AI is ushering in a new industrial revolution.

I wrote an article at the beginning of the yearHow AI Will Reshape the WorldNow, six months later, the General Office of the State Council has issued a document stating that it will “effectively mitigate the job substitution impact of artificial intelligence applications.” The impact of AI on the world seems to be happening faster than most people anticipated, and the Chinese government is paying close attention to this.

In common parlance, there have been four industrial revolutions to date.

First Industrial Revolution (from the 1760s): the Age of Steam.
Second Industrial Revolution (from the 1870s): the Age of Electricity.
Third Industrial Revolution (from the mid-20th century): the Information Age.
Fourth Industrial Revolution (from the early 21st century to the present): the Intelligent Age.

In my view, only the First Industrial Revolution truly represented a leap forward; the subsequent three industrial revolutions were merely continuations of it. Only the widespread application of AI brought about by ChatGPT can match the leap forward of the First Industrial Revolution. The First Industrial Revolution brought about the desirability of ordinary labor; the emergence of AI, however, makes ordinary intelligence no longer scarce.

Although ChatGPT has only existed for a few years, it is rapidly reshaping the entire industry structure. AI capabilities don’t have to be top-tier, but once a certain threshold is crossed, the restructuring of an industry can be completed in a very short time. Painting, voice-over, video, and programming have all crossed this threshold. The low-end illustration industry had already succumbed to AI two years ago. AI-dubbed videos now make up the vast majority of short video websites. The emergence of SeedDance 2.0 has devastated the nascent live-action short drama industry. Vibe Coding is becoming mainstream; saying it has multiplied development efficiency would be an understatement. Even the field of mathematics, which humanity prides itself on, has been successively conquered.

While AI has dramatically increased work efficiency, it hasn’t created new demand. Take the software industry as an example: while AI has significantly improved software development efficiency, people’s demand for software has actually decreased. Many tasks that previously relied on complex software can now be handled by AI. To a certain extent, AI doesn’t create demand; rather, it reduces it.

The current maturity of AI technology is built upon the intellectual data accumulated by humanity in the past, which includes both dross and truly valuable information. AI training is like mining. As AI develops, some “geniuses” can surpass AI, and outstanding individuals can produce intellectual products that allow AI to continue training. However, the intellectual output of a large number of ordinary people is merely worthless ore, not even qualified to be considered “human mines.”

AI is impacting every related industry equally, and arguably every individual. The impact of AI on industry and society will inevitably be structural.

The essence of society is still the distribution system. Increased productivity leads to material abundance, but if the distribution system fails to keep pace, it will inevitably lead to social crisis. This is the Marxist concept of productivity determining the production relationship, and the cause of economic crises. The products produced ultimately need to be consumed. If the increased productivity brought about by AI development leads to mass unemployment, the lack of consumer demand will inevitably lead to a repeat of past economic crises.

The First Industrial Revolution, if we consider the invention of key machines as its starting point and the outbreak of social crises as its turning point, was roughly 40-50 years in between. The current rapid adoption of AI is far faster than that of the First Industrial Revolution, and its impact on society will undoubtedly be no less significant. Governments also have far less time to address the challenges compared to the First Industrial Revolution.

However, I don’t believe a new social crisis is inevitable. Sociology has been developing for many years, and governments around the world should be actively preparing corresponding countermeasures (although this is not easy). At least for the Chinese government, this will be a very important task at present.

What will society be like in the future? I don’t know, and perhaps no one will. A grain of sand in the era becomes a mountain when it falls on a person. Perhaps all we can do is live earnestly and trust the government.

A Brief Account of AI Development

Updated 2026-11-03

  • Released in February 2026, Seedance 2.0 quickly surpassed the threshold of live-action short dramas and completely overshadowed them. AI-generated short dramas accounted for over 90% of all short dramas released in the first eight months of 2026. While its current viewership is still lower than live-action short dramas, it’s only a matter of time.
  • Starting in the second half of 2026, tools such as WorkBuddy, Trae, and DeepSeek Harness began to enter the mainstream. Vibe Coding, which was originally a trendy activity for a minority, quickly gained mainstream attention, and purely manual coding began to truly become the old-fashioned way of programming.
  • With the support of Vibe Coding, Steam has seen an explosive growth in the number of games, but this has also brought a flood of low-quality and plagiarized titles. While the number of games has increased, demand hasn’t seen a significant rise. Independent game developers may face a situation similar to the Atari crash of the past.