The post discusses an analysis of AI startups from 1999, particularly examining the boom and bust cycles of the AI industry. Key points include the technological limitations of AI in the 1980s, driven largely by custom processors and concurrent advancements in computing power, as suggested by Moore's Law. Current trends show a significant shift as AI technology is increasingly embedded into everyday applications, often under different labels like face recognition and autopilot systems, once they become normalized. Users express a belief in a continued upward trajectory for AI productivity, with investments expected to yield measurable returns. There’s also speculation about the potential for a repeating cycle of hype and disillusionment, drawing parallels with historical patterns in AI development.