A general who won a world war said this better than any pacifist ever has. It belongs at the center of this page:

“Every gun that is made, every warship launched, every rocket fired signifies, in the final sense, a theft from those who hunger and are not fed, those who are cold and are not clothed. This world in arms is not spending money alone. It is spending the sweat of its laborers, the genius of its scientists, the hopes of its children.”

— Dwight D. Eisenhower, “The Chance for Peace,” April 16, 1953. He was not a man who flinched from war. He commanded the invasion of Normandy. And still he called every weapon a theft from the people it was supposedly built to protect.

The line that matters most is the last one: the hopes of its children. Every bomb is not only a failure of the adults in the room in the present tense — it is money borrowed from a generation that had no vote in the decision. A single fighter plane is a set of schools that never got built. A destroyer is a set of hospitals that never opened. We are not just spending money we have; we are spending a future that isn’t ours to spend.

• Every military dollar carries an opportunity cost measured in schools, hospitals, and roads not built
• The bill for war is paid disproportionately by the generation with the least say in declaring it
• “We can’t afford either” is not a slogan — it is the honest ledger behind every appropriation

Leadership is not the same thing as command. The most durable influence this country ever bought, it bought by example — the Marshall Plan did more for American security than any weapon system, and the appeal of American institutions at their best generated more goodwill than force ever has.

So the position is straightforward. We need a stronger United Nations, not a weaker one, and we need to stop treating it as an inconvenience to be worked around. We need to submit ourselves to the international courts — the same courts we expect other nations to answer to. We cannot demand a rules-based order and then carve ourselves an exemption from the rules. That exemption is the tell. It is how the rest of the world learns that “international law” means “law for everyone but us.”

Being an American should mean being a civilized human being in the eyes of the world — someone who keeps the agreements he signs and answers to the same standard he holds others to.

• Recommit to a strengthened UN and to the international courts, including the jurisdiction we currently exempt ourselves from
• Adopt one governing rule for the whole tab: we will not ask of any government what we will not ask of ourselves
• Lead by example rather than by force — the only form of leadership that has ever lasted

“We cannot demand a rules-based order and exempt ourselves from the rules.”

There is accountability for Israel that has to happen, and I will not pretend otherwise to keep the peace at a dinner party. But the goal is not punishment. The goal is a durable end state, and there is only one on offer: Gaza and the West Bank become a state of their own, with real autonomy and a lasting, defined border. The 1967 lines are the negotiating baseline — not a foregone conclusion, but the starting point that has anchored serious proposals for half a century.

The parity principle. Whatever we extend to one side, we extend to the other. Equal aid, equal scrutiny, equal standing. If we are going to intervene, the United States sits without a thumb on the scale — only the facts. We are not a party to the grievance; we are supposed to be the honest broker, and an honest broker does not arm one side and lecture the other.

The naming mechanism. Watch what the word “state” does. Statehood carries rights — the right to defense, to a military, to a seat at the table, to be recognized as a peer rather than a problem. Withhold the name and you withhold the rights, while pretending you have withheld nothing. Keep Palestine “unidentified” and its fighters are “terrorists”; grant it the name and the exact same acts become the defense of a state. The label is doing the work that the facts should be doing. That double standard is nails on a chalkboard to any real sense of justice.

What Zionism was, and what this faction is. This is the distinction that separates documenting a hijacking from bigotry, so it has to be exact. Founding Zionism was a secular, nationalist project — Herzl was an atheist, and early Zionists argued not from a divine deed to the land but from a human and national right to survive rampant European antisemitism, with the location itself negotiable. The faction governing today argues from a religious-nationalist, title-to-the-land claim. Naming that gap is not an attack on Jewish people or on Israel’s existence; it is the observation that a refuge movement was captured by a settlement movement.

The gold coffin test. Smuggle a two-thousand-year-old solid-gold coffin out of Egypt and everyone agrees it is theft — the Met returned exactly such a coffin, the Coffin of Nedjemankh, once its forged export papers came to light. Bulldoze the houses that are standing today, plant your own settlement on the land, and suddenly it is “complicated.” The only thing that changed is whether anyone alive still has to answer for it. Recency should make a theft easier to judge, not harder. The complexity of the whole conflict does not dissolve the plain wrong of taking this specific deeded plot, this specific orchard, this specific home.

