Research & Perspectives
Expert analysis on enterprise security, identity management, and the practical application of AI automation.
We scraped and deduplicated 4,012 live AI and machine learning job adverts across LinkedIn, Reed and Totaljobs. Python still rules, large language model experience has become a baseline, and US employers pay far more than UK ones for the same work.
For most of computing history, shipping a working application meant hiring engineers. A growing share of internal business software is now built by the people who need it, working alongside AI agents, and that changes who counts as a builder.
AI inference costs have collapsed, yet total AI spending keeps climbing. A 160-year-old idea about coal, Jevons Paradox, explains why falling prices tend to expand demand and work rather than shrink them.
Two of AI's loudest voices predicted mass white-collar job losses, then quietly softened. A 160-year-old idea about coal explains why cheaper automation has so far meant more work, not less.
Predictions of mass AI job losses have started to soften. Bank machines, spreadsheets and a 160-year-old economic idea help explain why cheaper automation has tended to create work rather than remove it.
New insights published monthly. No spam, unsubscribe anytime.