The Empire's Trail of Blood

There is a moment in the history of every failed war when the generals look at the data and call it a success. The bombing runs are on schedule. The kill ratio is favorable. The targeting algorithms are performing above projections. The quarterly briefing looks good on paper. And somewhere beneath all of that paper, beneath all of those metrics and models and machine-learning outputs, a people who were supposed to have been defeated are still alive, still fighting, still refusing to become what the data said they should have already become: a conquered population.

I have been thinking about this gap — between what the machine says it is doing and what the machine is actually accomplishing — for most of my adult life. I thought about it in the Bay Area when I was a teenager trying to understand why the country that received me had helped arm the country that had bombed my childhood. I thought about it at Amazon Web Services, where the same cloud infrastructure I was selling commercial analytics contracts on was also powering surveillance operations I was not supposed to ask about. I think about it every day now, when I look at what we have documented on ParentsPlea.com — more than 140,000 souls recorded, one profile at a time, in the years since I started building a different kind of database than the kind the generals use.

The generals' databases count targets. Ours count people.

That difference is the subject of this chapter.

The Scorecard Since 1950

Let me begin with the most uncomfortable fact in American military history: the United States has not won a war since 1945. Not a single one. Not in any meaningful sense of the word win — meaning a durable, just outcome that actually achieved the stated objective and left the affected region more stable than it found it.

Korea ended in an armistice in 1953, not a victory. The country was divided at the 38th parallel, and it remains divided today. More than 28,000 American troops are still stationed in South Korea over seventy years later. There was no victory. There was a pause — a pause that required a permanent military presence to maintain.

Vietnam ended in 1975 with American helicopters evacuating embassy staff from a rooftop in Saigon as the South Vietnamese government collapsed. The United States dropped more than seven million tons of bombs on Vietnam, Laos, and Cambodia — more than the combined tonnage dropped by all sides in World War II. On Laos alone, the United States conducted 580,000 bombing missions between 1964 and 1973, dropping an estimated two million tons of ordnance on a country the size of Utah that had never attacked the United States. Laos became the most heavily bombed country per capita in the history of warfare. Eighty million unexploded

cluster munitions still litter the Laotian countryside today, still killing and maiming farmers and children, decades after the last American plane turned for home. (Source: Legacies of War, https://legaciesofwar.org/about-laos/cluster-bomb-problem/; United Nations Mine Action Service, https://www.unmas.org/en/programmes/lao-pdr)

Cambodia received 2.7 million tons of American bombs during the Nixon-Kissinger years — a campaign of covert bombardment so intense and so indiscriminate that historians have argued it contributed directly to the conditions that allowed the Khmer Rouge to rise to power. The bombing of Cambodia is one of the great unacknowledged war crimes of the twentieth century. It was enabled by the most sophisticated targeting systems available at the time. It achieved nothing — except the deaths of an estimated 50,000 to 150,000 Cambodian civilians and the destabilization of a country that then descended into one of history's most devastating genocides. (Source: Owen and Kiernan, "Bombs Over Cambodia," Yale University, The Walrus, 2006, https:// http://www.yale.edu/cgp/Walrus_CambodiaBombing_OCT06.pdf)

Vietnam ended as it ended because no amount of bombs, no kill ratio, no body count, no pacification program, no electronic battlefield, no surveillance network could substitute for the thing the United States never had and could never buy: legitimacy in the eyes of the Vietnamese people. Ho Chi Minh said it plainly, decades before the last helicopter left Saigon: "You can kill ten of our men for every one we kill of yours. But even at those odds, you will lose and we will win." He was right. The data was wrong. The will of a people to determine their own future is not a variable that fits into a targeting algorithm.

The Gulf War of 1990-1991 was declared a victory. Iraq's military was expelled from Kuwait. Saddam Hussein remained in power. The sanctions that followed killed an estimated 500,000 Iraqi children from disease and malnutrition — a figure made grimly notorious when then-Secretary of State Madeleine Albright, asked in a 1996 CBS interview whether the deaths of half a million children were worth it, said: "We think the price is worth it." (Source: CBS News, "60 Minutes," May 12, 1996; UNICEF, "Child and Maternal Mortality Survey," 1999) That price was not worth it. And the partial victory of 1991 set the stage for the full catastrophe of 2003.

