Industrial revolution of intellectual work

Before AI, education trained us to learn skills, find a job, apply those skills, and create value. It took years - sometimes decades - to acquire domain knowledge, only to become a skilled worker following boss’s commands.

Now, with AI, human labor’s role shifts from executor to planner. AI agents do the execution. In my daily programming work, 90% of tasks are done by AI - while I only take care of debugging and ops that it can’t yet handle well.

It’s as if everyone now has their own AI intern: you plan, it produces. The better your “intern,” the less time you need to spend.

• Foundation models like ChatGPT, Claude, Gemini, DeepSeek are improving the intern’s general intelligence - think of it as progressing from high school to university to graduate level.
• Techniques like RAG, Prompt Engineering, and Context Engineering are ways of teaching this intern specialized domain knowledge.
• The explosion of AI agents: PPT generation, image and video creation - makes this intern increasingly versatile, capable of handling more and more of the work. Ideally, one day, nearly all of it.

This model is quietly revolutionizing the structure of human society. Just as the Industrial Revolution transformed textile workers into machine supervisors, the AI revolution turns knowledge workers into intelligence supervisors. Where we once crafted every slide, line of code, or report by hand, AI now drafts most of it - faster, better, and often with the same level of quality.

The industrial revolution of intellectual work has already begun.

---