Follow the money, precisely. Israel is the largest cumulative recipient of US foreign aid in the post-WWII era — that is the plain historical fact, and I favor suspending military aid, in line with the position Sanders and Massie have taken. As for domestic influence, the defensible number is the electoral one: AIPAC’s PAC and its super PAC spent roughly $126.9 million in the 2023–24 cycle, making it the single largest PAC source of campaign money to members of Congress that cycle — concentrated on one foreign-policy question. That claim survives a fact-check. The vaguer “largest lobbying influence” does not, and I would rather be precise and unassailable than sweeping and wrong.

Answer the strongest objection honestly. The serious case for the status quo is not “Israel deserves it.” It is that Israel is a strategic asset — intelligence, a stable base in a volatile region, a proving ground for US-interoperable systems — and that abandoning a longtime ally signals unreliability to every other ally. The answer is not to dismiss that. The answer is to put it on the ledger: show the return on decades of spending against the domestic opportunity cost, and let the numbers make the argument. Change the environment, change the world — applied to a budget line.

• Two states, real autonomy, a defined border, 1967 lines as the negotiating baseline
• Separate the two mechanisms: the territorial baseline answers “where are the lines,” the accountability window answers “what happens for recent harms” — different questions, different instruments
• Suspend military aid; treat both sides by one standard
• The US as honest broker: facts, not favor

We are the only nation that has used a nuclear weapon on a civilian population, and we maintain one of the largest arsenals on earth while sanctioning other countries for pursuing the same capability. That is not a stable moral position. It is the lead-by-example doctrine failing in the one place it matters most.

• Announce a declared trajectory toward reduction — not overnight, but a direction the world can hold us to
• The deterrence argument assumes rational actors control every weapon at all times, and that assumption gets weaker every year, with more warheads in more hands and aging command systems
• Moral authority cannot coexist with reserving, for ourselves alone, the right to end the world

“We will not ask of any other government what we will not ask of ourselves.”

Here is the distinction that separates nuclear from every other weapon technology, and the one that makes our policy on it so incoherent: the same atom that builds the bomb also runs the cleanest large-scale power we have. Split it for a warhead and it is the end of the world. Split it for a reactor and it is a carbon-free city that never goes dark. “Nuclear” is not one thing to be feared. It is two things that have to be governed differently.

The irony is writing itself right now. The technology we are most afraid of weaponizing — artificial intelligence — is the single largest new customer for nuclear as clean energy. Every major AI company has signed nuclear deals in the last two years: Microsoft is paying to restart Three Mile Island, and Amazon, Google, and Meta are financing reactors directly, roughly ten gigawatts of committed capacity, because a single AI data center can draw the power of a mid-sized city and needs it around the clock. The bomb and the server farm run on the same discovery. That alone should end the reflex of treating every reactor as a warhead in waiting.

And nuclear energy is worth wanting on its own merits, not merely tolerating. Weigh it honestly. On emissions it is not close: over its full life cycle nuclear runs about a dozen grams of CO2 per kilowatt-hour, on par with wind and a fraction of solar, and a tiny fraction of coal or gas. On safety it is among the least deadly sources we have — on the order of a few hundredths of a death per terawatt-hour, comparable to wind and solar and thousands of times safer than coal, even counting Chernobyl and Fukushima in full. And on reliability it does the one thing renewables still cannot do alone: a US reactor runs near ninety percent of the time regardless of weather or nightfall, where solar sits around a quarter and wind around a third. You need roughly four gigawatts of solar nameplate to match the yearly output of one gigawatt of nuclear — plus the storage to move it to when the sun is down.

The costs are real and I will not hide them: reactors are expensive and slow to build, and the accident risk is low-probability but high-consequence. As for the spent rods — the honest knock on nuclear — “spent” is nearly a misnomer. Roughly ninety-five percent of the material in a used fuel rod is still usable, and France, Japan, and others already reprocess it into fresh fuel; the frontier work, in fast reactors, aims to close the loop further and shrink what is left to a fraction. The barrier now is more political and economic than physical. So the fair verdict is not nuclear instead of renewables. It is nuclear and renewables, together, in place of fossil fuel — nuclear as the firm, always-on floor, renewables as the cheap bulk on top. Both beat burning carbon, and only one of them keeps a hospital lit at three in the morning in a windless January.

So when a country says it wants reactors, not warheads, the empiricist answer is not reflexive suspicion. It is: verify it, and let energy be the product. A nation should be able to sell its fossil fuel rather than depend on it to keep the lights on. Two recent cases show what happens when we get this right, and then throw it away.

