19 世紀的作業系統：為何金錢與官僚制度無法解決文明問題，以及必須由甚麼取代

The 19th-Century Operating System: Why Money and Bureaucracy Cannot Solve Civilizational Problems and What Must Replace Them

For six thousand years, humanity has run its civilization on two orchestration algorithms—money and bureaucracy. The crises now threatening our species are not failures of will or wisdom. They are inevitable outputs of software that was never designed to handle them.

This is a systems engineering problem. And I want to make the case for why that reframe matters.

The Scaling Problem

Human coordination mechanisms evolved to work within a narrow bandwidth:

• Family loyalty
• Reciprocal reputation
• Institutionalized moral codes functioning as behavioral firmware

These are elegant solutions to the cooperation problem. But they operate reliably only within roughly 150 people—the cognitive horizon we call Dunbar's number.

Once communities scaled beyond that threshold, something had to replace the intimacy of mutual recognition. Two orchestration layers emerged, persisted, and came to govern nearly all of human civilization: money and bureaucracy.

Each is a distinct engineering answer to the same foundational question: How do strangers cooperate at scale?

And each, I want to argue, is broken in ways that are not correctable from within.

Money as a Lossy Compression Algorithm

Think about what a price actually is. It is a single integer. One number.

Into that number we have compressed an almost infinite set of dimensions:

• Carbon emitted in production
• Labor conditions across the entire supply chain
• Ecological impact of extraction
• Historical provenance of materials
• Power asymmetries shaping every transaction along the way

A price discards all of this.

That is not a correctable bug. It is a structural feature of any token-based, single-dimensional exchange system. The informational poverty of price is not something market reform can address, because no market reform can restore dimensions that the algorithm was never built to carry.

We are not dealing with a misconfigured system. We are dealing with a system doing exactly what it was designed to do—and that design is catastrophically inadequate to the complexity we now face.

The Hayekian Paradox

Friedrich Hayek gave us the most sophisticated defense of decentralized financial markets ever constructed. His argument:

• Prices aggregate and transmit distributed local knowledge
• No central planner could replicate this efficiency
• The system is self-correcting because it draws on the intelligence of every participant

I find this argument genuinely compelling. And I think it collapses under examination of money's actual informational architecture.

• Prices update slowly
• They transmit one dimension
• They have one-hop visibility

The dynamic, granular, multi-signal self-correction Hayek theorized requires an information channel rich enough to carry it. Price is not that channel.

The Anmaßung des Wissens—the pretension of knowledge—that Hayek attributed exclusively to central planners is equally a structural property of a system that can only ever see one number at a time.

Hayek was right about the problem. He was wrong about whether money solves it.

Compound Interest and the Death Spiral of Circulation

Money has a second fatal failure mode. This one is arithmetic rather than informational.

Compound interest and transaction asymmetry mathematically guarantee that tokens accumulate with a shrinking fraction of actors. A token hoarded by one billionaire circulates zero knowledge for the remaining eight billion people.

This means money progressively destroys the very information-circulation network that constitutes its only theoretical virtue. The distributed knowledge system Hayek celebrated is converted, through the ordinary operation of its own arithmetic, into an oligarchic one.

The mechanism eats itself.

No tax policy or redistributive program has ever reversed this dynamic at scale, because they operate within the system's logic rather than redesigning it.

Bureaucracy and the Centralization Trap

Bureaucracy occupies the opposite pole. It was engineered to serve the common good rather than particular interests—a genuinely different normative orientation.

But it purchases that orientation at the price of the same informational failure Hayek identified in central planning:

• Hierarchical structures cannot aggregate fast-moving, locally embedded, granular knowledge
• They are structurally slow, rigid, and blind to the complexity they are tasked with governing

Bureaucracy is not corrupt by nature. It is knowledge-impaired by design.

Good intentions encoded into a centralized architecture still produce the coordination failures we observe everywhere—from climate policy to pandemic response to supply chain regulation.

The Coordination Matrix

Map these two systems onto a two-axis matrix:

Money occupies the decentralized-particular-interest quadrant. Bureaucracy occupies the centralized-common-good quadrant.

The upper-right quadrant—decentralized and oriented toward the common good—is structurally empty.

This vacancy is not a policy failure or a historical accident. It is a geometric consequence of the engineering choices baked into both systems. No combination of market incentives and regulatory oversight can fill that quadrant, because the architecture of each system actively prevents it.

We have never built anything that lives there.

The Digital Inverse of Keynes

Keynesian economics understood that demand creation is the engine of financial-market stability.

Digitalization operates as its structural inverse.

When artificial intelligence and digital platforms satisfy human needs at near-zero marginal cost, they eliminate the economic transactions that would otherwise circulate money. They shrink the demand base on which financial markets depend.

This is not a temporary disruption to be absorbed by:

• Retraining workers
• Adjusting interest rates
• Incremental regulatory updates

It is a deflationary dynamic baked into the logic of digitalization itself, and the existing operating system has no mechanism to redirect it.

We are running 19th-century software on 21st-century hardware. The incompatibility is becoming catastrophic.

Civilizational Crisis as Operating System Failure

Climate change, extreme inequality, opaque supply chain ethics—these are not the results of human negligence or bad values.

They are intrinsic and predictable failure modes of an operating system that:

• Encodes no carbon dimension
• Mathematically concentrates wealth
• Compresses entire supply chains into a single price

Recognizing them as OS failures rather than moral failures is the necessary reframe.

It shifts the solution space from behavioral exhortation—consume less, care more, vote correctly—to systems engineering.

We do not fix a memory leak by asking the program to try harder. We rewrite the allocation logic.

Toward a Digital Orchestration Layer

A new orchestration layer must be designed to occupy the vacant upper-right quadrant of the coordination matrix.

Design requirements:

• Combine the decentralization that gives money its informational reach
• With a multidimensional encoding capacity that price has never possessed

Tokens in such a system would carry as native, irreducible dimensions—not optional metadata:

• Carbon provenance
• Labor conditions
• Reputation signals
• Common-good metrics
• Exchange value

Consider the token-ring network as an analogy: a protocol designed to ensure that information circulates to all nodes rather than concentrating at a few, with architectural guarantees against hoarding built into the circulation mechanism itself.

This is not utopianism. It is an engineering specification.

The question is whether we are willing to treat it as one.

Conclusion

The question before civilization is not whether to reform money and bureaucracy.

It is whether we possess the intellectual courage to recognize them as legacy code—and to architect, for the first time in six thousand years, an orchestration layer equal to the complexity of the problems we have created.

That is a systems design problem. And systems design problems have solutions.