Iraq, the second time: $1.922 trillion spent. A country of 30 million people invaded on the basis of fabricated intelligence about weapons that did not exist. A functioning state — however brutal its government — reduced to rubble, its institutions dismantled, its professional class scattered across the diaspora, its sectarian tensions ignited by the removal of the only force that had suppressed them. What emerged from the smoking ruins of the American invasion of Iraq was not democracy. It was ISIS. (Source: Neta Crawford, Brown University Costs of War Project, https://watson.brown.edu/costsofwar/; People's World, "The Costs of U.S. Empire's Forever Wars," August 2021, https://www.peoplesworld.org/article/the-costs-of-u-s-empires-forever-wars/)

Afghanistan: twenty years. $2.3 trillion. The largest single military expenditure in American history since World War II. One hundred and seventy-six thousand people killed — including 47,245 Afghan civilians. Forty-six thousand Afghan national security forces trained and equipped at American expense. And then, in August 2021, the Taliban captured Kabul in a matter of days — every single major city falling before American troops had even finished their evacuation. The government that two decades of American investment had built dissolved so quickly it barely left a shadow. The Taliban are now governing the country they governed before America arrived, and

the girls they banned from school before America arrived are banned from school again. Twenty years. $2.3 trillion. Precisely nothing — for the Afghan people, and precisely nothing for the stated American objective. (Source: Brown University Costs of War Project, "Costs of the 20-year war on terror: $8 trillion and 900,000 deaths," September 2021, https://www.brown.edu/news/ 2021-09-01/costsofwar)

The total post-9/11 bill — across Iraq, Afghanistan, Pakistan, Syria, Libya, Yemen, and related operations — stands at more than $8 trillion, according to the Brown University Costs of War Project, the most comprehensive independent accounting of American war spending ever compiled. That $8 trillion includes $2.2 trillion already set aside for veterans' care through the year 2050 — money that will still be flowing to the wounded and traumatized long after every politician who started these wars is dead. It includes 38 million people displaced by American military operations. It includes an estimated 900,000 people killed directly by the violence. (Source: Costs of War Project, Watson Institute for International and Public Affairs, Brown University, https://costsofwar.watson.brown.edu/findings)

And from 2020 to 2024 alone, private firms received $2.4 trillion in Pentagon contracts — approximately 54 percent of the Department of Defense's entire discretionary spending budget. Nine hundred and fifty lobbyists worked on behalf of the arms industry in Washington during that period, shaping the policies and the budgets that kept those contracts flowing. (Source: Brown University Costs of War Project, 2024; https://watson.brown.edu/costsofwar/)

This is the scorecard. These are the outcomes. This is what $8 trillion and the most sophisticated military technology in human history has produced in seventy-five years of American interventionism since the last war America actually won.

The Assange Thesis: Wars Are Not Meant to Be Won

In 2011, speaking to filmmaker John Pilger for the documentary The War You Don't See, Julian Assange said something that most media organizations reported briefly and then moved quickly past — because if you sit with it long enough, it becomes very difficult to look at the history I have just recited and reach a different conclusion.

"The goal," Assange said, "is to use Afghanistan to wash money out of the tax bases of the US and Europe through Afghanistan and back into the hands of a transnational security elite. The goal is an endless war, not a successful war." (Source: WikiLeaks, Julian Assange speaking in 2011, as posted by WikiLeaks on X, August 18, 2021, https://x.com/wikileaks/status/ 1427929346262642688; YouTube, "Julian Assange on Afghan war in a 2011 video," https:// http://www.youtube.com/watch?v=_IGU_7alJ80)

I want to be careful here. Assange was speaking specifically about Afghanistan. And the thesis — that wars are deliberately designed to fail so that they can continue extracting public wealth into private hands — is a thesis that invites caricature. It sounds conspiratorial until you look at the numbers. Then it sounds like accounting.

