information brought the audience, but the unexpected new fact is what kept them around.

The same is true of digital experience. Humans are humans, whether online or off. But today’s online experience not only fails to deliver serendipity, it saddles us with the exact opposite. Algorithms, and even today’s machine learning AIs, all draw conclusions by analyzing past behaviors. By definition, that is a narrowing funnel—the same thing gets delivered more and more often. Surprise diminishes, confirmation bias grows, isolation into camps of the likeminded increases. It has turned out to be a far more negative human experience than anticipated, verging on a danger to individual health and social stability.

We focus on the technological vectors of happiness, so I am thinking about serendipity a lot. We see the path toward technologies of happiness as deriving from delightful moments, small quanta of time that let people step out of the humdrum daily norm toward more energy, confidence, calm, serenity, absence of anxiety, reduction of pain. Serendipity is the main line toward delightful moments.

What’s so difficult about producing serendipity is that it varies by culture and person. The Amish celebrate not changing; their serendipity would come in extremely small shifts. Any dramatic new thing would lie outside cultural norms and be impossible. Some people, my wife for example, like going to the same restaurant at the same time every week and ordering the same thing. Her conservative food behavior will only tolerate rare, small changes. Others, like me, are: Wakandan food? Never had it. Let’s try it! Serendipity, in other words, lies in the eye of the beholder. Get it wrong—go a step too far and propose something too far outside an individual’s comfort zone—and it becomes instantly cringe-worthy.

So how can technology actually help produce, rather than reduce, serendipity? Right now, the prime candidate seems to be neural networking. Neural networking, where probabilities are developed by running gazillions of options across a network of independent processors, is the “other white meat” of AI, alongside the much better-known machine learning. Unlike machine learning, neural networks don’t base their determinations by looking back at past information. Instead, they project many options into the uncertain future, and then compile all those outcomes to see a possible pattern moving forward.

Because it can see so many outcomes, and can see them extremely quickly, neural networking has greater serendipity potential than any other existing tech. The right serendipitous option has to be in that set of outcomes somewhere. The trick becomes identifying it.

Which implies that machine learning (the past) applied to the derivations of neural nets (the possible future) may be the best path toward individual serendipity.

That’s just personal speculation, though, not science. What is undeniable is that serendipity tech is one of the great opportunities for this generation. The team that can solve this problem will, literally, alter human experience and create value of historic proportions.

I can’t wait to meet and have us invest in that team!