Labor economists have long debated whether technological change favors generalists or specialists. For much of the past decade, breadth was the safer bet for engineers: knowing many languages and tools kept a candidate adaptable. Recent labor market data suggest this assumption may no longer hold. An analysis of tens of millions of technology job postings since 2024 points to a shift away from breadth and toward depth. Employers now ask for fewer distinct skills per role, but more years of experience in the skills that remain.

From Breadth to Depth

The signal is quantitative. Across a large sample of postings, the average number of distinct skills listed fell by about a quarter, from roughly thirty to twenty-one, over two years. In the same period, average required years of experience per skill rose slightly, from about 4.6 to 4.8. Together, these trends describe a market asking candidates to know less, but to know it better.

This marks a departure from the "laundry list" posting that drew criticism in past years, where dozens of loosely related requirements served as a filtering mechanism rather than an honest job description. Employers appear to be moving away from that pattern, trading broad enumeration for narrower, more demanding criteria.

Which Skills Are Gaining, and Which Are Fading

The composition of demand has shifted too. Generic categories like "software development" or "application development" appear in a shrinking share of postings. In their place, more specific competencies have gained ground: particular API frameworks, individual programming languages, cloud deployment, and applied machine learning.

Depth requirements have grown unevenly. Data organization and research, along with enterprise architecture design, show the steepest jumps in required experience, in some cases more than doubling since 2023. Meanwhile, some coding and deployment skills have seen required experience fall. This split makes sense: the skills gaining depth tend to involve judgment and architectural decision-making, while those losing ground tend to be more mechanical, the kind of work increasingly assisted by AI coding tools.

The Role of AI Adoption

The pattern sharpens when postings are split by whether they mention AI usage in the listed responsibilities. Controlling for seniority, industry, role, and posting length, AI-referencing postings list modestly fewer required skills and modestly more required experience than comparable postings that don't. The gap is small, a few percentage points either way, but it points to a plausible mechanism: as AI tools absorb routine, codifiable tasks, employers seem to be consolidating remaining human work around narrower, higher-judgment skills, and seeking people who have already mastered them.

Implications for the Labor Market

The takeaway is not simply good news for candidates, even though a falling skill count might first read that way. The skills retained in postings are disproportionately the ones built on years of accumulated expertise: data architecture, systems design, applied judgment. These are not the skills junior candidates are best positioned to offer. The market isn't becoming less demanding. It's becoming demanding in a different way, one that rewards depth over range.

This matters most for entry-level and early-career workers, whose traditional advantage has rested on adaptability rather than depth. If this trend continues, it points toward a labor market built around fewer, more specialized, more experience-intensive roles, a shift with consequences for career strategy, education, and workforce policy alike.

---

Based on analysis of technology job postings data published by Revelio Labs, September 2026.