Excerpted from Chapter Two
There is a curve that has followed me throughout my career. It is not a normal distribution curve; in fact, it doesn’t look anything like a bell. It is abnormal, yet I saw it regularly in science—first as a doctoral student, then throughout my research, and later beyond my primary field. As I began my business career, it appeared everywhere, but I couldn’t make sense of it. It is rarely spoken about in technology circles, yet it is persistent there as well. The curve appears in brains, ants, the internet, and virtually all other networks. It is a curve of success and looks like this:
Biological networks all follow similar paths and obey simple laws of nature. It should come as no surprise that technology’s greatest networks do so as well. What is surprising, however, is how predictable these networks are and how little of that predictability we actually use to advance technology. The biological basis of TCP is millions of years old, and we have understood it in terms of the brain for at least 100 years. Yet when it came time to “invent” TCP for the internet, we did it from scratch, the hard way. We don’t need to wonder whether the internet, or any other network, would have developed quicker “had we known” because we do actually know.
And it is not just the internet or even just technology that can benefit from an understanding of how networks function. All businesses, all consumers, all individuals can stay ahead of the turbulence and create an environment of success by understanding what lies ahead. It is predictable, but it is also pretty enigmatic.
These laws are easily understood and have profound implications that enable us to predict where a network is headed. Want to know whether your friends will be on Facebook? How about whether you will be using Google in five years? Is Apple going to remain the golden stock? What’s the next big thing? All of these questions have answers and they come in the form of what happens to biological networks.
There are three phases to any successful network: first, the network grows and grows and grows exponentially; second, the network hits a breakpoint, where it overshoots itself and overgrows to a point where it must decline, either slightly or substantially; finally, the network hits equilibrium and grows only in the cerebral sense, in quality or intelligence, rather than in quantity.
PHASE 1: GROWTH
Internets, ant colonies, and brains all start small, grow steadily, and then explode into hypergrowth. In nature, all species multiply as much as resources allow. This expansion may start linearly, but it quickly becomes exponential. Populations of plants, animals, yeast, and brain cells grow unencumbered until they reach the maximum quantity that the environment can sustain, the carrying capacity of an ecosystem.
If you put one bacterium in a Petri dish with some nutrients, the bacteria population will literally double every minute until the dish is completely full and can’t grow anymore, which only takes about an hour. In the human brain, we see a rapid expansion of neurons (called neurogenesis) in utero, where our brain size peaks at around 100 billion neurons. A fetus can generate an astronomical 250,000 neurons per minute.
There is a good evolutionary reason for this, and survival often depends on it. The world is a competitive place, and the best way to stomp out potential rivals is to consume all the available resources necessary for survival. Otherwise, the risk is that someone else will come along and use those resources to grow and eventually encroach on the ones you need to survive. The same is true of technology and business: if you don’t dominate a market, you will give potential upstarts an opportunity to grow and eventually compete with you. Monopolies prevent competition, which is as good in business as it is in nature—if , that is, you are the monopoly.
Remember the early days of the internet? It started as a network of only a few connected computers, growing slowly in the early days but expanding rapidly thereafter. Around the year 2000, the number of devices connected to the internet exploded, growing to five billion within eight years. There are now more devices connected to the internet than there are people on earth.
It is in this exponential growth phase that most networks die. In biology, species are weeded out here through natural selection. Very few organisms have the fitness to ultimately hit the growth curve that ensures sustainability. In technology, 95 percent of all innovations don’t make it through this critical phase. We can look back and clearly see growth in the early days of Google, Facebook, Twitter, and Instagram. But for each of these success stories, there were myriad others that flamed out before getting anywhere close to their carrying capacity (remember Eons.com, eToys, or AltaVista?). When an environment has excess carrying capacity, competitors will inevitably rise up and seize the opportunity to steal it. Just as it does in nature, Darwinian selection somehow selects out unfit technology as well.
PHASE 2: BREAKPOINT
Networks rarely approach their limits in a measured, orderly fashion. There are two reasons for this. First, exponential growth is hard to control, even for Mother Nature. Second, networks often don’t know the carrying capacity of their environments until they’ve exceeded it. This is a characteristic of limits in general: the only way to recognize a limit is to exceed it. It is for this reason that the breakpoint of a network—the time at which it exceeds the carrying capacity of the environment—is so critical.
Think about it: the only way to know that you should really only have two drinks at the company holiday party is because last year you had four. The only way for the city to determine an appropriate speed limit is to determine the unsafe speed and then subtract a few miles per hour. How are weight limits determined on elevators? How do we know the maximum oven temperature for a pizza? Because someone has exceeded the limit at least once.
