The Shift from Technical Execution to Evaluation
AI-generated image · Bay Street Wire
Princeton professor Arvind Narayanan argues that while AI may not automate jobs entirely, it will fundamentally revalue the skills required for researchers and developers.
As artificial intelligence continues to advance, the central concern for technical professionals is no longer just the existence of work, but the nature of the skills that remain valuable. In an invited talk scheduled for July 8, 2026, as first reported by Hacker News, Arvind Narayanan, a professor of computer science at Princeton University and director of the Center for Information Technology Policy, addresses the question of what will be left for humans to work on.
Narayanan posits an "AI as normal technology" thesis, which suggests that significant bottlenecks exist between the improvement of AI capabilities and the actual automation of jobs. According to the presentation abstract, AI should be viewed as an augmentation technology rather than one of pure automation. However, this shift will lead to the devaluation of purely technical skills.
Narayanan argues that human effort will migrate away from traditional problem solving and model development. Instead, the focus will shift toward tasks that are less verifiable, such as the creation of scaffolds, evaluation, and monitoring. In the research sector, Narayanan suggests effort will move toward conceptual progress and the act of asking questions. Within the industry, he notes that domain knowledge, relational skills, and normative and aesthetic judgment will increase in importance.