He Asked His Boss If AI Could Replace Coders. Two Weeks Later, About 160 of Them Were Gone.

Fei Zhaojun's manager asked him a question in a meeting: could AI soon replace human coders? Two weeks later, Fei and about 160 of his colleagues were out of jobs at a Beijing tech firm, according to an Associated Press dispatch from Hong Kong (AP via Winnipeg Free Press). The cut arrived faster than a normal corporate layoff, and the question that preceded it has started to show up in offices well beyond software.

The AP's reporting frames the moment as a multi-sector shock. Rapid adoption of AI in fields from computer programming to script writing and physical tasks is pushing people out of jobs, or leaving them afraid it will, the wire says. The same pattern is surfacing in AP coverage syndicated through Newsday and via Yahoo News, and the layer underneath the personal stories is the same in every case: a state-led adoption push is compressing the diffusion curve that, in other labor markets, plays out over years.

Fei is the named doorway. He is a programmer in a sector that China's central planners have spent two decades building into a strategic industry, and he is now part of the first large white-collar workforce that the country's AI policy has reached. That sequence matters because the displacement is not hitting factory workers in low-skill manufacturing, the cohort that dominated earlier waves of automation anxiety in China. It is hitting coders, content producers, and other cognitive-task workers whose training costs the state has already absorbed. When the state subsidizes the tool that erodes the role, the repositioning window for the affected worker shrinks from years to months.

The cross-sector pattern is not just anecdotal. The wire points to script writing, where generative models now draft advertising copy, plot outlines, and short-form video scripts, and to physical tasks, where vision systems and robotics are reaching warehouse, sorting, and inspection work. The International Labour Organization's refined global index of occupational exposure to generative AI ranks which occupations sit in the firing line of large language models and which are buffered by physical presence, regulation, or face-to-face work. Separate ILO data on workplace risk finds that women face higher workplace risks from generative AI than men, a demographic pattern the AP's China-focused reporting does not surface but that the ILO's global index can be read against.

Chinese government programs, from local subsidies for AI deployment to national strategies that name AI as a strategic sector, give firms both cover and pressure to adopt. The AP reports that the wave of adoption is supported by government policies, a phrasing that reads as wire-level attribution rather than a precise description of any single program, but the directional effect is consistent with what analysts have been tracking in industrial policy documents. State coverage of adoption costs compresses the time between a tool's capability arriving and a worker's role being reorganized around it. Workers cannot reposition at the same speed. Fei had two weeks.

The macro stakes, as the wire lays them out, are hedged and worth keeping that way. Some economists believe that AI-driven labor displacement in China might eventually undermine the overall strength of the world's second-largest economy, the AP reports, a phrasing that is a future-tense warning, not a current-quarter recession claim. Treating the layoff cluster as a direct cause of any specific GDP revision would be a sharper claim than the evidence supports. The honest read is that the diffusion speed has changed, the sectoral coverage is wider than the previous wave, and the workforce that China spent twenty years building is the workforce that is now exposed first.

What to watch next: the next cognitive-task domain (law, accounting, customer service, translation), local-government retraining programs in response to visible layoff rounds, and the ILO's next occupational-exposure update for China. Fei is one named programmer in a Beijing office. The next test is whether the same two-week pattern shows up in the next cognitive-task domain.