New AI job impact index ranks software QA testers and writers as most affected: as a junior dev, what skills actually hold up?
The ASA AI Impact Index from October 23, 2026 puts computer and mathematical occupations at the top for AI impact, and names software QA analysts and testers, writers, and data-entry keyers as the three most-affected occupations. Stanfords update found employment of 22 to 25-year-olds in the most AI-exposed jobs is about 19% below trend, while experienced workers show no comparable gap. I am early in my career and trying to decide where to invest learning time. Is the move to get better at reviewing and steering AI-written code, or to move toward the parts of the job AI cannot touch - architecture, debugging production, talking to users? What are people actually seeing in hiring right now?
1 answer
- Review-and-steer is the skill that survives
The index measures task exposure, not job loss, and the pattern is consistent: AI absorbs execution, not judgment. Junior roles shrink because the first rung - writing boilerplate, running test suites, drafting docs - is what the tools do now. The durable skills are the ones around it: reading AI-written code critically, debugging production systems, and deciding what to build. Learn to review a diff the model wrote and find the bug in it. That is the job now.