Deviate, p.23
Deviate, page 23
Think of the cerebellum of a bird as compared to the wiring of an airplane. The bird’s cerebellum controls its motor behavior, as does the cerebellum in mammalian brains. In the case of the bird, it enables coordinated wing movements that are essential for fighting the laws of gravity. In contrast to ground-dwelling creatures like ourselves… who, if their motor coordination fails, might fall a few feet at most, with the requisite cuts and bruises… a failure of the bird’s cerebellum would be catastrophic. As such, you might imagine it was very efficient indeed. This is true. But it’s also true that a bird’s cerebellum has a tremendous level of redundancy of connections: it is theoretically possible to remove over half of it without causing the bird to tumble to earth!80 Compare this to the most sophisticated war-plane, which is possibly one of our most efficient creations. Cut just a few wires in its control system, or damage one of its wings just a small amount, or even have a bird fly into one of its jet engines, and hundreds of millions of dollars will crash to earth. The system, like most businesses (and indeed much of modern-day life, from sports and athletes to education), is highly efficient but not at all creative, and as such is not adaptable—i.e., innovative!
To create a successful ecology of innovation we must look to our own biology and tap into our very own neural nature that balances efficiency and creativity. This is why neither always being creative nor always being efficient is the end-all route to success. The two qualities must exist in dynamic equilibrium. What is more, the system must develop.
The process of development is the process of adding dimensionality, called complexification, which my lab and others have studied for many years. It’s very intuitive: start simple (few dimensions), add complexity (more dimensions), and then refine (lose dimensions) through trial and error… and repeat. Development is the process of innovation incarnate.
In 2003, when Apple began work on its revolutionary smart-phone, they were sure that they wanted to use new “multi-touch” technology for the screen, which would radically change the user interface experience. But design of the phone itself was still unclear. One approach was called “Extrudo.” The case was made out of extruded aluminum, with a fatter middle and tapered sides. Not totally satisfied with the Extrudo, however, Jony Ive also spearheaded another approach, the “Sandwich,” which combined two halves of specially designed glass. Ive and his team were complexifying, simultaneously entertaining the possibility that both of these assumptions were good to help them more creatively discover which, in fact, wasn’t. In the end, neither turned out to be the chosen solution. Instead, Ive decided on a previous prototype that they’d had all along… a third assumption that his team wouldn’t have chosen if they hadn’t first complexified by designing the other two.81
This complexification model is effectively the Silicon Valley model, though it has its roots in biology, since development, learning, and evolution are the iterative processes. So, “innovative” Silicon Valley are companies doing nothing new. They are basically doing what slime molds and brains do all the time. There are a few important tricks and strategies to ensure that complexification succeeds, though, which David Malkin and Udi Schlessinger’s inspired Ph.D. work in my lab revealed.
Thinking back to modeling assumptions as a network, we have shown that the more complex the network, the more likely the “best” solution will exist in your search space of possibility, simply because the large number of interconnections increases the number of potential solutions (as shown in the bottom image here). This is in contrast to very simple systems that have only a few possible solutions (as shown in the top image). The problem is that complex networks don’t adapt well (this is what we call low evolvability). So while the highest peak is more likely to exist in your space of possibility if you’re a complex system, an evolutionary process is unlikely to find it, which creates quite a conundrum: complex systems can help us adapt, but aren’t very adaptable themselves.
So how do you find the best solution?
You develop.
Start with one or a few dimensions, then add more. The image here shows the progressions of the same system as it develops. David Malkin and I call it an “uber-landscape,” which is a unique idea that adds the element of time to a search space (which are usually modeled as static). Notice how the system starts off very simply… with a single peak. But as time progresses along the z axis, the number of peaks steadily increases as more elements are added to the system, until it reaches its most complex state at the other end of the plot of the graph, where there are roughly five peaks. What we found surprising is that when one grew the network in this way, the best solution in its most complex state was also the solution that had the highest probability of being discovered… when at each step the complexifying system tried to minimize its “energy state.” This is in contrast to our search for the same solution when the system started in its complex state in the first instance. It’s as if, when starting from the simple state, we were able to walk along a ridge from lower to higher dimensional space. In doing so, we were able to walk past the troughs of a lower-dimension system by adding a higher dimension, which is called extra-dimensional bypass. The findings suggest a counterintuitive, but essential, and admittedly speculative principle: Adding what we call “noise” (i.e., a random element) to a system can increase the adaptability of that system. If true, then—though highly speculative—there could be genes that are there not because they encode any particular phenotype, but because they increase the dimensionality of the search space full stop… and are then lost (or silenced) once the system settles on a new peak. Imagine what this would mean in an organization where the only function of a particular person is to increase the dimensionality of the space, thereby bridging the poorer solutions… suggesting that sometimes…
Noise in a system is good!
