Since Claude Code launched in February 2025, US software development job postings on Indeed have grown 15% while overall postings fell 7%. The mechanism is counterintuitive but coherent: agentic AI tools are increasing the surface area of software problems worth solving, which requires more humans to scope, direct, and validate the work. The displacement narrative that has dominated workforce strategy conversations for three years is not just incomplete, it is directionally wrong for this sector. Any brand that froze technical hiring on AI-displacement logic in 2025 has just been outpaced.
Two stories today point to the same structural split. Software hiring is up 15% in the agentic AI era, while overall job postings fell: the people actively building with AI tools are not being replaced, they are being multiplied.
Meanwhile, the EU's action against Meta reframes passive attention infrastructure as a legal liability, not a business model. The split is between organisations that use AI as a force multiplier for human capability and those that use it as a substitute for engagement. The former are growing headcount; the latter are growing regulatory risk.
For people who’d rather be early and wrong than late and safe.