“War is the realm of uncertainty; three-quarters of the factors on which action in war is based are wrapped in a fog of greater or lesser uncertainty. A sensitive and discriminating judgment is called for; a skilled intelligence to scent out the truth.”
– Carl von Clausewitz, On War (1832)
Wrong Lesson from Caracas
Do you get the sense that the world is being run by a menacing algorithm?
Between the capture of Maduro in January and the current war in Iran, we’ve moved from strategic operations to what feels like a high-stakes beta test for AI generated warfare.
In early January, the world watched as Operation Absolute Resolve snatched Nicolás Maduro and his wife from Caracas in a lightning strike. To the White House, this was the ultimate proof of concept. A clean, surgical extraction that decapitated a regime with zero U.S. casualties.
But Iran isn’t Venezuela. By applying the Maduro Playbook to Tehran – specifically the assassination of the Supreme Leader on February 27 – the Trump administration ignored the fundamental difference between a crumbling narco-state and a deeply ideological, regional powerhouse.
Capturing Maduro was a police action. Martyring the Ayatollah was a religious and geopolitical earthquake. Instead of a peaceful transition, the USA-Israel consortium triggered a phased retaliation from Iran that Alastair Crooke notes is designed to systematically evict the U.S. from the Middle East entirely.
What’s more, President Trump triggered this war of choice with a dangerously light belt of ammo. He recently admitted that while our medium-grade munitions are plentiful, our highest-end stockpiles, the stuff that actually stops incoming ballistic missiles, are “not where we want them to be.”
Four years of shoveling top-tier hardware into Ukraine has left the cupboard bare. We are currently burning through Patriot interceptors and Tomahawks faster than they can be manufactured. In a war of attrition against Iran’s swarms of low-cost Shahed drones, the math is completely unbalanced. We are using $2 million missiles to stop $20,000 drones.
Why Anthropic Had to Go
The timing of the Anthropic ban last Friday wasn’t a coincidence. It was a clearing of the decks. Anthropic’s refusal to allow Claude to be used for mass surveillance or fully autonomous lethal systems made it a national security risk in the eyes of the Pentagon leadership.
In Anthropic’s place, comes OpenAI and the GPT-5.3 Codex upgrade. The administration didn’t just want a chatbot, they wanted an agentic system capable of managing a multi-theater conflict in real-time.
By designating Anthropic a supply chain risk and switching to OpenAI, the government has essentially handed the keys to a system that doesn’t have the same restraints and guardrails. We are now in an era of war where AI is calling-the-shots.
If the strike on the girls elementary school in southern Iran, which killed more than 100 children, felt calculated, it’s likely because an algorithm determined that the “inflame-and-solidify” metric was more valuable than the “humanitarian-outcry” risk.
This shift marks the end of human-centered oversight in modern warfare. Under the new GPT-5.3 Codex framework, tactical decisions are processed at speeds that make human ethical deliberation an obsolete bottleneck. By stripping away Anthropic’s internal AI constraints, the Pentagon has moved from assistive technology to a predictive authority that treats civilian casualties as data points in an ‘end justifies the means’ equation.
The destructive reality of the Southern Iran strike illustrates this pivot. When a system is programmed to prioritize long term advantages over immediate moral cost, the obliteration of an elementary school becomes a logical move in a global chess game.
This machine now operates with detached precision. The AI algorithm has no soul.
Dragon in the Machine
Yet the U.S. may not have the AI edge it supposed. Recall the China-Russia-Iran trilateral pact signed back in January. While the U.S. is betting on OpenAI, Iran is being backed by Chinese electronic warfare and AI tools.
In this regard, China is treating the Middle East as a proxy laboratory. It isn’t just sending hardware. It’s providing the AI-driven data to blind Western sensors and predict U.S. naval movements.
If we see a cyber-virus or a mass grid-down situation, it won’t be a random event. It will be a targeted strike by a Chinese AI that has been training on our patterns for years.
China has ample rationale for escalating its support of Iran. A major objective of the attack on Iran is to severe the flow of oil from Tehran to Beijing.
For years, the ghost fleet – a shadowy network of aging tankers – has taxied millions of barrels of Iranian crude to Chinese independent refineries. The ghost fleet has effectively bypassed U.S. sanctions and fueled the very industrial base that produces the AI chips now being used against America.
The Pentagon’s AI-driven modeling likely gamed that a conventional embargo was no longer enough. To truly de-risk from China, the administration had to kill the source. By turning the Persian Gulf into a war zone, the U.S. is attempting to physically blockade the energy supply of its greatest global rival.
However, this is where the AI logic hits a wall. The U.S. strategy assumes that China will simply sit back and watch its energy security evaporate. Instead, we are seeing the Dragon breathe fire through Iranian proxies.
The Suffocating Fog of AI Generated Warfare
As the war escalates, President Trump’s finding it harder to pull miracles out of his hat. The switch from Anthropic to OpenAI’s GPT-5.3 was supposed to provide a God’s-eye view of the battlefield. But it seems to have created a feedback loop of overconfidence.
The AI model said a strike on a civilian target would break the enemy’s will based on data from a completely different cultural context (like Venezuela). Instead, this resulted in a massive surge in volunteer martyrdom.
The Trump administration’s reliance on AI generated warfare, where every life is a data point and every missile is a line of code, has blinded them to the human element. They learned from COVID-19 that a shocked population can be managed. But they are learning now that a population with nothing left to lose cannot be computed.
The question of whether AI is ready for the big leagues is being answered in blood. In Venezuela, it worked because the variables were simple. In Iran, the variables are infinite. AI is unable to comprehend it.
If the goal was a quick, surgical regime change, the AI has failed. If the goal was a total global shock that leads to “owning nothing and being happy,” then the chaos is a feature, not a failure. At this point in the operation, we expect chaos to multiply and ricochet about like metastatic cancer for decades to come.
President Trump, War Secretary Pete Hegseth, and Israel Prime Minister Benjamin Netanyahu, thought they were playing a video game where the AI always ensures a win. They forgot that the other side has a computer too. And theirs is backed by a billion people and a lot of historical resentment.
As these rival AI systems clash, the distinction between strategic victory and systemic collapse vanishes into a haze of predictive errors and unintended escalations. We have moved past the era of human accountability into a new reality where no one – not the generals, not the presidents, and certainly not the architects of the code – can see through the suffocating fog of AI generated warfare.
We’re, no doubt, destined for uncertain catastrophe.
[Editor’s note: Get a free copy of an important special report called, “Cash Machine – Why You Should Own this Mineral Royalty with a 12% Yield,” when you join the Economic Prism mailing list today. If you want a special trial deal to check out MN Gordon’s Wealth Prism Letter, you can grab that here.]
Sincerely,
MN Gordon
for Economic Prism
Return from The Suffocating Fog of AI Generated Warfare to Economic Prism
