Dylan Litman 检视 Meeker 1995–1997 年《Internet Trends》中的 86 项预测，并以 Statista、美国人口普查与政府报告等资料对照，将其分为保守、准确、部分准确或错误。虽然不少预测属质性判断、评分难免主观，但整体呈现清楚的趋势：许多看似过度乐观的推估，后来不仅方向正确，幅度还往往被现实超越。
例如她对全球渗透率、使用者数、电商与数位广告的预估多被实际结果超过；她也提醒投资热潮会伴随波动、随后出现洗牌与整并。她的错误主要在于难以押对赢家公司：对哪些企业会主导线上零售与浏览器市场的判断多有偏差。作者因此总结：描绘技术长期路径通常比挑出最终受益的公司容易，而短期狂热可能退潮，但长期影响往往能远超最雄心勃勃的预测。
The piece revisits the dot-com boom and bust, noting how the Nasdaq’s dramatic rise to a March 2000 peak and subsequent collapse is often cited today as a cautionary tale for AI enthusiasm. Yet, with hindsight, the authors argue that one of the era’s most criticized boosters—Morgan Stanley analyst Mary Meeker—was surprisingly accurate about the internet economy’s direction and scale, if anything understating its eventual growth.
Dylan Litman reviews 86 predictions from Meeker’s influential Internet Trends reports (1995–1997) and compares them with later data from sources like Statista and U.S. government statistics, rating each as conservative, accurate, partially accurate, or wrong. While many calls were qualitative and thus partly subjective to grade, the broad pattern is clear: forecasts that seemed wildly optimistic at the time often proved directionally right and, in many cases, modest relative to what actually happened.
Meeker’s projections on adoption, users, e-commerce, and digital advertising were frequently surpassed, and she also warned that investor enthusiasm would swing into periods of depression, followed by shakeouts and consolidation—an insight consistent with the dot-com crash. Where she struggled was in picking the winning companies, misjudging which firms would dominate key markets even when she correctly identified the markets themselves. The takeaway for AI is that it may be easier to chart a technology’s broad trajectory than to pick its champions, and that long-run impacts can outstrip even ambitious forecasts despite short-term volatility.