Original work, written down.
Research and working notes documenting original work in AI-native software engineering, organizational intelligence, and software architecture.
Independent research, not formally peer reviewed. Each note includes a provenance statement.
Monowire by Samiel: Human ↔ Machine Semantic Divergence Detection
A prior-art review and unverified research hypothesis not a claimed capability checked against intent-drift detection, requirements traceability, and grounding theory. A narrow, real gap, and a falsifiable experiment to test it.
Read the researchMonowire by Samiel: Research Map
Every architectural mechanism in Monowire checked one at a time against the existing literature classified as established, narrowly different, or genuinely worth a full paper. Two of nine mechanisms clear the bar.
Read the researchSkill-Governed Execution Graphs for Multi-Agent Software Engineering: A Monowire by Samiel Study
A hypothesis about assigning procedural knowledge upstream, during organizational documentation, rather than leaving execution agents to discover it at runtime checked against a fast-moving 2026 literature on skill compilation and execution graphs.
Read the researchDocumentation as an Intermediate Representation for AI-Native Software Engineering
A proposed workflow Idea, Conversation, Documentation, Knowledge Compilation, Creation tested honestly against the existing literature rather than assumed to be new.
Read the researchOrganizations Don't Have an AI Problem. They Have a Memory Problem.
The real bottleneck in AI adoption isn't model choice or context length. It's what happens to an organization's knowledge after the meeting ends.
Read the research