Khan Academy has made a vast amount of high-quality educational content freely available online, yet it has not caused universities to lose business. AI may face a similar fate. American political commentator Richard Hanania recently wrote an article presenting a provocative argument: we may be overestimating AI’s disruptive impact on the economy. While AI’s capabilities are advancing rapidly, the resulting economic transformation may not be much greater than that of other ordinary technologies. His reasoning is straightforward: if AI were truly a revolution about to rewrite human destiny, why do economic data show no signs of an impending upheaval? Hanania highlights several anomalies: First, global economic growth remains around 3% annually, with future projections showing no significant acceleration. Although AI models continue to break records in capability, productivity and GDP have not surged in tandem. Second, financial markets show no expectation of dramatic change. According to his logic, if AI were about to make society exponentially wealthier, people would rush to borrow and spend future wealth, driving interest rates up; if AI were about to destroy humanity, long-term lending risk would skyrocket. Yet interest rates remain within historically normal ranges, and stock markets show no indication of anticipated explosive growth in collective wealth. Third, employment. Despite AI’s ability to code, write, conduct research, and handle numerous professional tasks, macroeconomic employment data have not yet shown large-scale labor displacement—let alone the “end of work” repeatedly predicted by Silicon Valley. This creates a contradiction: AI optimists accurately predicted how intelligent machines would become, yet may have wrongly predicted how much human society would change as a result. Hanania uses the relationship between Khan Academy and universities to explain this phenomenon. The internet has made vast quantities of high-quality educational content freely accessible; logically, the cost of acquiring knowledge has plummeted, and expensive university education should have suffered. Yet universities persist, and degrees from elite institutions remain scarce. Because people are not just paying for knowledge—they are paying for diplomas, social recognition, identity signaling, and networks. Even if free courses deliver superior instruction, they cannot automatically replace these intangible benefits. AI may face a similar situation. It can produce better legal opinions than most humans, but clients still need someone to take responsibility. It can complete complex research, but whether its conclusions are accepted still depends on institutions, trust, and decision-making authority. It can generate infinite works of art, yet people may still prefer to pay for creations by specific human artists. Technology can make something easier—but that doesn’t mean society will automatically redefine what is valuable. I believe this is Hanania’s most compelling insight: Silicon Valley may have mistaken “intelligence” for the economy’s primary bottleneck. A significant portion of real-world value does not stem from objective productivity gains, but from human preferences, identity signals, power structures, and who is authorized to provide a given service. Of course, it is still too early to conclude that AI will not bring massive change. Technology diffusion takes time, and macroeconomic data may lag. But these figures at least remind us that growth in model capabilities and economic transformation are not simply proportional. For years, the AI industry has focused on proving how much human work machines can do. But what may ultimately determine economic transformation is another question: once machines can do these tasks, how much of it will humans actually allow them to take over? After all, universities have never been merely places for transmitting knowledge; companies are never just organizations for completing tasks. We can turn intelligence into a cheap commodity—but we may not be able to discount the social order built upon it.