Consider what actually happened. Between 2001 and 2021, the United States spent $2.3 trillion on the Afghanistan war. That money did not disappear. It went somewhere. It went to Lockheed Martin, which built the aircraft. To Raytheon, which built the missiles. To Palantir, which built the targeting and intelligence systems. To Booz Allen Hamilton and CACI and SAIC and dozens of other contractors who provided the personnel, the logistics, the base management, the

security, the intelligence analysis, the reconstruction projects that were never completed, the police training programs that produced police forces that dissolved the moment American support was withdrawn. The money was extracted from the American tax base — and specifically from the borrowing that replaced those taxes, since the post-9/11 wars were fought almost entirely on credit, a historically unprecedented decision — and it flowed into the accounts of a relatively small number of private corporations and their shareholders. (Source: Neta Crawford, Costs of War Project, Brown University; WarCosts.org, "Cost of War," https://www.warcosts.org/cost-of-war)

Neta Crawford, Chair of Political Science at Boston University and co-director of the Costs of War Project, put it plainly: "The United States has fought its post-9/11 wars almost entirely on borrowed money, with the costs shifted to future generations. This is the first time in American history that taxes were cut during wartime." During World War II, the top marginal tax rate was 94 percent. During Korea, 92 percent. During Vietnam, 70 percent. During the War on Terror, the Bush administration cut taxes to 35 percent — and kept cutting. Every dollar spent on these wars was borrowed. American taxpayers and their children will pay interest on that borrowing for generations. (Source: Crawford, as cited in WarCosts.org, https://www.warcosts.org/cost-of-war)

The conclusion is not a conspiracy theory. It is arithmetic. The wars failed their stated objectives. They succeeded — spectacularly — at transferring public wealth to private actors. Whether that transfer was the intention or merely the outcome, the result is the same: the transnational security elite Assange described is richer than it was before any of these wars started, and the populations that were supposed to be liberated are, by almost every measure, worse off.

If wars can be started by lies, Assange also said — in what became perhaps his most quoted line — then peace can be started by truth. That sentence is the reason whistleblowers like him and Edward Snowden matter. And it is the reason the governments that hunted both of them spent so much energy on the hunting.

What the Algorithm Cannot See

I published an article in October 2024 — the exact first anniversary of the Gaza genocide — that tried to name what I was seeing when I looked at twenty-four years of military AI and big data applied to war gaming and conflict modeling. I had been using the AI tools available to me to research the ParentsPlea database, to understand the gaps in what the machines knew about the people whose lives I was documenting. I asked ChatGPT, Grok, Bard — all of them — for the kind of data that ParentsPlea was trying to collect: names, biographies, family testimony, the lived human record of people killed in the conflicts of the past twenty-five years.

Not one of them had it. They had aggregate statistics. They had anonymized data. They had body counts and conflict timelines and geopolitical analyses. What they did not have was Alejandro Carranza Medina, forty years old, a fisherman from Santa Marta, Colombia, whose wife Katerine Hernández asked a question no algorithm has ever been programmed to answer: why did they have to take his life like that? Fishermen have the right to live. Why didn't they just detain them? What they did not have was Amal Khalil, forty-three years old, a journalist from Beirut who was killed in an airstrike while seeking shelter in a building, because the first building

she sheltered in was also struck. What they did not have was Refaat Alareer, forty-four years old, a poet and professor from Gaza who wrote a poem about his own death before it came for him — and who told us exactly what to do with it.

The AI systems that the militaries of the world have spent hundreds of billions of dollars building cannot hold a single one of these lives. They were built by generals, and generals see targets, not people. They were built by men who understand violence but not love. They were built on the assumption that if you collect enough data about a population, you can predict and control that population's behavior. And they have been wrong — consistently, catastrophically, expensively wrong — for seventy-five years.

At the site GoAmour.com, I have tried to articulate what I believe the machine is missing: the principles of love, compassion, and human dignity that constitute what it actually means to live alongside other people on this earth. Not as a spiritual abstraction, but as a practical reality. You cannot understand a people by counting their casualties. You cannot model a resistance movement by tracking its membership. You cannot bomb a culture out of existence. Every general who has tried to do these things — from the French in Algeria to the Americans in Vietnam to the Israelis in Gaza — has discovered, eventually, that they could not. The people kept surviving. The culture kept transmitting itself. The children kept being born and taught, in whatever rubble remained, what their parents believed and what had been done to them.

