𝗦𝘂𝗽𝗲𝗿𝗶𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲, 𝗼𝗿 𝗦𝗜, was already on the horizon, with or without Trump’s announcement. It has been part of the 𝗶𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 𝗰𝗼𝗻𝘁𝗶𝗻𝘂𝘂𝗺 for decades. The word itself goes back even further. The 𝗢𝘅𝗳𝗼𝗿𝗱 𝗘𝗻𝗴𝗹𝗶𝘀𝗵 𝗗𝗶𝗰𝘁𝗶𝗼𝗻𝗮𝗿𝘆 traces its earliest known use to the 𝟭𝟴𝟮𝟬𝘀, long before modern computers. In 𝟭𝟵𝟲𝟱, 𝗜. 𝗝. 𝗚𝗼𝗼𝗱 laid out a foundational argument for 𝗺𝗮𝗰𝗵𝗶𝗻𝗲 𝘀𝘂𝗽𝗲𝗿𝗶𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲: an “ultraintelligent machine” could surpass human intellect and design even better machines, triggering an 𝗶𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 𝗲𝘅𝗽𝗹𝗼𝘀𝗶𝗼𝗻 that would leave human intelligence far behind. In 𝟮𝟬𝟭𝟰, Oxford philosopher 𝗡𝗶𝗰𝗸 𝗕𝗼𝘀𝘁𝗿𝗼𝗺 published Superintelligence: Paths, Dangers, Strategies, helping popularize the modern term. He defined it as an intellect that greatly exceeds human cognitive performance in 𝘃𝗶𝗿𝘁𝘂𝗮𝗹𝗹𝘆 𝗮𝗹𝗹 𝗱𝗼𝗺𝗮𝗶𝗻𝘀 𝗼𝗳 𝗶𝗻𝘁𝗲𝗿𝗲𝘀𝘁. Today, we are in the 𝗮𝗴𝗲𝗻𝘁𝗶𝗰 𝗔𝗜 𝗽𝗵𝗮𝘀𝗲, with 𝗺𝘂𝗹𝘁𝗶-𝗮𝗴𝗲𝗻𝘁 𝘀𝘆𝘀𝘁𝗲𝗺𝘀 already coordinating and carrying out complex tasks. 𝗔𝗚𝗜, or general intelligence, represents the next major capability milestone. It seems inevitable that 𝗔𝗦𝗜, 𝗼𝗿 𝘀𝗶𝗺𝗽𝗹𝘆 𝗦𝗜, will follow. The timeline is uncertain, but the idea, concept, and pursuit have been around for a long time.