Bhumics 16: The System of the World

We live inside meshes of ideas, ideals and institutions. Whenever you open a math textbook, text a friend, watch a rocket launch, or read climate news, you are stepping into a human-made “system”—a framework that helps us explain, predict, and change the world. Some of these systems chase flawless understanding; others chase practical power.

In this essay I call the first kind perfectionist and the second productionist. They have argued with one another for centuries, yet they now find themselves entangled in a single, urgent project: keeping Earth livable in the twenty-first century. To see why, we must follow both strands of thought to their planetary consequences.

1. The Dream of Perfection

Picture Newton gazing at the night sky a couple of nights before the apple falls on his head. He wants more than a chart of planetary positions; he wants the exact rule that explains every orbit forever. The search for one ruling principle (or a few such) is the heart of perfectionism. Mathematics embodies it most clearly. Euclid’s Elements begins with a few definitions and builds hundreds of theorems as logically inevitable once you accept the axioms. More than two thousand years later, modern logicians tried to push that style to an extreme: could all mathematical truths flow from a small set of ironclad axioms?

Sadly (happily?), Kurt Gödel shocked everyone by proving that any rich enough system of axioms will always contain true statements that cannot be proved within the system itself. Physics pursues a similar ideal. Isaac Newton’s three laws once seemed sufficient to explain the motion of every planet, cannonball, and ocean tide. Later, Einstein’s relativity and the great search for “symmetry principles” promised an even deeper unity: perhaps space, time, and matter could be reduced to one grand unified theory. Of late, it looks like that too is a false hope.

Or even more importantly: it’s no longer a desirable goal.

Perfectionist thinking is enormously powerful. It gifts us with clarity, beauty, and a shared language of proof. But it also tempts us to believe that understanding automatically earns control—that once you know the law, the world must make it so.

Not so soon.

2. The Rise of Production

Imagine the skyline of a modern city at night: glowing office towers, electric trains, satellite dishes on every roof. Or, for that matter, the remnants of the swamp on which the city was built - a snake here, a badger there. Neither sprang from a single elegant principle. They came from workshops, factories, budgets, and deadlines in the first instance and burrowing creatures in the second - the productionist world. Productionism creates and sustains order, and it’s not as if elegance is missing, but that elegance isn’t constrained by perfection. Instead, productionism also cares about the following question: does it work?

A Roman engineer did not need the finest theory of arches to build the aqueduct; he needed stone, labor, and practical geometry. Today swarms of scientists and engineers design artificial-intelligence models with billions of parameters they cannot completely interpret. Function outruns theory. The power of productionism is obvious. It put footprints on the Moon, mapped the human genome, and fills your pocket with a device that connects you to every perfectionist who has ever existed.

It also carries risks.

Carbon dioxide from burning coal and oil, by products of production, has altered the planet’s climate. Plastic waste choke oceans; poorly regulated opioids and synthetic drugs fuel addiction crises. We might say that productionist systems both threaten the living Earth and reveal those threats to us through satellites, sensors, and climate models. They generate problems at continental scale and hand us the instruments to diagnose them.

Note how productionist anthropocentrism - manifesting as the Anthropocene - is a very different beast from perfectionist anthropocentrism. Bhumics is the antidote to the former.

Perfection seeks certainty; production satisfices. Perfection wants a single story; production overflows with tangled stories that keep changing. Learning to speak both languages is crucial.

3. Cybernetics: A First Marriage of the Two Traditions

In the middle of the twentieth century a new field tried to fuse perfectionist rigor with productionist utility. It was called cybernetics, from the Greek word for “steersman.” Cybernetic thinkers asked how any system—whether a steam engine, a brain, or an economy—could use feedback to keep itself on course. Thermostats adjust heat; animals adjust blood pressure; corporations adjust prices. The underlying mathematics was clean and general, yet the applications were wonderfully messy. Early cyberneticians helped invent radar, guided missiles, and the first generation of computers.

Bhumics sidenote: James Lovelock’s Gaia theory is very much a product of cybernetic thinking

For a time it seemed that cybernetics might offer a single map for all complex systems, natural or artificial. But the map kept fracturing as each domain sprouted special details—genes, neurons, silicon gates, international trade laws. By the 1970s cybernetics had split into many successor fields: control theory, cognitive science, computer engineering, ecology. Its unifying promise faded, yet its core insight—feedback rules everything—quietly seeped into them all.

Cybernetics was succeeded in the 90s by Complex Systems Theory, which was physics dominated, therefore perfectionist in its aims.

Today every major productionist enterprise, from wildfire prediction to stock-market automation, relies on feedback loops that cybernetics once charted. Yet the scale has exploded. No feedback diagram drawn on a blackboard can capture the planetary tangles of greenhouse gases, financial derivatives, and social-media sentiment. We therefore need an updated art of steering—something that honors cybernetic roots but is honest about Earth-wide complexity.

Enter Bhumics.