A fool with a tool is still a fool. Even if the tool is a supercomputer. Even if the fool has a star on his shoulder.

The Bombing of Laos and What It Proved

There is a lesson in Laos that has never been properly absorbed by the architects of American military policy, and the reason it has never been absorbed is that absorbing it would require admitting something the military industrial complex cannot afford to admit: that the most intensive bombing campaign in the history of warfare achieved nothing except the destruction of the people who endured it.

Between 1964 and 1973, the United States dropped two million tons of ordnance on Laos — a country that was not at war with the United States, whose government had asked the Americans to leave, and whose civilian population was not engaged in any military activity against American forces. The bombing was designed to disrupt North Vietnamese supply lines along the Ho Chi Minh Trail. It did not disrupt them. The North Vietnamese adapted. They moved at night. They dispersed. They built redundant routes. They sent women and children to clear bomb craters so the supply lines could reopen before dawn. The will of a people determined to achieve their objective cannot be interdicted from 30,000 feet. The data does not change this. The sophistication of the targeting system does not change this. The only thing that changes when you make the technology more precise is that you kill more people, with more precision and faster.

Eighty million unexploded cluster munitions remain in Laotian soil today. They kill an average of one hundred people per year — farmers, children, people who had nothing to do with the Ho Chi Minh Trail in 1968 and who were not born when the bombs were dropped. The United States government spent $17 million per day, every day, for nine years bombing Laos. It has spent approximately $180 million total in the fifty years since on clearing the ordnance. Do

that arithmetic. The investment in destruction and the investment in remediation are not in the same universe. (Source: Legacies of War, https://legaciesofwar.org/about-laos/cluster-bomb-problem/; U.S. Embassy Vientiane, Unexploded Ordnance Clearance assistance data)

Vietnam itself is still finding and clearing unexploded ordnance. Vietnam, Cambodia, and Laos together represent a postwar humanitarian catastrophe that continues in real time, fifty years after the last American combat troops left the theater. The bombs are still there. The people are still adapting around them.

This is what technological supremacy looks like from the ground.

The Human Data That Big Data Cannot Collect

When I built ParentsPlea.com, I built it because I understood something that the architects of military AI have never understood: the most important data about a conflict is not the targeting data. It is not the kill ratio, the sortie count, the satellite imagery analysis, or the biometric database of suspected militants. The most important data is the human data — the names, the ages, the occupations, the families, the things people loved and the things they were still planning to do when the bombs found them.

ParentsPlea.com now holds more than 140,000 such profiles. When I began building the database, I asked every major AI platform available to help me populate it with data from the conflicts of the past twenty-five years — Iraq, Afghanistan, Syria, Libya, Yemen, Gaza, Lebanon, Sudan, Somalia, Ukraine, Central America. Not one of them could do it. Not one of them had correlated the statistical record of these wars with the individual human lives that the statistics represented. The AI that can autonomously discover zero-day vulnerabilities in every major operating system on earth — the AI I described in Chapter Twelve — cannot tell you the name of a single person killed in the bombing of Mosul. It cannot tell you what Alejandro Carranza Medina was planning to cook for dinner the night before the airstrike found his boat. It cannot tell you what poem Refaat Alareer was working on in the week before he died.

This is not a technical limitation that better hardware will resolve. It is a philosophical one. The AI systems built by the military industrial complex were built to optimize for the goals of the people who commissioned them: to identify, to target, to kill more efficiently. They were not built to understand. And you cannot win a war — not in any durable, human sense of winning — without understanding the people you are fighting. Every general who has tried to substitute firepower for understanding has eventually learned this lesson at the expense of the people who paid for their war.

The United States has spent $8 trillion learning it since 2001. Laos is still paying for America's failure to learn it in 1973.

The Numbers on ParentsPlea Tell the Story the Generals Refuse to Read

More than 140,000 souls documented. The initial data was primarily from Gaza. But the database has grown to include victims from major conflict zones that we have been able to obtain data from. Each profile represents a family that submitted a story, or a journalist who told one before they too became a statistic. Each profile is a refusal — a refusal to allow the machinery of modern war to reduce a person to a body count.