Biological networks almost always exceed their limits by growing too large for the carrying capacity of their environments. In ecology this is called “overshoot,” and it’s true in technology as well as nature. How do ants know they’ve reached their maximum suitable colony size? The colony gets a little oversized, which results in too much congestion, noise, and confusion. This is how the ants know it’s time to start sending fertile ants out of the colony to reproduce elsewhere.
The brain does a similar culling, by shedding neurons in addition to over half its connections. By the time a child is five years old, there are nearly 1,000 trillion connections. Through a process of selective pruning, the 1,000 trillion connections shrink to roughly 100 trillion by adulthood.
So in ants and brains, Phase 2 is best described as an “overshoot and pruning” or an “overshoot and collapse.” So why is the breakpoint so important? Because once you overshoot the carrying capacity, everything changes. The most important thing is to determine where the breakpoint actually resides and act accordingly. The goal is to identify the breakpoint and reduce the friction that the overshoot causes.
Carrying capacity is elastic: if you overshoot too far beyond the breakpoint, your capacity will drop proportionally in the opposite direction. In those cases, the reduction is truly a catastrophic collapse. But if you identify the breakpoint and limit the growth beyond it, the network will merely shrink back to a respectable level. While exciting to investigate, no one wants to be part of a catastrophic collapse, which often ends up being fatal for both biological and technological networks.
Consider what happened to MySpace. It grew out of control, growing from zero to 100 million accounts in three years. The average user went from a handful of friends to 200 friends, acquaintances, and complete strangers in the same time period. The navigation frames grew from just a few links to 15 in the main bar and 28 more in the services box. MySpace pages became cluttered with automatically playing songs, videos, glitzy wallpaper, and other widgets. Basically, it got congested, noisy, and too confusing to navigate—MySpace grew too far beyond its breakpoint. The graph of MySpace’s collapse looks very similar to that of the St. Matthew Island reindeer.
PHASE 3: EQUILIBRIUM
Unless there is a natural disaster, biological networks generally don’t fail in such a dramatic fashion. For that, it takes some human interference. Remember, the reindeer were brought to St. Matthew Island by people; Mother Nature never put reindeer there. And MySpace was certainly a human invention.
Ant colonies, various other animal species, brains, and internets are all networks, and as such they follow the same pattern of growth, breakpoint, and equilibrium. They start out small and grow explosively to the point where they overshoot and collapse. A successful network has only a small collapse, out of which a stronger network emerges wherein it reaches equilibrium, oscillating around an ideal size.
At the phase of equilibrium, networks continue to grow, but in terms of quality instead of quantity. When the size of a network slows, other things speed up—like communication, intelligence, and consciousness. At this point, the real magic begins.
This last network phase is poorly understood, even by biologists. We are just starting to learn about equilibriums in biological systems, let alone in technology. When Deborah Gordon first discovered these properties in ant colonies, she learned that the size of a colony remains stable, that it has no central leadership, and that it becomes intelligent at equilibrium. But no one yet knows why or how this happens.
Part of the reason is that people dismiss the notion that intelligence can come from a network. When we talk about ourselves, it is easy to call us intelligent beings. It comes naturally, even if it’s anthropocentric. But when we talk about our neurons in our brains, the discussion starts to get a bit dicey. It is hard for us to believe that the human mind could emerge out of something as simple as neuronal firings. For many people, the brain seems beyond a scientific explanation. While some may find it hard to believe, the evidence supporting the science of neurons is now irrefutable.
The science of neurons brings into question the most fundamental beliefs about intelligence. The reality is that if we are willing to accept that neurons are what make us smart, then all entities with sufficient neurons must be capable of intelligence, rationality, and even consciousness. Of course, the term “sufficient” is up for debate, as always. We don’t attribute intelligence to a sea slug with its measly 18,000 neurons or to an individual ant at 250,000. But how about a mouse at 75 million or a housecat at 1 trillion?
It is worth considering whether an ant colony, with its trillions of neurons, is deserving of our consideration. Could it be that the ant itself is a mere part, that the organism is the colony? The answer to this question has profound implications for other networks. If we accept that the brain is a smart network even though individual neurons are simpletons, and that the ant colony is intelligent even though an individual ant isn’t, we’re acknowledging that networks possess intelligence beyond the sum of their parts. And if that is the case, then the internet as a collective unit could also gain intelligence, rationality, and consciousness once it reaches equilibrium.