AI agents aren't replacing software engineering but expanding it far beyond code, researchers argue

2026-04-26 · Source: The Decoder · Field: Technology & Digital — Software Development & Engineering, Artificial Intelligence & Machine Learning · Depth: Intermediate, short

Summary

Researchers from Chalmers University of Technology and the Volvo Group argue that AI agents are not making software developers obsolete but are instead expanding the scope of software engineering. They introduce the concept of "semi-executable artifacts," such as prompts, workflows, and decision routines, which shape system behavior but rely on human or probabilistic interpretation. To illustrate this expansion, they propose a "Semi-Executable Stack" model comprising six rings. This stack ranges from traditional code (ring 1) at the core to social and institutional factors like the EU AI Act (ring 6) at the periphery. The study suggests that the developer's role is evolving from primarily writing code to focusing on strategic decisions, validation, and ongoing system maintenance, particularly in the outer rings where engineering methods are less developed.

Key takeaway

For CTOs and VPs of Engineering evaluating AI integration, recognize that AI agents fundamentally expand the engineering surface area beyond traditional code. Your teams should shift focus from mere code generation efficiency to establishing robust engineering practices for "semi-executable artifacts" and the outer rings of the "Semi-Executable Stack," including governance, validation, and maintenance, to avoid local productivity gains that miss broader organizational redesign opportunities.

Key insights

AI agents expand software engineering beyond code, necessitating new methods for "semi-executable artifacts."

Principles

• AI agents expand software engineering scope.
• Scale and everyday AI deployments drive value.
• Nuanced judgment becomes more valuable.

Method

The "Semi-Executable Stack" model organizes software engineering into six rings, from executable code to societal fit, highlighting areas where human interpretation and new engineering methods are increasingly critical.

In practice

• Focus on validating and governing AI systems.
• Address prompt drift through testing and monitoring.
• Develop engineering practices for outer rings.

Topics

• AI Agents
• Software Engineering
• Semi-Executable Stack
• Prompt Engineering
• Developer Role Shift

Best for: CTO, VP of Engineering/Data, Director of AI/ML, Software Engineer, AI Engineer, Research Scientist

Related on AIssential

• The Semi-Executable Stack: Agentic Software Engineering and the Expanding Scope of SE — Software engineering's scope expands beyond code to semi-executable artifacts, requiring new models for understanding AI's impact.
• The World Needs More Software Engineers — AI agents will expand software engineering demand across the economy by making more projects economically viable.
• From Coder to Manager: Navigating the Shift to Agentic Engineering with Notion Co-Founder Simon Last — Notion is evolving into an agent-centric platform, shifting from human-direct work to human-managed AI agents.
• Most Influential ArXiv (Software Engineering) Papers (2026-04 Version) — LLMs and AI agents are rapidly transforming software engineering, driving innovation in code generation, testing, and repair.
• Cognition’s Scott Wu says AI coding agents shouldn’t replace humans — AI coding agents like Devin aim to augment human programmers by automating "toil," not replace them.
• Steve Yegge on AI Agents and the Future of Software Engineering — AI is rapidly transforming software development, making traditional coding skills less critical while increasing engineer productivity.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Decoder.