On the Eve of AI Prosperity
2026-11-01 · 中文版
A few days ago I read an article by a big account on X, titled “On the Eve of the AI Bubble Bursting.” It was well written, and the facts were solid: Intel fell 28% in a single day, Google lost $500 billion in market value in a day, SanDisk fell 13% despite beating earnings, Reddit lost 29% because Google’s AI Mode answers questions directly and sends no traffic. His core argument is this: AI has raised productivity enormously — a year of coding now matches the past ten — but as of October 31, 2026, he sees no genuine “incremental value” that AI has created. Intelligence is no longer scarce, and what is not scarce has no value; with no middlemen earning a spread, the cycle of trade stalls; the money earned all flows back into AI infrastructure. So AI is neither 2B nor 2C, but 2me — a self-pleasing economy.
I have to admit, the pain points he describes are real. I use AI to write code and do analysis every day, so I know that feeling well: productivity is up, but the money has not followed. Blueberry output rose 39% and the price fell by half — the people buying blueberries are happy, but what about the people growing them? He is not wrong about that.
But I think his conclusion goes too far.
He equates “producers have not made money yet” with “society has no incremental value.” That leap is too big. Blueberries getting cheaper benefits blueberry buyers — that itself is value. AI makes things worth doing that were not worth doing before, lets a small team build what used to require a large company, and accelerates research, medicine, and education. All of that is incremental value; it just lands first in the hands of consumers and users, not on producers’ profit statements. Value does not only come in the form of “sellers earning more.”
More importantly, history has already run this script several times. Railways, electricity, the internet — which one did not go through bubble, then crash, then real value creation? When the railways were built, countless railway companies went bankrupt, yet the value the railways themselves created was enormous. When the internet bubble burst, a pile of companies died, but the internet itself did not die; it just changed hands to a new set of people to harvest. A bubble is not a disease. It is how infrastructure gets financed — except the price is that a group of people always has to fall in the darkest hour before dawn.
So my view is this: AI certainly has a bubble, but that is a good thing, and it is inevitable in a period of technological breakthrough. AI’s change to human productivity is enormous and irreversible. We are not standing on the eve of the bubble bursting. We are standing on the eve of AI prosperity. The two look alike — everything in ruins, no money in sight — but they point in opposite directions.
In this process, many people and companies will fail and lose confidence in the darkest hour before dawn. That is real, and it is the hardest part. I do not deny it. But those who make it through will have a bright future. The question is not whether AI has value. The question is whether you are still standing when the value arrives.
So people and companies should face it with a positive mindset. Not blind optimism, but deliberate action after seeing the pattern clearly: the bubble will burst, but the technology will not go backward; productivity has gone up, and it is not coming back down. What we should do now is build our skills and hold our ground in the dark, and wait for daylight.
Dawn is coming. Do not fall asleep before it breaks.