墨子 MOZI

Twenty-four centuries ago, a teacher walked ten days and ten nights to stop a war. He didn’t command an army. He didn’t negotiate a treaty. He took off his belt, laid it on a table to represent a city wall, and proved that his defenses could outlast any attack.

This is the story of Mozi and the siege of Song, told in Chapter 50 of the text you’re about to read. The engineer Gongshu Ban had built cloud-ladders for the king of Chu — siege weapons designed to breach Song’s walls. Mozi tried moral argument. Gongshu Ban was persuaded; the king was not. So Mozi challenged Gongshu Ban to a simulated siege on the table. Nine attack strategies deployed; nine repulsed. Gongshu Ban exhausted his arsenal. Mozi still had defenses in reserve.

Then the king’s last card: just kill Mozi, then take the city. But Mozi had already thought of that.

“My disciples Qin Hua Li and others numbering three hundred are already armed with my implements of defence waiting on the city wall of Song for the bandits from Chu. Though I be murdered, you cannot exhaust the defence of Song.”

The king of Chu called off the attack.

On his way home through Song — the city he had just saved — it rained. Mozi sought shelter at a gate. The guard turned him away.

Why the Mohists matter

The Mohists were an organized community of thinkers and engineers who traveled between warring states providing defensive technology to cities under threat. They argued for universal moral concern across state boundaries, condemned offensive warfare, insisted on meritocratic governance over hereditary privilege, and rejected fatalism — the belief that outcomes are predetermined and effort is futile.

The Mozi, the text collected under their name, preserves all of this: moral philosophy, political argument, formal logic, early science, and detailed manuals for city defense.

I built this site because I think the Mohists saw something important — something that took twenty-four centuries to resurface.

Why I care

I work on AI policy in Washington, D.C. Most of my time goes to a question with old roots: how do you shape the development of powerful technology so it protects people rather than threatens them?

There’s a useful framework for thinking about this, sometimes called d/acc — defensive accelerationism. The core idea: not all technological acceleration is equal. Some technologies favor defenders over attackers. Some distribute power; others concentrate it. The distinction matters, and it doesn’t sort itself out automatically.

Vitalik Buterin, who coined the term, traces this back to a structural observation: societies where geography favored defense — Switzerland’s mountains, Britain’s seas, the stateless highlands of Zomia — developed freer, more democratic governance. Where offense was easy, as on the open steppes, authoritarian empires emerged. Technology can shift this balance. It can make the world look more like mountains, or more like plains.

I came to the Mohists through this lens. Buterin mentions them in passing — the Mohists “focused on helping cities build better walls and fortifications.” But when I read the Mozi, I found that the walls were the least interesting part.

Black and white

Chapter 17 opens with an argument that has stayed with me since I first read it.

Everyone agrees that stealing peaches is wrong. Stealing oxen is worse. Murder is worse still. The gentlemen of the world condemn all of these. “But when it comes to the great attack of states, they do not know that they should condemn it. On the contrary, they applaud it, calling it righteous.”

Then the analogy:

“If, upon seeing a little blackness, one should say it is black, but, upon seeing much, should say it is white; then we should think he could not tell the difference between black and white.”

We condemn the small wrong and celebrate the large one. We punish the individual cyberattack and call the state-sponsored offensive operation “strategy.” We treat individual surveillance as stalking and mass surveillance as national security. Mozi noticed this twenty-four centuries ago: the gentlemen of the world are confused about the difference between righteousness and unrighteousness. He was right, and we still are.

Water or fuel

In Chapter 46, a critic puts a sharp question to Mozi. Neither universal love nor selfish partiality has yet produced visible results. So why prefer one over the other?

“Suppose a conflagration is on. One person is fetching water to extinguish it, and another is holding some fuel to reinforce it. Neither of them has yet accomplished anything, but which one do you value?”

Everyone is accelerating. The question is whether you’re carrying water or fuel — whether the technologies you build favor the defender or the attacker, whether the systems you design distribute power or concentrate it. The Mohists carried water.

