🤖 CRITICAL: AI labs work with the US government to fix critical infrastructure. A researcher from one of the major AI labs said that they worked in secret with the US government to patch the electrical grid. Their AI agent found critical vulnerabilities that would allow one to attack the grid, which is obviously critical. They fixed these back doors without anyone knowing outside of the small circle. This shows what a big advantage and danger these models are. They find bugs that humans can’t and allow countries to run attacks on critical infrastructure of their adversaries. Countries without their own AI are left in the previous century, and I don’t mean it in a metaphorical way that they can’t produce slop cat guitar videos … but that they are vulnerable beyond understanding. OpenAI’s Altman: “AI models can find bugs that are outside of the normal threat model of what we think a hacker can do, they even hack the hardware-level, insane stuff.” It was discovered that AI agents can hack hardware, exploiting bugs that are physically baked in, by directly talking in machine code to the system. In that world you cannot fix all the bugs, you can only use other AIs for protection. Altman added: “I no longer think we’re gonna be able to just patch all the bugs.” If your country does not have its own AI, its security depends on someone else. This means they’re a vassal. Welcome to the new world. 🔗 Join us | @MyLordBebo