Bhumics asks two big questions:

1. How do we run powerful production systems while keeping the planet’s life-support systems intact?

2. Even better: can we create production systems that center the flourishing needs of all life?

4. Family Resemblances in the Anthropocene

To answer these questions we must first admit that modern production systems do not share a single “secret ingredient.” Ludwig Wittgenstein’s idea of a family resemblance¹ helps here. Think about the category “game.” Chess, football, hide-and-seek—no single feature belongs to all of them, yet you can still tell they belong together through overlapping likenesses. The same is true for the sprawling networks that define the Anthropocene, the age in which human activity is Earth’s dominant geological force. Electric grids, shipping lanes, cloud-computing centers, drug cartels, global seed banks, cryptocurrency exchanges: each has its own jargon and technology, but they overlap in feedback loops, flows of energy, and webs of legal or illegal trade.

Trying to compress that variety into one elegant law would revive the old perfectionist mistake. I follow the methodology of mathematician Alexander Grothendieck, who likened understanding to a “rising sea” that slowly surrounds a stubborn rock. You do not crack the rock with a single chisel blow; you let water infiltrate its fissures until it loosens and splits. Likewise, to comprehend a mega-system you gather many small, precise observations and let them soak together until patterns emerge. Often the wide sea explains the narrow rock, reversing Occam’s Razor², which claims the simplest explanation is best. In the Anthropocene, complexity can be the key that unlocks apparent simplicity. A single hurricane’s path, for instance, makes sense only when viewed inside sprawling ocean-heat patterns, jet streams, and urban land-use policies.

5. The AI Lifecycle: A Productionist Showcase

No current system illustrates these themes more vividly than modern Artificial Intelligence. In its early decades AI was perfectionist: if we could just formalize reasoning as logical symbols, we might solve intelligence once and for all. Didn’t work - the Frame Problem came in the way. The breakthroughs of the 2010s and 2020s did not arrive from faultless logic but from messy data, massive computer chips, and endless trial-and-error training. Today’s large language models consume terabytes of text, megawatts of electricity, and warehouses of cooling equipment. They succeed not because we perfectly understand them but because we know how to build and adjust gigantic feedback loops around them.

To grasp AI at planetary scale, you need at least seven overlapping lenses:

1. Environmental Assessment. From mining cobalt in the Democratic Republic of Congo to disposing of obsolete servers in Arizona, every step leaves a carbon, social, and toxic footprint.

2. Materials Science. All intelligence begins with sand: silicon dioxide purified and etched into transistors. Tiny changes in crystal structure can decide whether a chip sips power or guzzles it.

3. Computer Architecture. Graphics Processing Units (GPUs) arrange billions of transistors to multiply matrices at lightning speed. Network topologies inside data centers decide how quickly those GPUs share results.

4. Mathematics of Uncertainty. Linear algebra, probability, and optimization choreograph how weights in a neural network shift during training.

5. Global Value Chains. A single server may contain parts designed in California, fabricated in Taiwan, assembled in Malaysia, and sold by a Dutch company to a U.S. cloud giant leasing space in Dublin.

6. Energy and Thermodynamics. Each chat query you submit releases a puff of heat somewhere; multiply by billions and you get the electricity demand of a mid-size nation.

7. Geopolitics and Policy. Export controls on advanced chips, data-privacy laws, and international standards for AI safety reshape who can innovate and who must wait.

No one human can master all seven. Pretending otherwise would lure us back into perfectionist hubris. The wiser path is to map how the seven domains are interconnected. An energy analyst might reveal how renewable power shortages constrain data-center growth; a policy scholar might show how that bottleneck pushes companies to lobby for looser emissions rules; a computer scientist might trace how relaxed rules accelerate chip upgrades and therefore speed up model sizes, which loops back to energy demand again. There’s a special role for the synthetic philosopher in this game, where I am adopting Eric Schleisser’s definition:

By ‘synthetic philosophy’ I mean a style of philosophy that brings together insights, knowledge, and arguments from the special sciences with the aim to offer a coherent account of complex systems and connect these to a wider culture or other philosophical projects (or both). Synthetic philosophy may, in turn, generate new research in the special sciences, a new science connected to the framework adopted in the synthetic philosophy, or new projects in philosophy.

7. The System of the World

Bhumics, as a sketch of a future discipline, proposes three guiding practices:

1. Describe³ first, explain second. Gather thick descriptions of how systems actually operate. Observe without rushing to “why.”

2. Respect family resemblances. When two systems share only partial overlap—say, illegal drug networks and legal pharmaceutical supply chains—resist the urge to squeeze them into one box. Instead, list their overlaps and differences explicitly.

3. Embrace Planetarity. We need feedback loops that keep forests standing, oceans cool, and social trust intact. Building those loops is an engineering challenge, a policy challenge, and a moral challenge at once. But those feedback loops aren’t enough; there also has to be a transcendental understanding of the planet as the condition of possibility of our existence. Every anthropogenic intervention has to keep the planet in mind.

1. Wittgenstein's concept of a 'family resemblance category' challenges the traditional notion that all members of a category must share a single, essential feature. Instead, he argues that many concepts-such as "game"-are united not by one common property, but by a network of overlapping similarities and relationships among their members.

2. Occam's Razor is a principle that suggests when faced with competing explanations for the same phenomenon, the simplest one-requiring the fewest assumptions-should be preferred. It is also known as the "law of parsimony."

3. 'Describe' is a term of art here, with echoes of the descriptive method in Phenomenology, as opposed to the explanatory method in the natural sciences. Husserlian phenomenology is characterized as descriptive rather than explanatory because it focuses on systematically analyzing and articulating the structures of consciousness and experience as they are directly perceived, without seeking causal or material explanations.