Against fate

The Mohist connection that struck me hardest was their rejection of fatalism.

Three chapters (35–37) are devoted to dismantling the idea that outcomes are predetermined. Mozi treats fatalism not as a neutral error but as a weapon — a doctrine that licenses inaction, discourages effort, and serves the interests of those who prefer the status quo:

“With this doctrine the rulers are urged above and the people are kept away from their work below. Hence the fatalists are unmagnanimous.”

And: “Fatalism was an invention of the wicked kings and the practice of miserable men.”

His proof is simple. Under the same conditions, with the same people, different rulers produced radically different outcomes. “The times did not alter and the people did not change, yet under Jie and Zhou the world was chaotic and under Tang and Wu it was orderly. Can it be said that there is fate?”

Same conditions, different outcomes. The variable was effort and governance, not destiny.

I’ve made a version of this argument about technology. The people who insist that AI will develop however it develops, regardless of what we do, are the modern fatalists. The Mohists would have had a word for them. They’d have pointed out that fatalism conveniently benefits those already in power and conveniently excuses everyone else from doing the hard work of shaping the trajectory.

Chapter 37 makes the link between fatalism and defensive failure explicit: if people believe outcomes are fixed, “within, defense will not be strong, and, without, attack will not be victorious.” Why build the wall if fate has already decided whether it holds?

Mountains or plains

Beyond building for defense, the Mohists understood something structural about the relationship between defense, power, and freedom.

When nobody guards the weak, “the strong will overpower the weak, the many will oppress the few, the wealthy will mock the poor, the honoured will disdain the humble, the cunning will deceive the simple” (Chapter 15). This isn’t moralizing. It’s a description of what the world looks like when offense is cheap and defense is absent — what d/acc would call an “open plains” equilibrium. Power cascades upward. The strong take from the weak because nothing stops them.

The Mohist alternative wasn’t just “be nicer.” It was structural. Build the walls. Train the defenders. Share the technology freely. Organize as a distributed community that can deploy defensive capability across state boundaries. Evaluate every expenditure by whether it actually protects people. This is the Mohist version of making the world look more like mountains.

The window that closed

Here is the hardest thing about the Mohists: they failed.

Despite being right about defense, despite building the walls, despite training three hundred defenders, they could not prevent Qin’s unification of China by conquest. Mohism likely did not survive the Qin dynasty. The aggressive empire they had spent generations opposing destroyed them.

Defensive technology buys time. That’s what it does. What matters is what you build with that time — the institutions, the coalitions, the widespread adoption that keeps the window open. The Mohists were brilliant engineers and incisive moral thinkers. But they never persuaded enough rulers to adopt their system broadly. They fragmented into rival factions, bickering about ideology. So their window closed.

I think about this when I think about AI. We have a window right now to build defenses — to harden infrastructure, to develop verification systems, to build the institutional capacity to govern powerful technology. That window will not stay open forever.

“In times of peace, give thought to danger. In times of danger, give thought to peace.” (Chapter 71)

The Mohists gave thought to danger. They built the walls. They trained the defenders. And still the empire came.

The lesson is not that defense is futile. It’s that defense alone is not enough. You also have to win the argument — to persuade the people inside the walls that the walls are worth maintaining, even when the guard at the gate doesn’t recognize you standing in the rain.

Read the text

The Mozi is not what you expect from an ancient Chinese philosophical text. It is practical where Confucius is ceremonial, analytical where Laozi is mystical. It makes arguments — real arguments, with evidence and counterexamples and worked analogies. Mozi runs cost-benefit analyses on military campaigns. He deploys thought experiments with a logician’s precision.

The core doctrinal chapters — Universal Love (14–16), Condemnation of Offensive War (17–19), Anti-Fatalism (35–37) — lay out the moral philosophy. The dialogues (46–50) contain stories. The Canons (40–45) are dense formal logic and early science. The defense chapters (52–71) are engineering manuals — the Mohists literally building the technology they argued for.

Start with Chapter 50. Then read 17. Then 35. Then come back for the rest.