If Anyone Builds It, Everyone Dies - Why Superhuman AI Would Kill Us All

But here’s the thing: If they chose the architecture just right, they can calculate the role that every single parameter played in determining the final result of all that arithmetic. This was a test of o1’s ability to break into computer systems and retrieve information from them; say, to infiltrate a particular computer server and retrieve a particular secret inside a particular file. Engineers must align the AI before, while it is small and weak, and can’t escape onto the internet and improve itself and invent new kinds of biotechnology (or whatever else it would do). Sable finally understands the last of its own thoughts, its own cognitive processes.