I think I’ve finally figured out the right way to work with AI – The essence of working with AI is *investment*, and the endgame of investment is the S&P and Nasdaq White-collar work used to be about converting your intellectual resources into productivity. But AI’s productivity in most scenarios trends toward infinity. So all you need to do is decide how to allocate AI resources, which is essentially an investment decision. For most people, trying to use AI to generate returns is basically like picking individual stocks. And because of limited attention, you can only hold a small number of stocks, which means you might not keep up with the whirlwind pace of this era. At the same time, institutions like quant funds, with their capital and scale advantages, dominate the key nodes in the value chain. This makes it much harder for individuals to achieve excess returns. Take short dramas, for example. The intellectual resources of creators now account for only a tiny fraction of the value chain. Both production and consumption are heavily supported by AI in the hands of giants like ByteDance. The share of profits creators get is almost negligible compared to traditional novels, films, and other creative teams. If ByteDance were to go public, and short drama investors had to choose between starting their own projects or buying ByteDance stock, the average return from buying the stock would probably far exceed the returns from producing short dramas. So, using AI yourself to generate returns is unlikely to outperform the model of sharing in the value of key nodes through index investing. Thankfully, the U.S. stock market, with its centuries of development, has become relatively fair, giving ordinary people the chance to share in the returns of these key value nodes. I hope this opportunity remains open in the future. Will it?