AI, Inefficiency, Innovation

It occurs to me occasionally, as I see military branches and governments around the world leap headlong into a future maximally enabled by artificial intelligence… that we might be witnessing the outset of the absolute eradication of the most powerful force in the history of human progress, creativity, and innovation- Could AI be beckoning the death of serendipity?

Think of all the poor choices, the missteps, the bias-driven decisions, poor placements, and inefficiencies throughout human history that were the starting points for some of our greatest leaps forward, for creative explosions and exponential advances. I am tempted to say that nothing is ever accomplished by pure force of will or skill alone. Stephen Johnson’s book Where Good Ideas Come From brilliantly illustrates the powerful, absolutely essential role that serendipity plays in human inspiration and the accidental fusion of concepts into previously unimagined new forms. Malcolm Gladwell’s Outliers, while absolutely about the power and potential behind persistence and perseverance, is also about the serendipitous conditions that have enabled some of our greatest thinkers to be in the right place to apply that thinking to make great things happen. Serendipity is everything, including motivation, imagination, and skill. I like to remind myself I was simply blessed with the good fortune to have bad experiences that inspired me to try as hard as I do. I don’t believe I have ‘earned’ anything. Serendipity is everything.

The power of serendipity is demonstrated beautifully by certain concepts of evolution that I learned from Andreas Wagner’s wonderful book Life Finds a Way. I learned that natural selection can only make things more fit. It is no more than a one-way ticket to increased efficiency and greater fitness for survival. Those with less fitness for survival die off and the genetic alleles which make them more fit grow more prevalent in a population. It’s a simple process. But the points of greatest potential can’t be reached by way of pure advancement. There are innumerable low peaks on the jagged slopes that an organism can get trapped on, incapable of escaping by means of natural selection, because the only way out is down, and natural selection isn’t capable of making a thing less fit.

So natural selection on its own is only really capable of getting a thing stranded on low peaks in the fitness landscape, unable to advance because every direction is down. Wagner calls natural selection an “overzealous janitor”, because without a little messiness, evolution wouldn’t get us very far. Luckily, there are other forces at work, like genetic drift, which bring just the right amount of chaos into the mix. These forces, unlike natural selection, are random, and serve the function of increasing entropy so that an organism can be serendipitously knocked off of the low peaks they’ve been stranded on, giving them a chance to evolve further, rising gradually in fitness to the next peak.

This is one reason I start shouting when people tell me that Continuous Process Improvement (CPI) fits into the same category as innovation. I very strongly believe that it does not, because CPI is severely limited in its capacity to make things worse. It’s even in the title. It is like natural selection, in that the only acceptable direction is up. Innovation, on the other hand, is what I would consider the art of courting serendipity. Using practices like Design Thinking, we dare to do things in a worse way in pursuit of greater value– we recognize that some plants may take time to bear fruit, and that’s ok. That’s not to say that CPI doesn’t play an important role. We should also continue to seek efficiency in many contexts. It’s just that at the same time, we should recognize that pure pursuit of efficiency alone can get us trapped on those low peaks of the adaptive landscape.

So my question is this: What if the ubiquity of AI and our increasing application of algorithms that sterilize and put the disordered in order is like relying more and more on that efficiency-seeking force of natural selection? What if our subsequently decreasing agency is the loss of those forces of disorder that would bring us to new and better places, if only by the power of entropy?

I will not get selected by an AI for the job, because my talents for explosive advancement are hidden under the disheveled folds of my inefficiency. I am a force of chaos, which makes me extremely inefficient, but if you make the serendipitous mistake of hiring me, I will innovate the shit out of your processes. I, for one, do not welcome our new robot overlords. They lack the human power of abstraction. They fail to make important mistakes. They hide from entropy, rather than embracing it. They might be extremely efficient, but perhaps that will prove to be a greater impediment than a benefit, as we increasingly get stranded on low peaks in the adaptive landscape. The decisive advantage of the future is not efficiency, but serendipity.

My suggestion? Keep your artists around, to balance out the engineers. Infuse your culture with opportunities and incentives for divergence. Let mistakes happen, because oddly enough, they can prove to be more beneficial than successes. Do not pursue 100% efficiency, because efficiency is a narrow goal at a finite distance.

Court serendipity, and be sure not to drive her away.