About

The End of Software Engineering and the Rise of Agentic Engineering argues that software engineering is being redefined as agentic engineering. The unit of work is moving from hand-authored code to specifications, contracts, orchestrated agent activity, and verification of machine-produced artifacts.

This is a sourced practitioner series with a four-repository case study, not a systematic literature review. Empirical studies support measured claims. Pinned source code supports implementation claims. Neither is used as evidence for the other.

Engineering judgment becomes the control layer: practitioners define objectives and constraints, compose agents and tools, govern effects, and decide what evidence is sufficient for acceptance.

The five parts

Pillar	Role in the argument	Evidence base	Part
Claim	Definition and evidence for disciplinary redefinition	Historical and empirical sources	Part I
Anatomy	Primitives, boundaries, and composition operators	Four first-party systems	Part II
Loop	Bounded plan, act, and verify composition	Two code bases and one hosted product	Part III
Teams	Coordination patterns and shared-resource rules	Elixir and Rust runtimes	Part IV
Substrate	Context, memory, provenance, and governed data	Four first-party systems	Part V

Evidence posture

The papers report results before interpreting them, distinguish vendor claims from independent evidence, and state when a repository observation only proves that a mechanism exists. Productivity variation, quality risk, and review cost determine how the redefinition must be governed and where human control must remain explicit.

Author

Matthew Long, The YonedaAI Collaboration, YonedaAI Research Collective, Chicago, IL. matthew@yonedaai.com · yonedaai.com