Gergely Orosz (2025), The Pragmatic Engineer (blog / newsletter excerpt) — https://blog.pragmaticengineer.com/software-engineer-jobs-five-year-lo…
Key takeaways
- The author reports that software-developer job postings on Indeed are about 35% below January 2020 and about 3.5× below the mid-2022 peak, the largest boom-and-bust of any sector tracked, while all postings are up about 10%. He attributes most of the drop to the end of zero interest rates, with GenAI productivity expectations, over-hiring in 2021–22, and leaner teams as secondary and uncertain factors.
Seeded from the abstract and a skim, before the reading in NOTE-039. What follows is what the work says about itself.
(Blog post; no abstract.) Using Indeed’s aggregated postings index (January 2020 = 100) via FRED, the author shows software development vacancies at about 65% of the 2020 baseline in the US, with similar drops in the UK, France and Germany and an exception in Australia. He compares other sectors and weighs explanations: interest rates, US Section 174 amortization, over-hiring during the pandemic, GenAI tooling, and smaller teams. He also questions how representative Indeed’s data is.
Standing in the record
Filed from the survey of 2026-09-25 of work the anthology set aside as out of scope (tier C): 510 seconds of active reading over 1 session in the papers-feed tracker. It was filed Deferred, unread. NOTE-039 is the close reading of 2026-09-25, and it placed the work: Active — its headline figures reproduce from Indeed Hiring Lab’s postings series, which is the primary source and is not held here; several secondary claims (the sector comparison, the hospitality figure, the peak month) do not.
When seeded, before the reading in NOTE-039, the entry gave a priority for a deeper reading: low — a single 510 s session on a blog post the skim fully covers, and the triage says drop.
What the seeding asked a deeper reading to check (NOTE-039 is that reading):
- It is an attributed practitioner’s reading of labour-market data, not a study. The claims about AI’s labour effect are speculative by the author’s own account.
- If this topic is filed, the primary sources (Indeed Hiring Lab / FRED series) and peer-reviewed labour-economics work would be better anchors.
Access when seeded: I fetched and read the full post from blog.pragmaticengineer.com. The page metadata gives datePublished 2025-02-20T20:27Z and dateModified 2025-05-05. The post says it is an excerpt from a paid issue of “The Pulse” sent a week earlier, and that earlier issue was not reached. The charts are images (Indeed data via FRED) and were not inspected.