The piece reflects on the evolving landscape of AI and computing, set in late 2025. It opens with the author’s personal experience while running, which serves as a backdrop for broader technological themes. The author notes a shift in attitudes among engineers who initially doubted AI's capabilities, particularly the Compute Theory of Everything, which posits that intelligence is fundamentally tied to processing power rather than symbolic reasoning. Many engineers who were once skeptical have begun to adopt AI tools in their work, recognizing their effectiveness. The author cites Hans Moravec's foundational essay from 1976, "The Role of Raw Power in Intelligence," which argues that the evolution of intelligence is closely linked to computational resources. Moravec illustrates this point by comparing the speed of human optic nerves to that of early computers, highlighting a significant disparity that hampers scientific progress. He criticizes the AI research community for ignoring the compute deficit, suggesting it limits innovation. The essay emphasizes that true advancements in AI require acknowledging and leveraging the raw power of computational resources. Moravec's ideas hinge on historical context. He describes a period from 1960 to 1990 known as “the big freeze,” during which AI research stagnated due to limited computing resources despite advancements in hardware. This stagnation contrasts sharply with recent years, where rapid improvements in computational power have fostered breakthroughs in AI capabilities. The author suggests that understanding these dynamics is essential for grasping the current and future potential of AI technologies.