The case for hybrid work weakens once you split the data by tenure

The most useful thing in this material is a subtraction. The global hiring study the report leans on found early-career workers taking a substantially smaller share of new hires by 2025 in occupations easiest to do remotely ^[6], and the obvious rival explanation for that pattern is AI. When the researchers tested it, the AI effect largely disappeared and the remote-work association held ^[7]. That is the right test to run, because the two stories predict a similar aggregate curve and demand opposite responses from an employer. It is also not a clean verdict, for reasons the summary itself makes visible: it does not say how remote capability was scored, which countries and years the sample covers, or whether the firms that went remote were already hiring differently.

The proximity result is the closest thing here to a mechanism. Software engineers seated near their teammates received 18.3% more feedback on their code and produced higher-quality code ^[4], and the gains concentrated among younger and less-tenured staff ^[5]. That gradient is what makes the finding worth attention: if sitting nearby were merely a marker of being a strong engineer, you would not expect the advantage to fade as people accumulate experience. What the number does not measure is any career outcome. Feedback volume is a proxy for the thing the report actually cares about, which is tacit knowledge, hard to codify and picked up through observation and social interaction ^[15].

The Microsoft study, covering more than 61,000 employees, reported that companywide collaboration grew more siloed under remote work, with communication concentrating inside existing networks and thinning across organizational boundaries ^[13]. That result reads differently by tenure. An employee with a decade of contacts loses some breadth; an employee three weeks in has no existing network to concentrate into, and professional development depends partly on reaching people who supply feedback and opportunities ^[17].

None of this surfaces where firms look. Retention and performance survive hybrid arrangements in the existing research ^[2], and those are the outcomes with dashboards attached. Whether a junior employee has developed judgment, built relationships across the organization or learned to work a problem out alone is not instrumented ^[14]. The people concerned appear to register the tradeoff: only 23% of remote-capable Gen Z employees in the United States say they would prefer fully remote work, against 35% of older generations ^[9], a 12-point gap running against the usual assumption about young workers and flexibility ^[12]. Deloitte's 2025 survey puts half of Gen Z respondents wanting a manager who mentors them and 36% saying they receive it ^[10], a shortfall of 14 points ^[11].

So, a view with its condition attached. Hybrid work is better evaluated by tenure band than firm-wide, because the outcomes that hold up are measured largely on employees who already possess what a junior is trying to acquire. The condition is that this is a synthesis rather than an experiment, and the hiring evidence it rests on is described as emerging ^[6]. If entry-level professional roles keep landing on-site while flexibility accrues with seniority ^[8], that sorting is already underway, and the pipeline problem follows from it: an employer who buys finished skills rather than building them is relying on someone else to do the building ^[16].