Spinellis argues that cognitive work is going the way of horse riding: economically obsolete, retained as a hobby. The mechanism isn't vibes, it's three concrete factors — performance rising across perception, reasoning, judgement, and autonomous action; cost falling generation over generation, with old models kept only for backward compatibility; and 24/7 availability plus trillions in data center investment through 2030. He notes the resistance will be fiercer than past revolutions precisely because the displaced party is knowledge workers, who hold the organizational power this time. Regulation and taxation won't stop it, per precedent. His prescription for practitioners: stay current with AI advances, be adaptable, and build a purposeful life outside work. The closing frame is the actionable part — decouple your identity from "making a living by thinking" before the market does it for you. One data point worth savoring: early ChatGPT miscounted the a's in his name, claiming two. In under four years we went from failing a character count to solving hundreds of open math problems. By that slope, by 2030 the models will be solving problems that don't exist yet, and OpenAI will charge us per problem. Bryan Johnson is reportedly already thinking as a hobby; his thoughts are biologically 18 and in bed by 8:30.