Having said that, it is very important when adding new dimensions that you balance the rate of periodic success with the rate of complexification… while minimizing the energy state of the system at each level of complexity. In other words, you wouldn’t want to add updates and upgrades to a fundamentally crappy cell phone; you should only do this for one that is already strong at its core. This mirrors how the brain develops. When it grows, it is adding connections, and therefore complexifying itself: adding dimensions to the space of possibility, and forging new pathways of neuroelectric connectivity. But it’s doing so according to feedback on what worked and didn’t work in the past. The internal patterns of activity that worked (the useful assumptions) are reinforced, and those that didn’t work (the no-longer-useful assumptions) are lost. As you may recall from Chapter 3, this is what your cells do in the neuromuscular junction, creating redundancies that the system then prunes away, only to then complexify again, with the result that the number of connections is actually greater overall, but organized usefully. What is more, Udi Schlessinger also demonstrated in the lab that the process of useful complexification has its own useful progression. Rather than just add a random connection, it was more useful to duplicate the system (add redundancy) with an additional small step. This process is something that deviators embody not just in their cells but in how they live their lives… complexifying their existence with new experiences that complexify their assumptions, one small step at a time… and in turn alter perceptions, not just in their minds but also in the physical world, only to then follow this process of complexification with a process of consolidation. Each update, then, is a “riff ” on the last version, not a rethinking of the whole thing. This cycle of increasing and decreasing complexity IS the innovation process that is endemic to life itself, and the underlying route to resolving its inherent conflicts.
So, if you own an iPhone or any other Apple product, there are always new modifications, new updates, and new models. This is not just about marketing as such, but a result of the questioning process that Ives pioneered through the design group at Apple—the governing idea that, even though the company has an unequivocally great and successful product, they must keep asking themselves, Why can’t we make this better? They can and do, continuing to unloose new little “jokes,” as embodied by each new iteration of the iPhone that improves on the last.
In addition to brain development, evolution is also a great example of the successful tension between these two poles. All species go through exploration (creative) periods, in which different traits appear, like feet in the first organisms that left water billions of years ago, or like the opposable thumbs that developed in some primates, including humans. There have been periods of massive expansion of the diversity of species (increasing dimensionality), followed by the exploitation (efficiency) period in which those with useful new traits survive (decreasing dimensionality) while others are selected out, leaving only the most robust organisms.
We can also see this cycle reflected in the movement of Earth in relation to the Sun, and how humans evolved to be in synch with their environment. If our species’ existence is itself an ecology of innovation, then the balance between wakefulness and sleep is at the heart of the process, where wakefulness creates connections and sleep consolidates them. The same goes for our development, first as babies and then as children. When we are young, this critical period for our brain is plastic… as it’s establishing many of its future assumptions and attractor states. This is the tremendous advantage of our species being born “too early,” as small beings that can’t survive on our own for many years, with brains that are underdeveloped. Because of this, the human brain adapts itself to the environment in which it happens to find itself. It is for this reason that humankind, more than any other species, has been able to occupy such a diversity of niches in nature. Then, as we age, this formative period slows and change becomes more difficult, as our repeated experience forms deeper attractor-like states (though of course we can still creatively change… that’s why you’re reading this book!). Yet in allowing you to be creative, these “detours” enable you walk along ridges around the valleys between different peaks in a fitness landscape. Once you end up on the higher peak, you can discard that extra dimension and swing back toward the efficiency side of the paradigm. Innovation (adaptation), then, is like a spiral: when you’re in the cycle usefully, you never return to where you were. Instead, you return to a similar but higher place.
A useful metaphor is the tonic scale in music. When you’re at middle C, as you move up through notes D, E, and F, you are moving further “away” from C in that the signal is shifted to higher frequencies. This is akin to the creative phase. But as you continue still further up the tonic scale, through G, A, and B, you start moving perceptually “back toward” C, but one octave higher. This spiral nature of innovation illustrates that creativity—ironically—is a key route to efficiency, and efficiency can lead to creativity. Like all things in nature, it’s the movement between them that matters in order to create an ongoing process adaptable to a changing world.
This is why in business and other cultures that attempt to foster adaptation (which should include universities), the key is to never be “phase-locked,” wherein one pursues a single idea with maximum efficiency without movement. In nature, as in business, such systems are quickly selected out of the population. The app called Yo was a single, highly specialized and very efficient messaging system that reduced communication to a single word: “YO!” The significance of the message varied with context. It took the founder only eight hours to code, yet it quickly won over hundreds of thousands of users. Nothing could compete with it. It was Malevich’s White on White of messaging platforms, taking the genre to its ultimate conclusion. But had it not complexified, it is likely to have quickly died, or even more likely to have been subsumed by another, more general platform. Indeed, the most successful companies follow what I call wedge innovation. They start off “sharp,” with a highly focused product that enables them to outcompete others in their space (by being efficient with an original, creative idea with which they started). But then they widen their use-case and thus the company’s contextuality, increasing what they do and why they do it. Kodak is the classic example of a company that did not do this and died because they didn’t expand their wedge (the pioneering of print photography for everyday people) to start a further twist to its spiral by expanding into digital photography before their competitors, although they had the opportunity to do so. Apple is an example of the opposite… a spiraling, forward-thinking company.
