The Algorithm We Call Money: How Digital Technology Renders Our Oldest Coordination Systems Obsolete
When people hear "algorithm," they picture code, silicon, server farms. But strip the definition down: an algorithm is a formal procedure for solving a problem under constraints. That's it.
Now strip economics down to its bones: the systematic allocation of scarce resources under conditions of incomplete information.
These are not analogies. They are descriptions of the same underlying formal reality, arrived at from different directions. The largest optimization problems our civilization has ever run have been economic ones. I'm not speaking metaphorically. I'm saying computer science and economics share the same soul — and once you see that, the history of human coordination looks completely different.
Two Architectures, One Problem
Early network engineers faced a foundational design choice:
Circuit switching: A central authority pre-allocates a dedicated connection before communication begins. Orderly. Predictable. Catastrophically fragile if the central node fails.
Packet switching: Each data unit carries its own addressing information and navigates independently through available nodes. Messy. Decentralized. Robustly resilient.
Human societies — with zero knowledge of these engineering terms — built both.
- Money is packet routing. Every unit of currency carries embedded price signals that allow decentralized agents to make local allocation decisions without any global coordinator needing the full picture.
- Bureaucracy is circuit switching. It pre-allocates authority, establishes dedicated procedural pathways, and requires central coordination to function.
Neither alone is sufficient. The market cannot build a legal system. The state cannot price ten thousand commodities simultaneously. We have always run both architectures in parallel — not from ideological confusion, but from genuine architectural necessity.
The Token-Ring Failure Mode
In a token-ring network, all nodes share a single communication medium. A special authorization token circulates continuously. Any node wishing to transmit must:
- Acquire the token
- Transmit
- Release the token back into circulation
The system functions correctly — until one node seizes the token and refuses to release it. Every other node falls silent. The medium does not degrade gradually. It collapses.
This is what extreme wealth concentration actually is. Not primarily a moral failure (though it is that too) — but a technical failure mode.
The monetary signaling function — the capacity of price systems to coordinate behavior by transmitting information about relative scarcity and value — depends entirely on tokens circulating. A seized token is not a successful accumulation strategy within the system. It is the system's breakdown. The network doesn't slow down. It stops talking.
The Devil's Advocate: Can Negative Money Save the Signal?
The strongest objection runs like this:
Even if one node captures all existing tokens, we can extend the number line. Issue debt. Create credit. Generate new tokens through financial instruments and restore the signaling capacity of the system — without redistributing anything.
It sounds elegant. It nearly works.
But David Graeber's anthropological research disturbs this solution at its foundation. His argument:
- Debt is not an extension of money — it is historically anterior to it
- Credit relationships preceded coinage
- Debt carries enforcement mechanisms that monetary exchange does not: obligation, coercion, moral condemnation of the defaulter
When you solve the token-hoarding problem by expanding into negative money, you are not patching the monetary algorithm. You are quietly replacing a decentralized signaling system with something older, more personal, and considerably more coercive.
The architecture has changed. You just haven't updated the documentation.
When Constraints Dissolve: The Core Problem
Money and bureaucracy were never ideal systems in any absolute sense. They were optimal solutions to a specific constraint: the prohibitively high cost of storing, transferring, and processing information at scale.
Given those costs:
- Decentralized price signals were a genuinely brilliant design
- Standardized procedural authority was a genuinely brilliant design
- Both solved the problem that actually existed at the time
But digital technology is driving information costs toward zero.
When that happens, the optimization problem does not get easier. It changes entirely.
Our existing solutions are no longer solving the problem we have. They are solving the problem we used to have. This is the precise situation of a brilliantly engineered steam engine in a world that has discovered electricity. The engine is not broken. The constraint it was built to address — the need for heat to generate motion — has simply dissolved. The engine's excellence is now irrelevant.
The Right Question to Ask
Once you recognize money and bureaucracy as historically contingent algorithmic solutions — rather than natural features of human social life — a genuinely different question becomes available.
Not:
- How do we reform the tax code?
- How do we streamline regulatory compliance?
But: What coordination algorithms would we design from first principles if we began today?
This reframing is not utopian. It is engineering.
It shifts the conversation from:
- Political argument about redistribution and deregulation — conducted entirely within the assumptions of the old system
To:
- An architectural question about what problems human coordination actually needs to solve when near-zero information costs remove the constraints that shaped every instrument we currently possess
That question has not been seriously asked. It should be the central intellectual project of our moment.
The Compiler Has Changed
For six thousand years, humanity has been running two great algorithms — money and bureaucracy — without ever realizing they were algorithms at all.
The question before us is not whether to fix them. It is whether we possess the intellectual courage to recognize that we are the programmers, the algorithms are ours, and the compiler has changed.
The codebase is not sacred. It is legacy software. And legacy software, however brilliantly written for its original environment, does not get a pass simply because it once ran well.
Top comments (0)