“If every error has to be caught, explained and corrected, a system of the complexity of the living organism would not run for a millisecond.” — John Von Neumann in Theory of Self-Producing Automata
A very insightful question you should ask yourself is why hasn’t biological evolution invented the wheel? The reason why this is insightful because it puts into focus the difference between biological innovation and human technological innovation. Although we can recognize the similarities between evolution and technological innovation, we often overlook the differences. We should seek to understand both aspects and in so doing gain a deeper and richer understanding of how to build future advanced robust and sustainable organic systems.
Despite the rich diversity found in biology, we find only a single instance of the wheel that has been invented. The one instance found of a rotary system can be found in Flagellum. There are also instances of other species that roll around like balls (i.e. tumbleweeds, armadillos). The wheel, however, needs to move frictionlessly around its axis, in other words, it isn’t physically attached to its body. An organism needs to be attached to their appendages to provide a constant supply of nutrients. Shells, however, are examples of hardened structures that are grown biologically and eventually become detachable.
One other plausible reason is that the natural environment has very few flat structures to provide motivation for the invention of the wheel. Biological evolution works by constantly learning and adapting to existing environmental challenges. Biological innovation in analogous to the kind of evolution we find in agile development processes where the development of a minimal viable product (MVP) is driven by incremental feedback (i.e. challenges) from customers:
Biological evolution requires wheels to be useful for fitness to begin with. Then it incrementally evolves to make that wheel more useful. This would require that a wheel is from the very beginning detached, that there exists an environment that encourages the growth of much larger wheels, and that perfect symmetry is always useful. The analogy with agile development is that there aren’t any users that are in need of a small wheel. However, it is important to note that biology frequently employs exaptation to borrow designs that are invented elsewhere. Meaning the wheel could be invented elsewhere, perhaps biology millions of years in the future when humans no longer exist would adopt wheels.
The inability of biology to invent the wheel reveals to us the difference in the discovery method of natural evolution versus technological innovation. Technological innovation is driven by design in conceptual space, while natural evolution is constrained by the limits of physical space. A wheel is not only detached physically but it is also detached evolutionarily in its immediate usefulness. As the MVP diagram above illustrates, wheels are of no value unless the entire car is built. A baby in the womb is not constructed part by independent part but rather relives the evolutionary process by transitioning from more primitive organisms into a final human organism. Biology grows solutions while technology configures solutions together. This difference is depicted best by human designs that employ methods similar to biology:
This method of growing solutions implies that biology may not take the most efficient route. Biology carries plenty of baggage that may have been of use before but is not of any use in the final solution. In contrast, human technology is optimized to employ what is essential to a solution. However, this optimization process is also what makes it less robust.
The smallest cell has evolved to constantly pursue survival. According to Richard Dawkins, it is the gene that is selfish and not the actual organism. We see this collective survival behavior is ubiquitous, we can find examples in bacteria, in bees and in the behavior of cells that are part of a multi-cellular being like us. This is a consequence of specialization. In complex biological systems that consist of the composition of collections on nano-intelligent biological agents, effective functioning of the collection requires specialization of the capabilities of its constituents.
We can make an analogy with human societies. In a small hunter-gatherer tribe, there is already some specialization between the men and women of the tribe. This specialization increases once civilization emerges. In civilization, not everyone is a farmer (the producer of food). But rather, through the invention of commerce, human specialization becomes a sustainable activity. Money becomes that decentralized coordination mechanism that allows more efficient markets and leads to the viability of specialized roles. Money is a purely informational and abstract thing that binds present-day human civilization together. It’s not very romantic, unfortunate but it is true.
Actually, if we explore the origin of money, we begin to realize that it has its origins in the concept of debt. Debt is, in fact, an informational thing that codifies ongoing human relationships. In fact, most religions revolve around the concept of debt. The key message about religion revolves around paying one’s debt or about freeing oneself of debt. The difference between religions aligns with the dimension of centralization versus decentralization. On the centralized end, there is monotheism and at the decentralized end, there is panpsychism. Centralization demands conformity, while decentralization demands diversity.
