Three recent announcements circulated with the same data. Microsoft canceled its contract with Claude Code. An NVIDIA executive declared that computing is now worth more than employees. Goldman Sachs projected a twenty-four-fold increase in token consumption by two thousand thirty. From that data emerged three distinct readings. Some see efficiency. Others see displacement. Others see investment opportunity. None are technical. All three are political. And that difference matters more than it appears.

The question they activate isn't new. When wage labor stops organizing the social fabric, who pays whom, under what terms? Thomas Paine posed it in seventeen ninety-six. Milton Friedman reformulated it in nineteen sixty-two. Sam Altman reformulated it again in two thousand twenty-five. Two centuries and three reformulations later, there are three answers worth taking seriously.

The first is the most familiar: universal basic income. If artificial intelligence eliminates wages as a distribution mechanism, the state or the corporation pays a direct transfer. The logic seems clean. It solves the money, but doesn't solve the power. The OpenResearch study funded by Altman—three years and sixty million dollars—concluded that cash alone cannot address chronic health conditions, childcare, or high housing costs. Mexico has operational versions: Pensión Bienestar, Jóvenes Construyendo el Futuro. They stabilize. They don't redistribute power. The distance between stabilizing and redistributing is exactly the distance between containing a crisis and solving its cause.

The second is what I've explored: a gamified model with six elements—XP, SP, levels, auditable AI, rotating committees, and transparent surveillance. It preserves the market but places the decision lever in auditable human committees. The intuition is that the problem isn't in the market itself but in who controls its parameters. Without all three mechanisms operating together, the model collapses. It happens exactly like in the Black Mirror episode where a points scheme that promised meritocracy ends up being a more sophisticated extraction mechanism. The six unresolved problems of that proposal are documented. I document them because honesty about limitations distinguishes a serious proposal from a manifesto.

Game mechanics usually reflect the priorities of those who design the rules. Here it's no different. I still don't have a clear idea how to prevent certain incentives from becoming corrupted at scale, though I keep exploring the issue because historical records show that the devil is in the parameters.

The third answer is the most radical and, curiously, the only one with an empirical case at sectoral scale: eliminate money as a coordination mechanism and allocate goods directly through computation between need and availability. Fresco and the Venus Project are the known utopian versions. Cybersyn is something else. Cybersyn is the only moment in history when a state attempted national post-monetary coordination and left records of what happened.

On July thirteenth, nineteen seventy-one, Fernando Flores, technical manager of CORFO, wrote a letter to Stafford Beer. Months later, Beer arrived in Santiago. What they built together in the following two years was a four-module structure: a telex network connecting factories with the center, statistical analysis software, a national econometric model, and a hexagonal room with seven white chairs and screens that allowed real-time visualization of the economy's state.

In October nineteen seventy-two, the CIA financed a strike by forty thousand truckers with the explicit objective of paralyzing the Chilean economy. The model coordinated two hundred trucks and processed two thousand telex messages per day. National production dropped nine percent, but the sectors connected to the design absorbed the blow better than sectors outside it. Allende told Beer directly that without that model the strike would have brought down the government. The most conservative assessment by independent historians confirms at least that second part.

A month before the coup, Beer asked Allende how much worker control he expected the model would eventually transfer to workers. Allende replied: the maximum. On September eleventh, Pinochet seized power. The operations room was dismantled. Flores went to prison. Allende died. Beer was in London that day managing purchases of Chilean exports.

What Cybersyn confirms isn't that post-monetary coordination is easy or inevitable. It confirms that it's technically possible at sectoral scale. That's no longer speculation: it's historical record. The most valuable aspect of the case, however, is what Beer and Allende didn't do. They didn't present the model as technically neutral. They knew that each parameter was a political decision. They designed so that politics would be visible, auditable, and debatable. Transparency wasn't an accident. It was the design.

Three decades later, the conversation about artificial intelligence repeats the error that Beer and Allende lucidly avoided. Every decision about what to optimize, what to measure, and what to consider efficient is presented as a technical result. As if the algorithm reached its conclusions from some place outside politics. As if the mentioned companies simply described reality instead of constructing it.

Those three initial announcements don't describe a natural phenomenon. They describe decisions. Someone decided to cancel the contract. Someone decided to make the public declaration about computing versus employees. Someone decided to publish the token projection for two thousand thirty at that moment and with that framework. The question isn't whether artificial intelligence will transform work. That part already happened. The question is who's making the decisions about the parameters of the model that will coordinate us all.

Beer knew this in nineteen seventy-one. He designed his structure so that question wouldn't have a single possible answer.

The operations room was dismantled in nineteen seventy-three. The servers of current language models are in Virginia, Oregon, and Singapore. No one's going to dismantle them. What matters is whether someone will audit their parameters and whether that audit will be public.

Sources:

1. Eden Medina, Cybernetic Revolutionaries: Technology and Politics in Allende's Chile, MIT Press, 2011.

2. OpenResearch, Unconditional Cash Transfer Study, 2023. Results published at openresearch.com.

3. Stafford Beer, Brain of the Firm, Allen Lane, 1972.

4. Goldman Sachs Research, AI Infrastructure Demand Forecast, 2024.

5. Thomas Paine, Agrarian Justice, 1797.