Writings

A field report applying a scientific lens to a practical question: where is the efficient frontier of AI in software engineering today? This is not a copy-paste method. It is a set of principles, drawn from information theory and control theory and tested against 600 million tokens of real runs, meant to help engineering leaders judge where agentic AI pays and where it does not. The theory does not prove the conclusions; the runs do. The theory explains why they hold.

How do you scale a tech department? Whether you are a fast-growing startup, a corporate launching a new project, or integrating teams post-M&A, the principles remain the same. This article breaks down 'Organizational Mitosis', explaining why adding managers creates latency, why teams should not collaborate, and how to align software architecture with business goals to avoid building premature cathedrals.

Rigid engineering standards create an illusion of control while increasing systemic fragility. AI is driving code costs to zero: the true bottleneck is the cognitive overhead of bureaucracy. This article explores why 'perfect order' is an insurance policy for mediocrity and how to architect managed autonomy instead.

Software engineering is deterministic; humans are not. Many organizations attempt to manage human productivity with the same logic they use to monitor servers, creating 'measurement theaters' that destroy value. This article explores why measuring the means instead of the goal encourages gaming the system and why aligning team topologies with business KPIs is the only way to achieve real scalability.

When communication breaks down during a scale-up phase, the corporate reflex is to hire another layer of management. They create specialized silos and endless coordination roles. This is a fatal architectural flaw. Adding hierarchy does not solve misalignment. It simply forces decisions to escalate to people with zero context, adding massive latency to a broken system. This article explores why decisions must stay at the lowest possible level and why relying on alignment meetings is a symptom of failed engineering design.

Most organizations treat the 'Build vs. Buy' decision as a simple technical choice. In reality, it is a high-stakes capital allocation. This article explores why rebuilding non-core products is a form of capital destruction, how over-engineering creates certain expenses for hypothetical gains, and why a Head of Technology must act as a CFO of the technical engine.

Demotivation is a systemic failure that turns engineers into passive observers. When the search for meaning is stripped away by top-down solutions and rigid procedures, ownership vanishes. This article explores how to re-engineer motivation by focusing on direction rather than execution and ensuring every builder has skin in the game.

While everyone focuses on which roles will die because of AI, the real shift is the collapse of code production costs. Software is becoming a disposable commodity. This article explores the historical evolution of abstraction, the extinction of the super-specialist, and why injecting AI into a broken and highly coupled organization will only accelerate its failure.