Biology is massively decentralized by its very nature. Cells of all kinds are responsible for their own ongoing preservation. A living cell contains sufficient manufacturing complexity to take inputs to its environment and to generate the energy to sustain itself. Craig Venter’s team has created the first minimal synthetic cell (473-genes design known as JCVI-syn3.0). This cell was constructed using CRISPR technology by taking an existing biological cell and systematically removing all genes until only a minimal set of genes were left to sustain it functionally. Even with this minimal set, the system’s complexity is surprisingly beyond current scientific understanding, Craig Venter remarks:
We expected that maybe 5 percent of the genes would be of unknown function. We weren’t ready for 30 percent. I would have lost a very big bet.
Contrast this rugged individualism of cells with the brittle dependency of human invention. A car is a good example that reveals the brittleness of human technology. The car needs the energy to be of any use. To get that energy (i.e. fuel) it needs to be accessed from a vending point (i.e. a fuel station). That vending point isn’t an unlimited source of fuel, but rather a deposit that must be continually refreshed through a complex supply chain. That supply chain consists of gasoline container trucks that redistribute fuel from much larger depots. These depots hold gasoline which is a byproduct of crude oil. The inputs of this system come from oil refineries that process crude oil into usable fuel. Crude oil is delivered in container ships from the many sources around the world. Crude oil is extracted from the ground in oil fields consisting of many wells. So for a car to get its fuel, there is a massive complex supply chain that needs to exist. A biological cell has its own fuel manufacturing capabilities that can take raw material and convert it into usable fuel. The entire fuel refinement process is contained in a single cell.
Biology is massively decentralized, every cell has a high degree of energy autonomy. Because there is no-shared infrastructure, every component in a decentralized system requires its own infrastructure. Biology has massive redundancy. Biology makes the trade for robustness over global efficiency. This trade-off has one very important side-effect, it leads to more sustainable systems.
The waste products of biological systems are constantly recycled within an ecology. Kate Raworth describes Doughnut economics, an economic system that is more in tune with the biosphere:
Biological evolution intrinsically is aware of a very different kind of wheel. Actually, one that is self-perpetuating, biology is shaped in a sustainable cycle. The essence of the usefulness of a wheel is that it is self-repeating. In simpler terms, stuff that runs in a cycle. Evolution has discovered a more abstract kind of a wheel.
Natural evolution makes progress by solving environmental challenges in an incremental manner. Technological innovation, by contrast, is able to ignore the physical environment to create more efficient solutions. These solutions, like the wheel, include the re-invention of the physical environment. The wheel’s efficiency and robustness is a consequence of the assumption of flat roads. Human designs are bounded by the assumptions that a designer makes. These designs are solutions for a narrow domain. However, this leads to a lack of flexibility in unexpected scenarios. Human technology is brittle to black swan events.
This inability to design systems in an organic manner has lead to the intractable complexity and rigidness of existing information technology (IT). Large organizations, such as nation-states and militaries, are burdened with antiquated and inflexible IT systems. The operating systems that are run on our computers and handheld devices are exposed to an ever increasing number of vulnerabilities. IT infrastructure is crumbling as we speak and is in need of a complete overhaul. Unfortunately, the form of that overhaul is beyond present-day imagination. Present day human technology is systemically fragile and aren’t meant to adaptively react to the complexities of an ever-increasing complex environment.
He calls this “living computation”, where best effort computation is favored over correctness.
Dave Ackley (Well known for Boltzmann machines) proposes “Robust first computing” as a solution to this brittleness. Ackley challenges the underlying assumptions of conventional design which he calls “the iron grip of correctness”. He argues for a very different architecture that is not based on strict correctness but rather is inspired by the messy designs found in biology. Strict correctness maximizes the consequences of failure. Ackley argues for an alternative architecture that favors “strict indefinite scalability” where you have to give up on global hardware determinism. He calls this “living computation”, where best effort computation is favored over correctness. Correctness becomes more like an ‘ility’ than a requirement. His architecture flips the design goals such that systems must be as robust as possible, followed by correctness and then efficient as necessary.
What is desperately needed is technology that supports organic intelligence. Fortunately, we are seeing the primordial emergence of such technology. This technology is known as Deep Learning. Deep Learning captures the robustness of biological evolution and combines it with the creativeness of virtual design exploration. The wheel gives us a hint of the nuance of biological innovation. This nuance was circumvented by human minds that pursued repeatability and correctness in their inventions. Similarly, biological evolution may not invent Von-Neumann based computation that favors strict correctness. Symbolic computation for biological minds is implemented in a radically different way than in computers (see: Semiotics). Biological evolution on its own did not design a wheel because it did not seek the correctness and efficiency that a wheel requires.