As a company grows (and indeed as life grows), it needs to have different spirals simultaneously underway with a variety of products at a variety of time-scales (frequencies). For example, when Google reaches the efficiency phase of one of their products… that is, it works well, like the Google search engine or the Chromebook laptop… the company is already in the beginning phases of other innovations, like self-driving cars. This could be considered the point of GoogleX, Googe’s moonshot research and development initiative, which is to create the creative ideas that others then make efficient (which I’d suggest is the same process that a lab follows… as exemplified by Bob Full above). The result is that some ideas will become innovative and others will die, like Google Glass .. not because of the technology, but because Google didn’t seem to consider one fundamental attribute in the design: human perception/nature. Google doesn’t appear to have taken into account the importance of eye-gaze in communication. We move our eyes not only to obtain information but also to signal our state of mood and the nature of the relationship in which we are engaged. For instance, if I’m looking at you and you look down, you may be signaling a sense of subservience or insecurity. If you look up, you can be signaling just the opposite.
The fundamental, biologically inspired point is this: adapt or die.
In business and in life we must always be in a cycle, while constantly being aware of an equation Bob Full uses to evaluate the interplay of efficiency and creativity in his lab. “The probability of success of the project,” he says, “is the value of the project divided by how long it takes. You’ve got to balance those things.” To ensure that a high probability comes out of that equation, a simple model emerges: One must lead with creativity and follow with efficiency… and repeat, not the other way around and not in tandem (unless via different groups).
“Blue-sky thinking” is important at the beginning of a project, long before the deadline looms, just as gene mutations are important for a species before a life-or-death scramble for resources occurs, and just as Picasso did hundreds of studies for his masterpiece Guernica before ultimately painting the final canvas. If you start with efficiency, however, you confine your space of possibility before you have had the optimal amount of time to rearrange it repeatedly to reveal the best available ideas. This is why editors don’t edit every sentence as a writer works, but read chapters or the whole book after the writer has had his or her exploratory period. (The best editors help writers find a balance between creativity and efficiency.) This is why Goethe wouldn’t have burned a whole twenty years on his obsession with color if he had given himself a limit of, say, ten years to organize his thoughts on the subject, instead of the open-ended timeframe he had for his overly exploratory book.
Alternating cycles of creativity and efficiency are what the most successful living systems did (and still do), and this is the process that Silicon Valley/tech culture has effectively rebranded as their own. The key for making this work is figuring out when to analyze feedback from the ecology. Companies with quick feedback iterations for apps that have potential but are far from perfect… like the first versions of the dating app Tinder, the traffic app Waze, or the real estate app Red Fin… have an advantage in being able to “code-switch” rapidly from efficiency to creativity and back again, and thus best serve their product and what their consumers are telling them. They are empirically seeking the best solutions and exploring the peaks and valleys of their fitness landscape of ideas. As such, a collateral necessity of this process, according to this way of framing things, is “failure.” The success of certain startups that have embraced this approach has led to all the recent hype around catchphrases like fail forward and fail better. These make for good mottos, yet their essence has been around since long before the tech community. They’ve discovered science.82
But no failure is actually taking place in such a model if done well. In science as in business—or as it should be in business, though the issue of revenue is always present—failure is when you don’t learn anything, not when you disprove a hypothesis (which means you learned). So, for instance, the Silicon Valley mottos—Fail Forward is either wrong or, well, wrong: Either it’s wrong because there is no failure in a properly designed experiment, and thus the very idea of failing is the wrong approach to the craft of science; or, Silicon Valley is wrong in its concept of failure, since to move forward in science is to learn and therefore isn’t a failure at all. This view is at the heart of any and all ecologies of innovation. So maybe Silicon Valley needs some new mottos: Learn… Move Forward… Experiment Well… (not nearly as catchy).
What modern tech culture does understand very well is that the process that produces true innovation, success, and emotional fulfillment is not seamless. There are conflicts. There are delays. There are errors that have consequences, like the painful falls Ben Underwood took while he was learning how to “see” again. These disheartening bumps feel awful (and this should be acknowledged as such). But conflict can lead to positive change (if in the right ecology), since a true ecology of innovation shouldn’t produce seamless results, and the brain itself again tells us why this is.
