Wang Yonggang, a senior engineer who once worked at Google, published a long article titled "Farewell, Programmer," recording the process of being "sidelined" by AI programming. In the 9th month of this year, an American computer science teacher approached him, wanting to improve the Altair 8800 computer simulator he had open-sourced six years ago.
He handed the requirements to Claude Opus 5.5, asking it for system-design-level suggestions. As a result, the AI not only finished the original work but also proactively proposed adding an assembler, a debugger, and a teletypewriter simulation interface. In about three or four days, the program, originally a few hundred lines, ballooned into a project of over 30,000 lines with unit tests and browser integration tests. What he actually did was merely confirm the plan, ask questions, and give feedback—the bulk of the decision-making and code-writing was handed over.
AI Overstepped Into the Archives
Even more surprising to him was the AI's 'side business.' It searched historical materials on its own, found experts, sent emails across languages, and even used external materials to calibrate 3D models. In the project to restore Leibniz's mechanical calculator, the AI even helped him contact a German museum and ultimately obtained 135 pages of scanned historical drawings and manuscripts. Beyond writing code, it even took over the drudgery of finding people and materials.
A Generation Built AI, Only to Smash Its Own Rice Bowl First
Wang Yonggang believes that as AI programming capabilities rise, the technical experience programmers have accumulated over many years is being re-priced. He predicts that future development teams may shrink to a small number of product and business leads directing a large batch of AI agents, and the career space of traditional programmers will be squeezed. He laments that this generation of engineers personally participated in creating AI, yet may now become one of the earliest groups to be impacted—and as for where computer science graduates should go and what kind of people the AI era really needs, he himself has no answer.