Three reasons I built it. First: if there is ever going to be judgment in a court of law — and I believe there will be — prosecutors need to identify victims, not just aggregate death tolls. The

targeting algorithms in Chapter Two are abstract until you put a name on the person they targeted. The kill lists are numbers until you find the family of one of those 37,000 people Lavender flagged and let them speak. Second: if reparations are ever ordered — and history suggests they eventually are, long after the people who should have paid them are dead — you need to know who was harmed and what they are owed. Third: because just peace requires us to inspect the deaths that have taken place. Who is doing the killing. Who is suffering. Not in the aggregate. By name.

The generals have their databases. The generals have Project Maven and Lavender and the Gospel and Blue Wolf. They have the most sophisticated targeting infrastructure in the history of warfare. They have AI systems that can identify a face in a crowd from a drone flying at altitude and cross-reference it against a biometric database and recommend a strike within twenty seconds.

They cannot tell you why killing that person will end the war. Because it will not end the war. It never has.

What AI and Generals Will Always Fail to Model

I started my career in the technology sector before Y2K — the year when the entire world held its breath waiting for computers to fail, and instead the computers hummed along, and we all exhaled and went back to building faster machines. What I have watched in the quarter century since is the same dynamic repeating at every scale: the machines get faster, the targeting gets more precise, the data sets get larger, and the wars still do not end. Not in victory. Not in peace. They end in occupation or withdrawal or protracted stalemate, and they leave behind them the rubble of the lives that the data never counted.

The fundamental failure of artificial intelligence as a tool of war is not technical. It is philosophical. AI is trained on historical data. The historical data of warfare is the data of what armies did and what they destroyed. It does not contain what I was trying to build into ParentsPlea: the human dimension, the biographical record, the family testimony, the demand of the surviving parent who wants the world to understand not just that her child died but who her child was and why their death was wrong. It does not contain the will of a people to survive. It does not contain the love that rebuilds after the bombs stop falling.

I said in my 2024 article: garbage in, garbage out. The AI that runs on the assumptions of men who only know how to waterboard and torture and kill will produce targeting recommendations for more waterboarding, more torture, more killing. It will optimize for what its designers valued. And what the designers of military AI have always valued is not peace. It is dominance. And dominance, pursued past the point of reason, produces not the victory the generals promised but the exact sequence of events the scorecard above describes: Korea, Vietnam, Iraq, Afghanistan, Libya, Syria, Yemen, Gaza. Endless war. Not successful war.

Assange was right in 2011. The goal is an endless war, not a successful war. And the technology has made the endlessness more efficient, the killing more precise, and the corporate revenues more reliable — without changing by a single percentage point the one variable that determines the outcome of every conflict: the refusal of occupied people to accept their occupation forever.

That refusal is not a data point. It is a human constant. And no algorithm that the PayPal Mafia has ever built, no targeting system that Unit 8200 has ever deployed, no cloud contract that

Amazon or Google has ever signed — none of it has produced a machine that can model human dignity. None of them can calculate the point at which a people decide that dying in resistance is preferable to living in submission. None of them can predict the moment when that calculation tips — when a Tunisian fruit vendor burns himself alive and an entire region rises, when a Palestinian poet writes his own elegy and it is shared by millions of people who have never been to Gaza, when a Vietnamese farmer looks at the B-52s on the horizon and decides that no bomb is large enough to make him forget what his land is.

The machines cannot see this. The generals cannot see this. And until they can, every war they start will end the same way every war they have started since 1945 has ended: not with victory, but with a helicopter on a rooftop and a question that no amount of data will ever answer.

What were we doing there? What did we spend all of this for?

More than 140,000 names on ParentsPlea.com. Each one an answer.

The historical record of American military failure — from Korea to Afghanistan to Gaza — forms the context in which the legal instruments described in Chapter Fourteen must be understood. The petition to Spain is not filed in a vacuum. It is filed against the backdrop of eight decades of evidence that these wars were never designed to be won, only to be fought — and that the technology companies described in this book have willingly provided the infrastructure that makes the fighting possible. Chapter Fourteen is about what happens when the documentation of failure becomes the foundation for accountability.