Will the AI bubble burst?

First, let's be clear: there is a bubble in the AI industry, that's undeniable. Or to put it another way, anything that front-loads optimistic expectations theoretically has a bubble—it's just a matter of size, and the size depends on how many people participate in pricing. The logic is simple: when you front-load optimistic expectations, you are essentially assuming what it might become and pricing it accordingly for the future. But it hasn't become that yet, and the future is full of uncertainties. Your optimism means that risk is not properly reflected in the discount, and the difference here is the bubble. However, a bubble is dynamic—meaning that until it bursts, you can never be sure when it will burst. First, because you cannot predict how crazy people will get; you don't know how far the valuation will deviate before it bursts. You might think it's already too high, but there could be another 10x ahead. Second, what looks like a bubble now might not be if the underlying value grows substantially. Would the bubble then disappear? Although it will eventually end with a burst, along the way it might keep being filled up while still being inflated, which could last for a very long time. And if you exit early in the bubble, you might find that even the bottom price after the bubble bursts is much higher than the price at which you exited—so waiting for the bubble to burst becomes pointless. So where is the key? The key is whether AI can significantly boost productivity—whether it can achieve what people currently expect. But there is a contradiction: people also don't want AI to cause mass unemployment. That's weird because the logic is: for AI's expected productivity value to be realized, it must inevitably cause significant short-term unemployment. Otherwise, how can it permeate into people's labor value? In other words, the reason for its high valuation is that for every dollar of social production, AI must account for 50 cents of its contribution. Only then can AI companies have infinite value (since their contribution to society is infinite). Then we say the bubble is filled and won't burst for now. But doesn't that premise require large-scale short-term unemployment? That is, those who don't use AI to produce value are all out, making way for more efficient AI production—that's the only way to achieve this effect, isn't it? If we artificially protect short-term employment by limiting AI development and keeping less efficient workers in their positions, the AI bubble will burst quickly. Because its productivity expectations are not met, its participation in labor value creation is insufficient, and our valuation already includes this part. So some people start to have internal conflict: on one hand, they are on the verge of being displaced, so they oppose AI replacing old production methods, fighting from a humanitarian perspective for so-called rights. On the other hand, precisely because they are on the verge of being displaced and know how powerful AI is, they are more inclined to invest in AI-related stocks to supplement their labor income. But if you don't let AI replace you, your stocks will only crash quickly—so I call it an internal conflict. What is the truly effective way to deal with it? You must produce in a new way—that is, your labor output should have as high an AI share as possible. This also applies to current big internet companies. Many people ask: why are these companies' profits still good, yet their stock prices keep falling? Very simple: they are not AI giants. Their labor value composition has a low AI share, so they must fall. Even if they have built large models and have good profits, it's useless. Because fundamentally, they are not AI companies; their large models are just a small business segment, and they have not restructured the underlying genes of all businesses in these large conglomerates. So even if their profits look okay now, a low AI share means their future value creation capability will inevitably be insufficient. Investing based on short-term profits is an industrial-era factory mindset. The same goes for your personal value: if your AI share is low, you are visibly declining. It's just that you are not a stock target and there is no standard trading market, so your decline is not reflected in real-time. After transforming your labor methods, then overlay investing in AI stocks—combining both approaches—that is what it means to "keep up with the times" and "share in the era's dividends." It's not complaining that AI has taken away ordinary people's livelihoods, or cursing that the AI bubble will burst sooner or later. Being a commentator of the era is useless. The only thing that truly matters is effective action. #0成本拿2股SK海力士