And here is the part that bleeds straight into the weapons question — the part that has to stop. The same civilian pursuit — enrichment, reactors, a fuel cycle — gets called “clean energy” when an ally does it, or when our own tech giants do it to run their data centers, and “a weapons program” the moment the country doing it lands on the wrong side of Washington. The physics does not change. The label does. We let allies run the full fuel cycle — Japan reprocesses its spent fuel with our explicit blessing — while telling adversaries those same steps are proof of a bomb. We do not even reprocess our own spent fuel, having sworn it off decades ago over the very proliferation fear we now invoke against others. The rule was never the physics. It was the flag. It is the naming mechanism again, the same move this platform flags on the Palestine page: a word doing work the facts do not support, tracking a country’s relationship to us rather than what it is actually building. A nation’s reach for energy independence should not become a reason for war the day it stops taking our orders. Verify the program, inspect it, hold it to a hard standard — but the same standard for friend and adversary, or admit the standard was never about the technology at all.

North Korea, told honestly. The 1994 Agreed Framework froze North Korea’s plutonium program — the fast, proven route to a bomb — for roughly eight years, in exchange for two civilian light-water reactors and fuel oil. Now look at what actually killed it, because the standard story gets it backwards. The incoming administration had signaled its hostility to the deal before it ever took office, and the intelligence later used to justify tearing it up did not hold up the way it was sold. In 2002 the public claim was an active, production-scale uranium-enrichment program. By 2007 the government’s own North Korea intelligence chief had walked that back to “mid-confidence” that such a program even existed, and the aluminum tubes at the center of the case turned out to be dual-use. High confidence that Pyongyang wanted the capability; far weaker evidence it had a working program. It was the Iraq pattern — conclusion first, intelligence bent to fit — run in parallel by some of the same officials.

None of that makes North Korea trustworthy. They were pursuing enrichment, and they are a brutal regime. But put yourself in their shoes: your counterpart is years behind on the reactors it promised and has just elected people who campaigned on killing the deal. Hedging is not the same as unprovoked cheating, and the question of who broke faith first is genuinely two-sided. Then look at the ledger. The bomb North Korea finally tested in 2006 was plutonium — made from the very material the Agreed Framework had frozen, and unfrozen the moment the deal collapsed. A working freeze on the actual bomb material was traded away over an overstated claim about a different one. Even that flawed deal bought eight years. Abandoning it bought a nuclear-armed adversary within three.

Iran, the cleaner case. The 2015 deal capped Iran’s enrichment at 3.67 percent and put international inspectors inside its facilities — and those inspectors repeatedly certified Iran was complying, including a full year after the United States walked away in 2018. We walked anyway. Iran then climbed to 60 percent enrichment and a stockpile of roughly 440 kilograms, and in the summer of 2025 the whole thing ended where the deal was built to prevent it from ending: in war, with strikes on the very facilities the inspectors had been watching. Whatever you believe about Iran’s intentions, the deal measurably froze them and killing it measurably unfroze them. That is not ideology. That is the ledger.

The honest objection, and the answer. A fair critic says neither regime had clean hands — North Korea was pursuing enrichment on the side, and Iran had a hidden weapons history with limits set to expire on a timer. Granted. But “they might cheat” is an argument for tighter verification, not for tearing up the one mechanism that put inspectors on the ground. You do not fix a smoke detector’s imperfection by ripping it off the ceiling. And notice what actually killed both deals: not that they were failing on their own terms — inspectors were certifying compliance — but domestic politics at home.

Which is the real defect, and the thing to fix. Both agreements died to a single signature, because neither was a ratified treaty. Each was a political commitment one administration could reverse with a pen and a press release. A commitment as consequential as nuclear restraint should not evaporate every time the White House changes hands. I am not pretending we can bind a future elected government forever — a democracy has to be able to change its mind, and a deal no one can ever exit is its own kind of danger. The goal is not to make an agreement impossible to undo. The goal is to raise the cost of undoing it until walking away is an act of vandalism instead of a campaign applause line: ratify real treaties through the Senate where we can, embed the commitment in the UN and in our alliances so that exit means breaking with partners rather than merely reversing a predecessor, and tie our word to the allies whose trust we actually depend on.

“A deal one president can sign with a pen, the next can burn with a match. That is the defect we have to design out.”

