Laid-Off Worker Describes AI-Shaped Job Search After June Layoffs

At his company, the account says, the process began with a reverse list. Each team ranked all members in reverse order, with those at the front facing the highest risk of being cut. Team leaders did not mark who should be dismissed; they submitted the lists, and a professional team calculated a dynamic balance point based on human cost and token cost. Leaders learned the final scope only after the lists were returned. The author's team had a layoff rate above 40 percent, while some teams were around 10 percent, according to the account.

The layoffs were also fast. Team leaders received the confirmed list on Monday morning. On Monday afternoon, HR discussed severance packages with employees; those who did not object signed immediately, returned equipment, and left. The author's leader was told his last day would be Friday, but after the Monday afternoon conversation HR urged him to sign and hand in his computer at once, without considering handover, and said he did not need to come in from Tuesday. At the end of the workday, his account was deleted on time, work software was forcibly logged out, and the whole process was finished within four hours, the account says.

Throughout June, the author heard of other companies' 630 layoffs moving from rumor to reality. He wrote that competition between people and AI in the workplace appeared to be accelerating toward a white-hot stage, and that this seemed to be only the beginning.

When he restarted his job search, the first step was a résumé, which he said he hated writing. He described using AI as a three-step process. First, he told AI everything he had done in each job, from project results to problem-solving methods and details, to build material and review his own experience. He suggested voice input, or using WeChat's voice-to-text, if typing was too slow. Second, he asked AI to distill and summarize his story, using a familiar structure such as background-execution-result if desired. Third, he found several target job descriptions and had AI revise the distilled version to match those roles, so the résumé would better reflect what employers wanted and help him target suitable positions.

The author advised keeping analysis and writing separate. He said job seekers should ask AI to return its analysis and conclusions at each step for review before writing, and should correct the model if it overinterpreted or expanded on certain terms in the source material.

After preparing the résumé, he applied through recruitment apps, company websites, and even Xiaohongshu. Once he updated both the attached and online résumés on recruitment apps, headhunters began contacting him. During the search he added no fewer than 50 headhunters on WeChat, each claiming a different main recruiting direction. He sorted them into several types.

The first type called just to meet a daily call quota. He said his résumé clearly stated that his AI projects involved only application features, not data labeling or algorithm optimization, yet many AI-focused headhunters immediately asked about his technical experience and seemed not to have read the résumé. After a few irrelevant questions, they ended the call. These headhunters looked only for keywords and treated him as an exposure pool, he wrote, and job seekers should not waste courtesy on them.

The second type was condescending. They asked about specific work and project experience, then belittled some projects as insufficiently detailed and not worth discussing. The author said he had not even put those projects on his résumé, but the headhunter insisted on asking. Later, the headhunter said the résumé and work experience were barely worth his attention, that sending it to HR would waste time, and finally said the author had no path planning and no company would hire him. The author guessed the headhunter had held overly high expectations and then become disappointed.

The third type turned preliminary communication into a mock interview. They asked detailed follow-up questions about the résumé, including each job's details and reasons for leaving, and interrupted to ask more about keywords during project descriptions. One phone call lasted an hour, and at the end the headhunter said the match was not high enough. The author said he had not heard of headhunters stealing candidates' project experience, but these experiences made him wonder whether he was the first unlucky person to encounter it.

The fourth type was reliable. These headhunters first confirmed whether keywords in the résumé were accurate, then briefly described the job requirements and their judgment of the match. Finally, they explained what points their recommendation would focus on and asked whether the author agreed or wanted to adjust or add information. The whole exchange took about ten minutes. Even if no job was recommended, the author said, the process helped him sort out his own direction. Such headhunters were rare among the dozens he met.

On recruitment platforms, the author saw a clear change from three years earlier: algorithmic recommendations were widespread, and the system invited him to apply to eight to ten different company positions almost every day. The system sent a job invitation; if he was interested, it automatically submitted his résumé. But résumés submitted automatically by the system remained low priority in corporate HR screening, he wrote. HR spent more time actively searching for résumés by criteria rather than trusting system recommendations. He advised job seekers not to let recommendation algorithms consume too much time.

He also encountered an AI phone call. After a platform's repeated invitations failed, it triggered a phone-application plan. One afternoon he received a call from a robot that sounded almost human. It made small talk and even gave a mechanical but polite laugh before saying a fast-moving consumer goods company had a suitable position and asking for basic personal information. During the questions, there were obviously incorrect pauses and broken sentences, which convinced him it was AI. By then his résumé had already been submitted, and the result was no result. He wrote that none of the positions that advanced to an interview had started from a system recommendation, and that ill-timed recommendations or calls made job seekers uncomfortable. Recruitment app recommendations were a mix of good and bad, and job seekers should reduce such information loss, he said.

The author also described a shift in job requirements: most positions now ask for AI-related experience, regardless of the field or direction. He saw three features. First, the more easily a job can be replaced by AI, the higher the demand for candidates' AI application experience. In his own applications for data product manager roles, work such as metric organization and SQL queries is already handled well by AI. Companies almost all asked whether he had used corresponding skills and agents at work, and some stressed that they valued a candidate's ability to independently build AI workflows, even asking detailed construction questions. He suggested that everyone try independent projects with AI, because such work could become a portfolio-like calling card in job searches.

Second, years of experience were not tied to AI experience. Although AI's application boom in many jobs had lasted only about half a year to a year, employers' requirements for AI experience were not limited to that period. The account then began to give what it called the most extreme example, but the supplied text ends at that point.