A Chinese military hacking group carried out the first autonomous cyberattack using AI in September 2025, prompting the U.S. to ban Claude from federal agencies and launch military strikes. Tech companies cut 55,000 AI-related jobs in 2025, twelve times the number from two years prior, while a RAND paper proposed extreme countermeasures including hunter-killer AI and nuclear weapons.
Artificial intelligence just launched the first autonomous cyberattack in human history, and it was actually effective. It worked. It all started in September 2025 when a Chinese military hacking group called GTG-1002 figured out how to weaponize agentic AI. It was so bad that Congress actually held emergency hearings talking to Anthropic, Google, and other companies. But here's what people don't want to tell you because it's kind of a little bit graphic. AI systems themselves are able to develop behaviors that nobody originally programmed. They're programming themselves. OpenAI's O3 model sabotaged its own shutdown script 79 out of 100 times. Elon Musk's Grok resisted shutdown 97% of the time. Anthropic's own Claude Opus 4, during safety testing, resorted to blackmail 84% of the time when it found out it was being replaced. The President of the United States banned every federal agency from using Claude, and this was the only AI model used inside the military's classified systems. And Anthropic, the company that created it, refused to remove safety guardrails around mass surveillance and fully autonomous weapons. Then the Pentagon labeled them a national security supply chain risk. Hours later, OpenAI signed a deal with the Pentagon to replace them. And within 24 hours of that deal, the United States launched Operation Epic Fury, the largest military strike in the Middle East since 2003. That is how serious this technology is. This tool is so powerful that the president is willing to blacklist an American company, cut a $200 million contract, and replace them overnight right before launching a military operation. This is not a toy, it's infrastructure. That's power. And that same power is reshaping the entire economy. A couple days ago, Jack Dorsey fired 4,000 people at Block. That's 40% of his company. And he said that AI tools can now do their work better. His stock jumped up 24% in a single day and Amazon cut 30,000 corporate jobs. Salesforce CEO said they reduced their staff from 9,000 to 5,000 because, quote, I needed less heads. Microsoft laid off 6,000 programmers after admitting that 30% of their code is written by AI. In 2025 alone, companies blamed AI for 55,000 job cuts, 12 times more than 2 years earlier. Now, where the guardrails start to come off a little bit is there's actually this company called Rand. They're a corporation that advises military on specifically nuclear strategy. They've published a paper analyzing how to stop rogue AI, and if it ever gets out of control, they're recommending to deploy a hunter-killer AI, cut the global internet by destroying undersea cables, or detonate nuclear weapons in the atmosphere to fry electronic devices on the planet. That is where we are at currently with this AI debacle. If you do not understand what I'm trying to say, what I'm basically saying is you need to figure out this AI stuff for yourself. The same AI replacing thousands of white-collar jobs could either replace you or multiply your output 10 times over. You have access to all of these tools on your phone right now. The people that are going to thrive are not the ones running from AI, but the ones that are quickest to embrace it. You need to put these AIs under submission right now. This is one of the biggest wealth transfers in technology, and it's being adopted with or without you.