And step back far enough to see it is not one mistake but a habit. The same doctrine has now run twice, with a demonstration bolted in the middle. First George W. Bush named North Korea to the “axis of evil” and let the Agreed Framework die — and Pyongyang, watching, built the bomb. Then came Iraq: invaded on intelligence that proved false, its leader toppled. That was not a separate deal we broke; it was the billboard that taught every other named regime exactly what disarmament and cooperation buy you. And then a second administration tore up the Iran deal the previous one had signed, and Iran ran its enrichment to the brink. Two cycles, one design — inherit or strike a deal that is holding, tear it up for reasons at home, and act surprised when the other side sprints for the one insurance policy that has ever reliably stopped an American regime change. We taught that lesson ourselves, on purpose, twice.

We cannot undo what we did. Iraq happened. The frameworks are dead and the arsenals are real. But we can stop doing it — and stopping is the whole point of this page. Change the environment a cornered regime is reacting to, and you change what it concludes it must build to survive. That is not softness. It is the only nonproliferation strategy that has ever addressed the actual cause.

“We taught every adversary that a working bomb is the one thing that reliably stops a regime change — then called the result a nonproliferation crisis.”

Artificial intelligence has become the quintessential catchphrase of the mid-2020s, and it means a dozen different things to a dozen different people — often the same person will champion it in one breath and dread it in the next. Some of my own closest friends, plainspoken and MAGA-minded, are building the very data centers that power it, and they still carry a real fear of the thing they are building. On questions of war and foreign policy, that fear deserves a serious answer rather than a punchline.

AI is a tool. Strip away the mythology and that is all it is. A saw can frame a house or cut one down; the blade does not care which. The same capability that could help us solve dementia and Alzheimer’s — adding quality of life, not just quantity of years — can also guide autonomous weapons that neither think nor care, carrying conventional munitions or, worse, chemical agents capable of devastating a population. That is exactly what makes this hard: we cannot simply ban the technology, because the tool that cures is the tool that kills. The question in front of us is not whether to have AI. It is how to govern what people do with it.

The Hollywood distortion. The dystopia the movies sold us — the machine that wakes up and decides to kill us — is about as well-founded as aliens crossing light-years of empty space for the sole purpose of murdering the locals. It makes for a good movie and terrible policy. And that is the real hazard: in a democracy, voters elect leaders who write policy around these flawed versions of reality, so we end up governing the movie instead of the machine — bracing for a fantasy while the actual risks go unaddressed.

The real risk is concentration, not sentience. What is actually worth fearing is not a self-aware machine. It is a powerful tool in the wrong hands — a rogue state, or a single unstable individual who mistakes technical genius for a mandate to steer the world. We have had our share of unstable geniuses throughout history: brilliant, unwell people who acquired outsized power over the technology of their moment. Howard Hughes in his era, and the wealthiest technologist of ours, invite the same uncomfortable question — should any one person hold this much control over the tools reshaping everyone else’s life? You do not have to answer it to see that concentration itself, not sentience, is where the danger lives.

When deterrence stops working. For seventy years, arms control has rested on two things: attribution and verification. Attribution means you can tell who did it; verification means you can confirm whether a country is keeping its word. AI erodes both at once — and that is what makes it a genuinely new problem rather than the nuclear problem wearing a new coat. Start with attribution: a nuclear detonation leaves a return address — an isotopic signature, a launch plume, a trajectory. That return address is the whole basis of deterrence; I do not strike you because you will know it was me and strike back. An AI-guided attack need not leave one. A cyber weapon, an autonomous drone, or a pathogen designed with machine assistance can arrive with no fingerprint at all — and when the attacker cannot be identified, the fear of retaliation that has kept the worst weapons holstered stops working, and the bar drops for exactly the actors we should worry about most.

Then verification. With a nuclear program you can count centrifuges and measure enrichment; the physics leaves signs. AI leaves almost none. The same chips and the same building that train a cancer-research model can train a weapons-design model. A model is a file — it can be copied endlessly and carried across a border on a drive, and a system that looks harmless today can be turned into something dangerous with a modest amount of extra training. There is no enrichment level to measure, because we do not yet know what the dangerous threshold even is. You cannot inspect for what you cannot define. That is the real gap: it is not that we have failed to sign the treaties, it is that we do not yet know what the treaties should say.

An IAEA for artificial intelligence. This is where the United States should be leading. The world built the International Atomic Energy Agency to watch over the last world-changing technology; we now need its equivalent for this one — not a copy of it, because the old inspection playbook does not map, but a body built to answer questions we are only beginning to learn how to ask. The country that helped invent the problem is the country best positioned to help govern it. If we do not take the lead in writing those rules, someone else will write them for us — and we may not like the answers they arrive at.

“We don’t only need the treaties we know we need. We need the ones we can’t yet imagine.”