An article discussing 'feng shui refactoring'—superficial code reorganizations that rearrange structure without improving functionality or maintainability. The piece contrasts pseudo-refactoring (renaming, moving files, reorganizing directories) with genuine refactoring that removes duplication, clarifies business logic, or simplifies system behavior.
A code review system using two adversarial LLM agents—one finding vulnerabilities, one challenging findings with counter-evidence—to overcome the problem of LLMs agreeing with themselves and to produce zero-noise validated security findings.
This article argues that while AI excels at code generation, it cannot make architectural and engineering decisions, resulting in poorly-structured codebases shaped by prompt sequences rather than deliberate design. The lack of decision-making creates technical debt that compounds over time, requiring human architects to provide oversight and establish consistent patterns.