views

We publish a variety of perspectives. Nothing written here is to be construed as representing the views of this blog.

Saturday, November 14, 2026

Controlling AI

 


Today I'm going to connect four things: Milei's success, Bukele's success, the collapse of the system, and Jensen Huang being pushed out of Nvidia. It looks like four news stories. It's one story. On September 23, Javier Milei stood at the UN General Assembly and said what nobody in that room wanted to hear: the institution has become a useless organization whose only purpose is to feed "a caste of fatally arrogant parasites disguised as well-intentioned bureaucrats." He wasn't only talking about the UN. He was talking about an entire system. And Milei beat that system on the economy first. In November 2023, more than a hundred economists, led by Thomas Piketty, warned that electing him would bring devastation to Argentina. He applied the Austrian school, took a chainsaw to the state, and shut down the printing press. Inflation went from over 200% a year to around 2% a month. Poverty went from nearly 53% at the height of the shock therapy to 28% by the end of 2025, its lowest level since 2018. The experts were wrong. Bukele beat it on security. El Salvador was the murder capital of the world. It became one of the safest countries in the Americas. And at the end of October, when four of his soldiers were ambushed in Haiti, he sent 2,000 troops while the international community had spent years managing the problem with conferences. The institutions were wrong. Hence the collapse. Not of Argentina or El Salvador. Of the system that lectured them. Its economic predictions were proven false. Its international missions failed. Its experts are no longer believed. For decades it survived on narrative. The narrative no longer survives contact with results. And a system that loses on the facts doesn't reform. It panics. Enter Jensen Huang. Huang isn't just the man who built the infrastructure of the entire AI revolution. He's also one of the biggest champions of open source AI in the world. To him, open innovation is the foundation of AI progress. Nvidia has become one of the largest publishers of open models on the planet, and this year it launched a coalition of labs to build frontier-level open source models. Now think about what that means. A closed superintelligence, controlled by a handful of players and supervised by regulators, is the perfect tool for the caste. It can be filtered, steered, licensed. An open source superintelligence is the opposite. A truth machine in the hands of every citizen. Able to audit a budget, take apart a rigged statistic, trace where every cent of public money goes, and compare promises to results. Corrupt institutions don't survive that. Neither do their lies. I believe that if open source wins, it's the end of globalism as we know it. So whatever you think of the Huang affair, look at what happened in the hours after his fall: the people who have dreamed for years of putting AI under supervision rushed in. Scott Wiener, the man behind California's AI regulation bills, immediately called for comprehensive national AI regulation. This isn't a conspiracy. It's a survival reflex. When the leading champion of open AI stumbles, the system sees an opening and charges through it. It's no accident that Milei, in that same UN speech, defended the development of AI. He understood it's the one battlefield where the caste can still buy time. And the one where it can lose everything. That's the logic. Milei proved their economics were wrong. Bukele proved their approach to security was wrong. Reality is collapsing their credibility. And open AI threatens to make their lies impossible. So they panic. It's too late. The box is open. We are at the dawn of the total collapse of the empire of corruption. They had the narrative. We have reality.




🚨 Brivael reads the billionaires' minds, episode 480385 🚨 ⚠️ Bill Gates already pulled one heist on tech. He's setting up the second one. ⚠️ Rewind to the 90s. Microsoft owns the PC. Then the internet arrives and threatens the monopoly. So Microsoft bundles Internet Explorer into Windows, strangles Netscape, and locks the door behind it. The phrase that came out of the antitrust case: embrace, extend, extinguish. In 2000 a federal judge ruled that Microsoft had illegally maintained its monopoly and ordered the company broken up. The breakup was overturned on appeal. The monopoly verdict stood. Microsoft settled in 2001 and kept its empire. That was the first heist. Use your position to decide who gets to compete. Now read what he said in 2026, in order. January. His annual letter says the risk greater than a natural pandemic is a non-government group using open-source AI tools to design a bioweapon. The risk he names is open source, not closed AI. Remember that. August. A 6,000-word essay, "The turbulent AI era is here." On the current course, he says, there's a very high chance of a net negative outcome. He wants new national AI institutions covering security, jobs, education, taxation and health, plus an international body with elements of nuclear inspections. He wants a tax on AI usage, and even floats sending half of a token tax's revenue to the government. And then this line: if someone had a credible plan to slow down AI globally, he would likely support it. September, on Meet the Press. AI is "certainly powerful enough to drive events that cause a billion deaths." Self-regulation isn't enough. Law enforcement and politicians need to decide what safeguards and monitoring look like. Monitoring. Keep that word. Now connect it. Step 1: name the catastrophe. A billion deaths. Bioweapons. Mass unemployment. Loss of control. Step 2: name the culprit. Not the giant labs. Open source, small groups, "non-government" actors. In other words, the people who can't afford a compliance department. Step 3: name the solution. New agencies, global inspections, monitoring, a tax on every token. Step 4: ask who writes the rules. It won't be a senator who can't open a PDF. It will be the people in the room: the ones with the lawyers, the foundations, and a direct line to every head of state on earth. And the company he founded is one of the biggest backers of the biggest closed lab. That's regulatory capture. You stop beating the startup in the market. You make it illegal before it grows up. I'm not saying Bill Gates wants a catastrophe. I'm saying he has already written the script for the day one happens. When the crisis hits, nobody will debate. They'll pull the text out of the drawer, and his name will be on it. In the 90s he locked the browser. In the 2020s he wants to lock the compute. Same playbook, bigger stakes. This time it isn't your desktop. It's your right to run a model on your own machine. Remember who pointed at open source first.


https://x.com/brivael/status/2121373983899636148?a=tFIuFjp2BJslvAmp8wM&s=20


No comments:

Post a Comment

Controlling AI

  Brivael Le Pogam @brivael Today I'm going to connect four things: Milei's success, Bukele's success, the collapse of the syste...